Tim Belzer – Page 26

California’s newly signed AI law just gave Big Tech exactly what it wanted

AI, AI development tools, AI ethics, AI law, AI regulation, AI safety, Andreessen Horowitz, Anthropic, Biz & IT, California, Fei-Fei Li, Gavin Newsom, Google, Jack Clark, machine learning, Meta, openai, Policy, Scott Wiener / Tim Belzer / September 30, 2025

On Monday, California Governor Gavin Newsom signed the Transparency in Frontier Artificial Intelligence Act into law, requiring AI companies to disclose their safety practices while stopping short of mandating actual safety testing. The law requires companies with annual revenues of at least $500 million to publish safety protocols on their websites and report incidents to state authorities, but it lacks the stronger enforcement teeth of the bill Newsom vetoed last year after tech companies lobbied heavily against it.

The legislation, S.B. 53, replaces Senator Scott Wiener’s previous attempt at AI regulation, known as S.B. 1047, that would have required safety testing and “kill switches” for AI systems. Instead, the new law asks companies to describe how they incorporate “national standards, international standards, and industry-consensus best practices” into their AI development, without specifying what those standards are or requiring independent verification.

“California has proven that we can establish regulations to protect our communities while also ensuring that the growing AI industry continues to thrive,” Newsom said in a statement, though the law’s actual protective measures remain largely voluntary beyond basic reporting requirements.

According to the California state government, the state houses 32 of the world’s top 50 AI companies, and more than half of global venture capital funding for AI and machine learning startups went to Bay Area companies last year. So while the recently signed bill is state-level legislation, what happens in California AI regulation will have a much wider impact, both by legislative precedent and by affecting companies that craft AI systems used around the world.

Transparency instead of testing

Where the vetoed SB 1047 would have mandated safety testing and kill switches for AI systems, the new law focuses on disclosure. Companies must report what the state calls “potential critical safety incidents” to California’s Office of Emergency Services and provide whistleblower protections for employees who raise safety concerns. The law defines catastrophic risk narrowly as incidents potentially causing 50+ deaths or $1 billion in damage through weapons assistance, autonomous criminal acts, or loss of control. The attorney general can levy civil penalties of up to $1 million per violation for noncompliance with these reporting requirements.

California’s newly signed AI law just gave Big Tech exactly what it wanted Read More »

On Dwarkesh Patel’s Podcast With Richard Sutton

Dwarkesh / Tim Belzer / September 30, 2025

This seems like a good opportunity to do some of my classic detailed podcast coverage.

The conventions are:

This is not complete, points I did not find of note are skipped.
The main part of each point is descriptive of what is said, by default paraphrased.
For direct quotes I will use quote marks, by default this is Sutton.
Nested statements are my own commentary.
Timestamps are approximate and from his hosted copy, not the YouTube version, in this case I didn’t bother because the section divisions in the transcript should make this very easy to follow without them.

Full transcript of the episode is here if you want to verify exactly what was said.

Well, that was the plan. This turned largely into me quoting Sutton and then expressing my mind boggling. A lot of what was interesting about this talk was in the back and forth or the ways Sutton lays things out in ways that I found impossible to excerpt, so one could consider following along with the transcript or while listening.

(0: 33) RL and LLMs are very different. RL is ‘basic’ AI. Intelligence and RL are about understanding your world. LLMs mimic people, they don’t figure out what to do.
1. RL isn’t strictly about ‘understanding your world’ except insofar as it is necessary to do the job. The same applies to LLMs, no?
2. To maximize RL signal you need to understand and predict the world, aka you need intelligence. To mimic people, you have to understand and predict them, which in turn requires understanding and predicting the world. Same deal.
(1: 19) Dwarkesh points out that mimicry requires a robust world model, indeed LLMs have the best world models to date. Sutton disagrees, you’re mimicking people, and he questions that people have a world model. He says a world model would allow you to predict what would happen, whereas people can’t do that.
1. People don’t always have an explicit world model, but sometimes they do, and they have an implicit one running under the hood.
2. Even if people didn’t have a world model in their heads, their outputs in a given situation depend on the world, which you then have to model, if you want to mimic those humans.
3. People predict what will happen all the time, on micro and macro levels. On the micro level they are usually correct. On sufficiently macro levels they are often wrong, but this still counts. If the claim is ‘if you can’t reliably predict what will happen then you don’t have a model’ then we disagree on what it means to have a model, and I would claim no such-defined models exist at any interesting scale or scope.
(1: 38) “What we want, to quote Alan Turing, is a machine that can learn from experience, where experience is the things that actually happen in your life. You do things, you see what happens, and that’s what you learn from. The large language models learn from something else. They learn from “here’s a situation, and here’s what a person did”. Implicitly, the suggestion is you should do what the person did.”
1. That’s not the suggestion. If [X] is often followed by [Y], then the suggestion is not ‘if [X] then you should do [Y]’ it it ‘[X] means [Y] is likely’ so yes if you are asked ‘what is likely after [X]’ it will respond [Y] but it will also internalize everything implied by this fact and the fact is not in any way normative.
2. That’s still ‘learning from experience’ it’s simply not continual learning.
3. Do LLMs do continual learning, e.g. ‘from what actually happens in your life’ in particular? Not in their current forms, not technically, but there’s no inherent reason they couldn’t, you’d just do [mumble] except that doing so would get rather expensive.
4. You can also have them learn via various forms of external memory, broadly construed, including having them construct programs. It would work.
5. Not that it’s obvious that you would want an LLM or other AI to learn specifically from what happens in your life, as opposed to learning from things that happen in lives in general plus having context and memory.
(2: 39) Dwarkesh responds with a potential crux that imitation learning is a good prior or reasonable approach, and gives the opportunity to get answers right sometimes, then you can train on experience. Sutton says no, that’s the LLM perspective, but the LLM perspective is bad. It’s not ‘actual knowledge.’ You need continual learning so you need to know what’s right during interactions, but the LLM setup can’t tell because there’s no ground truth, because you don’t have a prediction about what will happen next.
1. I don’t see Dwarkesh’s question as a crux.
2. I think Sutton’s response is quite bad, relying on invalid sacred word defenses.
3. I think Sutton wants to draw a distinction between events in the world and tokens in a document. I don’t think you can do that.
4. There is no ‘ground truth’ other than the feedback one gets from the environment. I don’t see why a physical response is different from a token, or from a numerical score. The feedback involved can come from anywhere, including from self-reflection if verification is easier than generation or can be made so in context, and it still counts. What is this special ‘ground truth’?
5. Almost all feedback is noisy because almost all outcomes are probabilistic.
6. You think that’s air you’re experiencing breathing? Does that matter?
(5: 29) Dwarkesh points out you can literally ask “What would you anticipate a user might say in response?” but Sutton rejects this because it’s not a ‘substantive’ prediction and the LLM won’t be ‘surprised’ or “they will not change because an unexpected thing has happened. To learn that, they’d have to make an adjustment.”
1. Why is this ‘not substantive’ in any meaningful way, especially if it is a description of a substantive consequence, which speech often is?
2. How is it not ‘surprise’ when a low-probability token appears in the text?
3. There are plenty of times a human is surprised by an outcome but does not learn from it out of context. For example, I roll a d100 and get a 1. Okie dokie.
4. LLMs do learn from a surprising token in training. You can always train. This seems like an insistence that surprise requires continual learning? Why?
Dwarkesh points out LLMs update within a chain-of-thought, so flexibility exists in a given context. Sutton reiterates they can’t predict things and can’t be surprised. He insists that “The next token is what they should say, what the actions should be. It’s not what the world will give them in response to what they do.”
1. What is Sutton even saying, at this point?
2. Again, this distinction that outputting or predicting a token is distinct from ‘taking an action,’ and getting a token back is not the world responding.
3. I’d point out the same applies to the rest of the tokens in context without CoT.
(6: 47) Sutton claims something interesting, that intelligence requires goals, “I like John McCarthy’s definition that intelligence is the computational part of the ability to achieve goals. You have to have goals or you’re just a behaving system.” And he asks Dwarkesh is he agrees that LLMs don’t have goals (or don’t have ‘substantive’ goals, and that next token prediction is not a goal, because it doesn’t influence the tokens.
1. Okay, seriously, this is crazy, right?
2. What is this ‘substantive’ thing? If you say something on the internet, it gets read in real life. It impacts real life. It causes real people to do ‘substantive’ things, and achieving many goals within the internet requires ‘substantive’ changes in the offline world. If you’re dumb on the internet, you’re dumb in real life. If you die on the internet, you die in real life (e.g. in the sense of an audience not laughing, or people not supporting you, etc).
3. I feel dumb having to type that, but I’m confused what the confusion is.
4. Of course next token prediction is a goal. You try predicting the next token (it’s hard!) and then tell me you weren’t pursuing a goal.
5. Next token prediction does influence the tokens in deployment because the LLM will output the next most likely token, which changes what tokens come after, its and the user’s, and also the real world.
6. Next token prediction does influence the world in training, because the feedback on that prediction’s accuracy will change the model’s weights, if nothing else. Those are part of the world.
7. If intelligence requires goals, and something clearly displays intelligence, then that something must have a goal. If you conclude that LLMs ‘don’t have intelligence’ in 2025, you’ve reached a wrong conclusion. Wrong conclusions are wrong. You made a mistake. Retrace your steps until you find it.
Dwarkesh next points out you can do RL on top of LLMs, and they get IMO gold, and asks why Sutton still doesn’t think that is anything. Sutton doubles down that math operations still aren’t the empirical world, doesn’t count.
1. Are you kidding me? So symbolic things aren’t real, period, and manipulating them can’t be intelligence, period?
Dwarkesh notes that Sutton is famously the author of The Bitter Lesson, which is constantly cited as inspiring and justifying the whole ‘stack more layers’ scaling of LLMs that basically worked, yet Sutton doesn’t see LLMs as ‘bitter lesson’ pilled. Sutton says they’re also putting in lots of human knowledge, so kinda yes kinda no, he expects that new systems that ‘learn from experience’ and ‘perform much better’ and are ‘more scalable’ to then be another instance of the Bitter Lesson?
1. This seems like backtracking on the Bitter Lesson? At least kinda. Mostly he’s repeating that LLMs are one way and it’s the other way, and therefore Bitter Lesson will be illustrated the other way?
“In every case of the bitter lesson you could start with human knowledge and then do the scalable things. That’s always the case. There’s never any reason why that has to be bad. But in fact, and in practice, it has always turned out to be bad. People get locked into the human knowledge approach, and they psychologically… Now I’m speculating why it is, but this is what has always happened. They get their lunch eaten by the methods that are truly scalable.”
1. I do not get where ‘truly scalable’ is coming from here, as it becomes increasingly clear that he is using words in a way I’ve never seen before.
2. If anything it is the opposite. The real objection is training efficiency, or failure to properly update from direct relevant experiences, neither of which has anything to do with scaling.
3. I also continue not to see why there is this distinction ‘human knowledge’ versus other information? Any information available to the AI can be coded as tokens and be put into an LLM, regardless of its ‘humanness.’ The AI can still gather or create knowledge on its own, and LLMs often do.
“The scalable method is you learn from experience. You try things, you see what works. No one has to tell you. First of all, you have a goal. Without a goal, there’s no sense of right or wrong or better or worse. Large language models are trying to get by without having a goal or a sense of better or worse. That’s just exactly starting in the wrong place.”
1. Again, the word ‘scaling’ is being used in a completely alien manner here. He seems to be trying to say ‘successful’ or ‘efficient.’
2. You have to have a ‘goal’ in the sense of a means of selecting actions, and a way of updating based on those actions, but in this sense LLMs in training very obviously have ‘goals’ regardless of whether you’d use that word that way.
3. Except Sutton seems to think this ‘goal’ needs to exist in some ‘real world’ sense or it doesn’t count and I continue to be boggled by this request, and there are many obvious counterexamples, but I risk repeating myself.
4. No sense of better or worse? What do you think thumbs up and down are? What do you think evaluators are? Does he not think an LLM can do evaluation?

Sutton has a reasonable hypothesis that a different architecture, that uses a form of continual learning and that does so via real world interaction, would be an interesting and potentially better approach to AI. That might be true.

But his uses of words do not seem to match their definitions or common usage, his characterizations of LLMs seem deeply confused, and he’s drawing a bunch of distinctinctions and treating them as meaningful in ways that I don’t understand. This results in absurd claims like ‘LLMs are not intelligent and do not have goals’ and that feedback from digital systems doesn’t count and so on.

It seems like a form of essentialism, the idea that ‘oh LLMs can never [X] because they don’t [Y]’ where when you then point (as people frequently do) to the LLM doing [X] and often also doing [Y] and they say ‘la la la can’t hear you.’

Dwarkesh claims humans initially do imitation learning, Sutton says obviously not. “When I see kids, I see kids just trying things and waving their hands around and moving their eyes around. There’s no imitation for how they move their eyes around or even the sounds they make. They may want to create the same sounds, but the actions, the thing that the infant actually does, there’s no targets for that. There are no examples for that.”
1. GPT-5 Thinking says partly true, but only 30% in the first months, more later on. Gemini says yes. Claude says yes: “Imitation is one of the core learning mechanisms from birth onward. Newborns can imitate facial expressions within hours of birth (tongue protrusion being the classic example). By 6-9 months, they’re doing deferred imitation – copying actions they saw earlier. The whole mirror neuron system appears to be built for this.”
2. Sutton’s claim seems clearly so strong as to be outright false here. He’s not saying ‘they do more non-imitation learning than imitation learning in the first few months,’ he is saying ‘there are no examples of that’ and there are very obviously examples of that. Here’s Gemini: “Research has shown that newborns, some just a few hours old, can imitate simple facial expressions like sticking out their tongue or opening their mouth. This early imitation is believed to be a reflexive behavior that lays the groundwork for more intentional imitation later on.”
“School is much later. Okay, I shouldn’t have said never. I don’t know, I think I would even say that about school. But formal schooling is the exception. You shouldn’t base your theories on that.” “Supervised learning is not something that happens in nature. Even if that were the case with school, we should forget about it because that’s some special thing that happens in people.”
1. At this point I kind of wonder if Sutton has met humans?
2. As in, I do imitation learning. All. The Time. Don’t you? Like, what?
3. As in, I do supervised learning. All. The. Time. Don’t you? Like, what?
4. A lot of this supervised and imitation learning happens outside of ‘school.’
5. You even see supervised learning in animals, given the existence of human supervisors who want to teach them things. Good dog! Good boy!
6. You definitely see imitation learning in animals. Monkey see, monkey do.
7. The reason not to do supervised learning is the cost of the supervisor, or (such as in the case of nature) their unavailability. Thus nature supervises, instead.
8. The reason not to do imitation learning in a given context is the cost of the thing to imitate, or the lack of a good enough thing to imitate to let you continue to sufficiently progress.
“Why are you trying to distinguish humans? Humans are animals. What we have in common is more interesting. What distinguishes us, we should be paying less attention to.” “I like the way you consider that obvious, because I consider the opposite obvious. We have to understand how we are animals. If we understood a squirrel, I think we’d be almost all the way there to understanding human intelligence. The language part is just a small veneer on the surface.”
1. Because we want to create something that has what only humans have and humans don’t, which is a high level of intelligence and ability to optimize the arrangements of atoms according to our preferences and goals.
2. Understanding an existing intelligence is not the same thing as building a new intelligence, which we have also managed to build without understanding.
3. The way animals have (limited) intelligence does not mean this is the One True Way that intelligence can ever exist. There’s no inherent reason an AI needs to mimic a human let alone an animal, except for imitation learning, or in ways we find this to be useful. We’re kind of looking for our keys under the streetlamp here, while assuming there are no keys elsewhere, and I think we’re going to be in for some very rude (or perhaps pleasant?) surprises.
4. I don’t want to make a virtual squirrel and scale it up. Do you?
The process of humans learning things over 10k years a la Henrich, of figuring out a many-step long process, where you can’t one-shot the reasoning process. This knowledge evolves over time, and is passed down through imitation learning, as are other cultural practices and gains. Sutton agrees, but calls this a ‘small thing.’
1. You could of course one-shot the process with sufficient intelligence and understanding of the world, what Henrich is pointing out is that in practice this was obviously impossible and not how any of this went down.
2. Seems like Sutton is saying again that the difference between humans and squirrels is a ‘small thing’ and we shouldn’t care about it? I disagree.
They agree that mammals can do continual learning and LLMs can’t. We all agree that Moravec’s paradox is a thing.
1. Moravec’s paradox is misleading. There will of course be all four quadrants of things, where for each of [AI, human] things will be [easy, hard].
2. The same is true for any pair of humans, or any pair of AIs, to a lesser degree.
3. The reason it is labeled a paradox is that there are some divergences that look very large, larger than one might expect, but this isn’t obvious to me.

“The experiential paradigm. Let’s lay it out a little bit. It says that experience, action, sensation—well, sensation, action, reward—this happens on and on and on for your life. It says that this is the foundation and the focus of intelligence. Intelligence is about taking that stream and altering the actions to increase the rewards in the stream…. This is what the reinforcement learning paradigm is, learning from experience.”
1. Can be. Doesn’t have to be.
2. A priori knowledge exists. Paging Descartes’ meditator! Molyneux’s problem.
3. Words, written and voiced, are sensation, and can also be reward.
4. Thoughts and predictions, and saying or writing words, are actions.
5. All of these are experiences. You can do RL on them (and humans do this).
Sutton agrees that the reward function is arbitrary, and can often be ‘seek pleasure and avoid pain.’
1. That sounds exactly like ‘make number go up’ with extra steps.
Sutton wants to say ‘network’ instead of ‘model.’
1. Okie dokie, this does cause confusion with ‘world models’ that minds have, as Sutton points out later, so using the same word for both is unfortunate.
2. I do think we’re stuck with ‘model’ here, but I’d be happy to support moving to ‘network’ or another alternative if one got momentum.
He points out that copying minds is a huge cost savings, more than ‘trying to learn from people.’
1. Okie dokie, again, but these two are not rivalrous actions.
2. If anything they are complements. If you learn from general knowledge and experiences it is highly useful to copy you. If you are learning from local particular experiences then your usefulness is likely more localized.
3. As in, suppose I had a GPT-5 instance, embodied in a humanoid robot, that did continual learning, which let’s call Daneel. I expect that Daneel would rapidly become a better fit to me than to others.
4. Why wouldn’t you want to learn from all sources, and then make copies?
5. One answer would be ‘because to store all that info the network would need to be too large and thus too expensive’ but that again pushes you in the other direction, and towards additional scaffolding solutions.
They discuss temporal difference learning and finding intermediate objectives.
Sutton brings up the ‘big world hypothesis’ where to be maximally useful a human or AI needs particular knowledge of a particular part of the world. In continual learning the knowledge goes into weights. “You learn a policy that’s specific to the environment that you’re finding yourself in.”
1. Well sure, but there are any number of ways to get that context, and to learn that policy. You can even write the policy down (e.g. in claude.md).
2. Often it would be actively unwise to put that knowledge into weights. There is a reason humans will often use forms of external memory. If you were planning to copy a human into other contexts you’d use it even more.

Sutton lays out the above common model of the agent. The new claim seems to be that you learn from all the sensation you receive, not just from the reward. And there is emphasis on the importance of the ‘transition model’ of the world.
1. I once again don’t see the distinction between this and learning from a stream of tokens, whether one or two directional, or even from contemplation, where again (if you had an optimal learning policy) you would pay attention to all the tokens and not only to the formal reward, as indeed a human does when learning from a text, or from sending tokens and getting tokens back in various forms.
2. In terms of having a ‘transition model,’ I would say that again this is something all agents or networks need similarly, and can ‘get away with not having’ to roughly similar extents.

So do humans.

Sutton claims people live in one world that may involve chess or Atari games and and can generalize across not only games but states, and will happen whether that generalization is good or bad. Whereas gradient descent will not make you generalize well, and we need algorithms where the generalization is good.
1. I’m not convinced that LLMs or SGD generalize out-of-distribution (OOD) poorly relative to other systems, including humans or RL systems, once you control for various other factors.
2. I do agree that LLMs will often do pretty dumb or crazy things OOD.
3. All algorithms will solve the problem at hand. If you want that solution to generalize, you need to either make the expectation of such generalization part of the de facto evaluation function, develop heuristics and methods that tend to lead to generalization for other reasons, or otherwise incorporate the general case, or choose or get lucky with a problem where the otherwise ‘natural’ solution does still generalize.
“Well maybe that [LLMs] don’t need to generalize to get them right, because the only way to get some of them right is to form something which gets all of them right. If there’s only one answer and you find it, that’s not called generalization. It’s just it’s the only way to solve it, and so they find the only way to solve it. But generalization is when it could be this way, it could be that way, and they do it the good way.”
1. Sutton only thinks you can generalize given the ability to not generalize, the way good requires the possibility of evil. It is a relative descriptor.
2. I don’t understand why you’d find that definition useful or valid. I care about the generality of your solution in practice, not whether there was a more or less general alternative solution also available.
3. Once again there’s this focus on whether something ‘counts’ as a thing. Yes, of course, if the only or simplest or easiest way to solve a special case is to solve the general case, which often happens, and thus you solve the general case, and this happens to solve a bunch of problem types you didn’t consider, then you have done generalization. Your solution will work in the general case, whether or not you call that OOD.
4. If there’s only one answer and you find it, you still found it.
5. This seems pretty central. SGD or RL or other training methods, of both humans and AIs, will solve the problem you hand to them. Not the problem you meant to solve, the problem and optimization target you actually presented.
6. You need to design that target and choose that method, such that this results in a solution that does what you want it to do. You can approach that in any number of ways, and ideally (assuming you want a general solution) you will choose to set the problem up such that the only or best available solution generalizes, if necessary via penalizing solutions that don’t in various ways.
Sutton claims coding agents trained via SGD will only find solutions to problems they have seen, and yes sometimes the only solution will generalize but nothing in their algorithms will cause them to choose solutions that generalize well.
1. Very obviously coding agents generalize to problems they haven’t seen.
2. Not fully to ‘all coding of all things’ but they generalize quite a bit and are generalizing better over time. Seems odd to deny this?
3. Sutton is making at least two different claims.
4. The first claim is that coding agents only find solutions to problems they have seen. This is at least a large overstatement.
5. The second claim is that the algorithms will not cause the network to choose solutions that generalize well over alternative solutions that don’t.
6. The second claim is true by default. As Sutton notes, sometimes the default or only solution does indeed generalize well. I would say this happens often. But yeah, sometimes by default this isn’t true, and then by construction and default there is nothing pushing towards finding the general solution.
7. Unless you design the training algorithms and data to favor the general solution. If you select your data well, often you can penalize or invalidate non-general solutions, and there are various algorithmic modifications available.
8. One solution type is giving the LLM an inherent preference for generality, or have the evaluator choose with a value towards generality, or both.
9. No, it isn’t going to be easy, but why should it be? If you want generality you have to ask for it. Again, compare to a human or an RL program. I’m not going for a more general solution unless I am motivated to do so, which can happen for any number of reasons.

Dwarkesh asks what has been surprising in AI’s big picture? Sutton says the effectiveness of artificial neural networks. He says ‘weak’ methods like search and learning have totally won over ‘strong’ methods that come from ‘imbuing a system with human knowledge.’
1. I find it interesting that Sutton in particular was surprised by ANNs. He is placing a lot of emphasis on copying animals, which seems like it would lead to expecting ANNs.
2. It feels like he’s trying to make ‘don’t imbue the system with human knowledge’ happen? To me that’s not what makes the ‘strong’ systems strong, or the thing that failed. The thing that failed was GOFAI, the idea that you would hardcode a bunch of logic and human knowledge in particular ways, and tell the AI how to do things, rather than letting the AI find solutions through search and learning. But that can still involve learning from human knowledge.
3. It doesn’t have to (see AlphaZero and previously TD-Gammon as Sutton points out), and yes that was somewhat surprising but also kind of not, in the sense that with More Dakka within a compact space like chess you can just solve the game from scratch.
4. As in: We don’t need to use human knowledge to master chess, because we can learn chess through self-play beyond human ability levels, and we have enough compute and data that way that we can do it ‘the hard way.’ Sure.

Dwarkesh asks what happens to scaling laws after AGI is created that can do AI research. Sutton says: “These AGIs, if they’re not superhuman already, then the knowledge that they might impart would be not superhuman.”
1. This seems like more characterization insistence combined with category error?
2. And it ignores or denies the premise of the question, which is that AGI allows you to scale researcher time with compute the same way we previously could scale compute spend in other places. Sutton agrees that doing bespoke work is helpful, it’s just that it doesn’t scale, but what if it did?
3. Even if the AGI is not ‘superhuman’ per se, the ability to run it faster and in parallel and with various other advantages means it can plausibly produce superhuman work in AI R&D. Already we have AIs that can do ‘superhuman’ tasks in various domains, even regular computers are ‘superhuman’ in some subdomains (e.g. arithmetic).
“So why do you say, “Bring in other agents’ expertise to teach it”, when it’s worked so well from experience and not by help from another agent?”
1. Help from another agent is experience. It can also directly create experience.
2. The context is chess where this is even more true.
3. Indeed, the way AlphaZero was trained was not to not involve other agents. The way AlphaZero was trained involved heavy use of other agents, except all those other agents were also AlphaZero.
Dwarkesh focuses specifically on the ‘billions of AI researchers’ case, Sutton says that’s an interesting case very different from today and The Bitter Lesson doesn’t have to apply. Better to ask questions like whether you should use compute to enhance a few agents or spread it around to spin up more of them, and how they will interact. “More questions, will it be possible to really spawn it off, send it out, learn something new, something perhaps very new, and then will it be able to be reincorporated into the original? Or will it have changed so much that it can’t really be done? Is that possible or is that not?”
1. I agree that things get strange and different and we should ask new questions.
2. Asking whether it is possible for an ASI (superintelligent AI) copy to learn something new and then incorporate it into the original seems like such a strange question.
  1. It presupposes this ‘continual learning’ thesis where the copy ‘learns’ the information via direct incorporation into its weights.
  2. It then assumes that passing on this new knowledge requires incorporation directly into weights or something weird?
  3. As opposed to, ya know, writing the insight down and the other ASI reading it? If ASIs are indeed superintelligent and do continual learning, why can’t they learn via reading? Wouldn’t they also get very good at knowing how to describe what they know?
  4. Also, yes, I’m pretty confident you can also do this via direct incorporation of the relevant experiences, even if the full Sutton model holds here in ways I don’t expect. You should be able to merge deltas directly in various ways we already know about, and in better ways that these ASIs will be able to figure out.
  5. Even if nothing else works, you can simply have the ‘base’ version of the ASI in question rerun the relevant experiences once it is verified that they led to something worthwhile, reducing this to the previous problem, says the mathematician.
Sutton also speculates about potential for corruption or insanity and similar dangers, if a central mind is incorporating the experiences or knowledge of other copies of itself. He expects this to be a big concern, including ‘mind viruses.’
1. Seems fun to think about, but nothing an army of ASIs couldn’t handle.
2. In general, when imagining scenarios with armies of ASIs, you have to price into everything the fact that they can solve problems way better than you.
3. I don’t think the associated ‘mind viruses’ in this scenario are fundamentally different than the problems with memetics and hazardous information we experience today, although they’ll be at a higher level.
4. I would of course expect lots of new unexpected and weird problems to arise.

It’s Sutton, so eventually we were going to have to deal with him being a successionist.

He argues that succession is inevitable for four reasons: Humanity is incapable of a united front, we will eventually figure out intelligence, we will eventually figure out superhuman intelligence, and it is inevitable that over time the most intelligent things around would gain intelligence and power.
1. We can divide this into two parts. Let “it” equal superintelligence.
2. Let’s call part one Someone Will Build It.
3. Let’s call part two If Anyone Builds It, Everyone Dies.
  1. Okay, sure, not quite as you see below, but mostly? Yeah, mostly.
4. Therefore, Everyone Will Die. Successionism is inevitable.
5. Part two is actually a very strong argument! It is simpler and cleaner and in many ways more convincing than the book’s version, at least in terms of establishing this as a baseline outcome. It doesn’t require (or give the impression it requires) any assumptions whatsoever about the way we get to superintelligence, what form that superintelligence takes, nothing.
6. I actually think this should be fully convincing of the weaker argument that by default (rather than inevitably) this happens, and that there is a large risk of this happening, and something has to go very right for it to not happen.
7. If you say ‘oh even if we do build superintelligence there’s no risk of this happening’ I consider this to be Obvious Nonsense and you not to be thinking.
8. I don’t think this argument is convincing that it is ‘inevitable.’ Facts not in evidence, and there seem like two very obvious counterexamples.
  1. Counterexample one is that if the intelligence gap is not so large in practical impact, other attributes can more than compensate for this. Other attributes, both mental and physical, also matter and can make up for this. Alas, this seems unlikely to be relevant given the expected intelligence gaps.
  2. Counterexample two is that you could ‘solve the alignment problem’ in a sufficiently robust sense that the more intelligent minds optimize for a world in which the less intelligent minds retain power in a sufficiently robust way. Extremely tricky, but definitely not impossible in theory.
9. However his definition of what is inevitable, and what counts as ‘succession’ here, is actually much more optimistic than I previously realized…
10. If we agree that If Anyone Builds It, Everyone Dies, then the logical conclusion is ‘Then Let’s Coordinate To Ensure No One Fing Build It.’
11. He claims nope, can’t happen, impossible, give up. I say, if everyone was convinced of part two, then that would change this.
“Put all that together and it’s sort of inevitable. You’re going to have succession to AI or to AI-enabled, augmented humans. Those four things seem clear and sure to happen. But within that set of possibilities, there could be good outcomes as well as less good outcomes, bad outcomes. I’m just trying to be realistic about where we are and ask how we should feel about it.”
1. If ‘AI-enhanced, augmented humans’ count here, well, that’s me, right now.
2. I mean, presumably that’s not exactly what he meant.
3. But yeah, conditional on us building ASIs or even AGIs, we’re at least dealing with some form of augmented humans.
4. Talk of ‘merge with the AI’ is nonsense, you’re not adding anything to it, but it can enhance you.
“I mark this as one of the four great stages of the universe. First there’s dust, it ends with stars. Stars make planets. The planets can give rise to life. Now we’re giving rise to designed entities. I think we should be proud that we are giving rise to this great transition in the universe.”
1. Designed is being used rather loosely here, but we get the idea.
2. We already have created designed things, and yeah that’s pretty cool.
“It’s an interesting thing. Should we consider them part of humanity or different from humanity? It’s our choice. It’s our choice whether we should say, “Oh, they are our offspring and we should be proud of them and we should celebrate their achievements.” Or we could say, “Oh no, they’re not us and we should be horrified.””
1. It’s not about whether they are ‘part of humanity’ or our ‘children.’ They’re not.
2. They can still have value. One can imagine aliens (as many stories have) that are not these things and still have value.
3. That doesn’t mean that us going away would therefore be non-horrifying.
“A lot of it has to do with just how you feel about change. If you think the current situation is really good, then you’re more likely to be suspicious of change and averse to change than if you think it’s imperfect. I think it’s imperfect. In fact, I think it’s pretty bad. So I’m open to change. I think humanity has not had a super good track record. Maybe it’s the best thing that there has been, but it’s far from perfect.” “I think it’s appropriate for us to really work towards our own local goals. It’s kind of aggressive for us to say, “Oh, the future has to evolve this way that I want it to.””
1. So there you have it.
2. I disagree.
“So we’re trying to design the future and the principles by which it will evolve and come into being. The first thing you’re saying is, “Well, we try to teach our children general principles which will promote more likely evolutions.” Maybe we should also seek for things to be voluntary. If there is change, we want it to be voluntary rather than imposed on people. I think that’s a very important point. That’s all good.”
1. This is interestingly super different and in conflict with the previous claim.
2. It’s fully the other way so far that I don’t even fully endorse it, this idea that change needs to be voluntary whenever it is imposed on people. That neither seems like a reasonable ask, nor does it historically end well, as in the paralysis of the West and especially the Anglosphere in many ways, especially in housing.
3. I am very confident in what would happen if you asked about the changes Sutton is anticipating, and put them to a vote.

Fundamentally, I didn’t pull direct quotes on this but Sutton repeatedly emphasizes that AI-dominated futures can be good or bad, that he wants us to steer towards good futures rather than bad futures, and that we should think carefully about which futures we are steering towards and choose deliberately.

I can certainly get behind that. The difference is that I don’t think we need to accept this transition to AI dominance as our only option, including that I don’t think we should accept that humans will always be unable to coordinate.

Mostly what I found interesting were the claims around the limitations and nature of LLMs, in ways that don’t make sense to me. This did help solidify a bunch of my thinking about how all of this works, so it felt like a good use of time for that alone.

Discussion about this post

On Dwarkesh Patel’s Podcast With Richard Sutton Read More »

zr1,-gtd,-and-america’s-new-nurburgring-war

ZR1, GTD, and America’s new Nürburgring war

car review, Cars, Chevrolet Corvette ZR1, Corvette ZR1X, Features, First drive, Ford Mustang GTD, Nordschleife, Nurburgring / Tim Belzer / September 29, 2025

Drive quickly and make a lot of horsepower.

Ford and Chevy set near-identical lap times with very different cars; we drove both.

Credit: Tim Stevens | Aurich Lawson

There’s a racetrack with a funny name in Germany that, in the eyes of many international enthusiasts, is the de facto benchmark for automotive performance. But the Nürburgring, a 13-mile (20 km) track often called the Green Hell, rarely hits the radar of mainstream US performance aficionados. That’s because American car companies rarely take the time to run cars there, and if they do, it’s in secrecy, to test pre-production machines cloaked in camouflage without publishing official times.

The track’s domestic profile has lately been on the rise, though. Late last year, Ford became the first American manufacturer to run a sub-7-minute lap: 6: 57.685 from its ultra-high-performance Mustang GTD. It then did better, announcing a 6: 52.072 lap time in May. Two months later, Chevrolet set a 6: 49.275 lap time with the hybrid Corvette ZR1X, becoming the new fastest American car around that track.

It’s a vehicular war of escalation, but it’s about much more than bragging rights.

The Green Hell as a must-visit for manufacturers

The Nürburgring is a delightfully twisted stretch of purpose-built asphalt and concrete strewn across the hills of western Germany. It dates back to the 1920s and has hosted the German Grand Prix for a half-century before it was finally deemed too unsafe in the late 1970s.

It’s still a motorsports mecca, with sports car racing events like the 24 Hours of the Nürburgring drawing hundreds of thousands of spectators, but today, it’s better known as the ultimate automotive performance proving ground.

It offers an unmatched variety of high-speed corners, elevation changes, and differing surfaces that challenge the best engineers in the world. “If you can develop a car that goes fast on the Nürburgring, it’s going to be fast everywhere in the whole world,” said Brian Wallace, the Corvette ZR1’s vehicle dynamics engineer and the driver who set that car’s fast lap of 6: 50.763.

“When you’re going after Nürburgring lap time, everything in the car has to be ten tenths,” said Greg Goodall, Ford’s chief program engineer for the Mustang GTD. “You can’t just use something that is OK or decent.”

Thankfully, neither of these cars is merely decent.

Mustang, deconstructed

You know the scene in Robocop where a schematic displays how little of Alex Murphy’s body remains inside that armor? Just enough of Peter Weller’s iconic jawline remains to identify the man, but the focus is clearly on the machine.

That’s a bit like how Multimatic creates the GTD, which retains just enough Mustang shape to look familiar, but little else.

Multimatic, which builds the wild Ford GT and also helms many of Ford’s motorsports efforts, starts with partially assembled Mustangs pulled from the assembly line, minus fenders, hood, and roof. Then the company guts what’s left in the middle.

Ford’s partner Multimatic cut as much of the existing road car chassis as it could for the GTD. Tim Stevens

“They cut out the second row seat area where our suspension is,” Ford’s Goodall said. “They cut out the rear floor in the trunk area because we put a flat plate on there to mount the transaxle to it. And then they cut the rear body side off and replace that with a wide-body carbon-fiber bit.”

A transaxle is simply a fun name for a rear-mounted transmission—in this case, an eight-speed dual-clutch unit mounted on the rear axle to help balance the car’s weight.

The GTD needs as much help as it can get to offset the heft of the 5.2-liter supercharged V8 up front. It gets a full set of carbon-fiber bodywork, too, but the resulting package still weighs over 4,300 lbs (1,950 kg).

With 815 hp (608 kW) and 664 lb-ft (900 Nm) of torque, it’s the most powerful road-going Mustang of all time, and it received other upgrades to match, including carbon-ceramic brake discs at the corners and the wing to end all wings slung off the back. It’s not only big; it’s smart, featuring a Formula One-style drag-reduction system.

At higher speeds, the wing’s element flips up, enabling a 202 mph (325 km/h) top speed. No surprise, that makes this the fastest factory Mustang ever. At a $325,000 starting price, it had better be, but when it comes to the maximum-velocity stakes, the Chevrolet is in another league.

More Corvette

You lose the frunk but gain cooling and downforce. Tim Stevens

On paper, when it comes to outright speed and value, the Chevrolet Corvette ZR1 seems to offer far more bang for what is still a significant number of bucks. To be specific, the ZR1 starts at about $175,000, which gets you a 1,064 hp (793 kW) car that will do 233 mph (375 km/h) if you point it down a road long enough.

Where the GTD is a thorough reimagining of what a Mustang can be, the ZR1 sticks closer to the Corvette script, offering more power, more aerodynamics, and more braking without any dramatic internal reconfiguration. That’s because it was all part of the car’s original mission plan, GM’s Brian Wallace told me.

“We knew we were going to build this car,” he said, “knowing it had the backbone to double the horsepower, put 20 percent more grip in the car, and oodles of aero.”

At the center of it all is a 5.5-liter twin-turbocharged V8. You can get a big wing here, too, but it isn’t active like the GTD’s.

Chevrolet engineers bolstered the internal structure at the back of the car to handle the extra downforce at the rear. Up front, the frunk is replaced by a duct through the hood, providing yet more grip to balance things. Big wheels, sticky tires, and carbon-ceramic brakes round out a package that looks a little less radical on the outside than the Mustang and substantially less retooled on the inside, but clearly no less capable.

The engine bay of a yellow Corvette ZR1. — A pair of turbochargers lurk behind that rear window. Credit: Tim Stevens

And if that’s not enough, Chevrolet has the 1,250 hp (932 kW), $208,000 ZR1X on offer, which adds the Corvette E-Ray’s hybrid system into the mix. That package does add more weight, but the result is still a roughly 4,000-lb (1,814 kg) car, hundreds less than the Ford.

’Ring battles

Ford and Chevy’s battle at the ‘ring blew up this summer, but both brands have tested there for years. Chevrolet has even set official lap times in the past, including the previous-generation Corvette Z06’s 7: 22.68 in 2012. Despite that, a fast lap time was not in the initial plan for the new ZR1 and ZR1X. Drew Cattell, ZR1X vehicle dynamics engineer and the driver of that 6: 49.275 lap, told me it “wasn’t an overriding priority” for the new Corvette.

But after developing the cars there so extensively, they decided to give it a go. “Seeing what the cars could do, it felt like the right time. That we had something we were proud of and we could really deliver with,” he said.

Ford, meanwhile, had never set an official lap time at the ‘ring, but it was part of the GTD’s raison d’être: “That was always a goal: to go under seven minutes. And some of it was to be the first American car ever to do it,” Ford’s Goodall said.

That required extracting every bit of performance, necessitating a last-minute change during final testing. In May of 2024, after the car’s design had been finalized by everyone up the chain of command at Ford, the test team in Germany determined the GTD needed a little more front grip.

To fix it, Steve Thompson, a dynamic technical specialist at Ford, designed a prototype aerodynamic extension to the vents in the hood. “It was 3D-printed, duct taped,” Goodall said. That design was refined and wound up on the production car, boosting frontal downforce on the GTD without adding drag.

Chevrolet’s development process relied not only on engineers in Germany but also on work in the US. “The team back home will keep on poring over the data while we go to sleep, because of the time difference,” Cattell said, “and then they’ll have something in our inbox the next morning to try out.”

When it was time for the Corvette’s record-setting runs, there wasn’t much left to change, just a few minor setup tweaks. “Maybe a millimeter or two,” Wallace said, “all within factory alignment settings.”

A few months later, it was my turn.

Behind the wheel

No, I wasn’t able to run either of these cars at the Nürburgring, but I was lucky enough to spend one day with both the GTD and the ZR1. First was the Corvette at one of America’s greatest racing venues: the Circuit of the Americas, a 3.5-mile track and host of the Formula One United States Grand Prix since 2012.

A head-on shot of a yellow Corvette ZR1. — How does 180 mph on the back straight at the Circuit of the Americas sound? Credit: Tim Stevens

I’ve been lucky to spend a lot of time in various Corvettes over the years, but none with performance like this. I was expecting a borderline terrifying experience, but I couldn’t have been more wrong. Despite its outrageous speed and acceleration, the ZR1 really is still a Corvette.

On just my second lap behind the wheel of the ZR1, I was doing 180 mph down the back straight and running a lap time close to the record set by a $1 million McLaren Senna a few years before. The Corvette is outrageously fast—and frankly exhausting to drive thanks to the monumental G forces—but it’s more encouraging than intimidating.

The GTD was more of a commitment. I sampled one at The Thermal Club near Palm Springs, California, a less auspicious but more technical track with tighter turns and closer walls separating them. That always amps up the pressure a bit, but the challenging layout of the track really forced me to focus on extracting the most out of the Mustang at low and high speeds.

The GTD has a few tricks up its sleeve to help with that, including an advanced multi-height suspension that drops it by about 1.5 inches (4 cm) at the touch of a button, optimizing the aerodynamic performance and lowering the roll height of the car.

A black Ford Mustang GTD in profile.

Heavier and less powerful than the Corvette, the Mustang GTD has astonishing levels of cornering grip. Credit: Tim Stevens

While road-going Mustangs typically focus on big power in a straight line, the GTD’s real skill is astonishing grip and handling. Remember, the GTD is only a few seconds slower on the ‘ring than the ZR1, despite weighing somewhere around 400 pounds (181 kg) more and having nearly 200 fewer hp (149 kw).

The biggest difference in feel between the two, though, is how they accelerate. The ZR1’s twin-turbocharged V8 delivers big power when you dip in the throttle and then just keeps piling on more and more as the revs increase. The supercharged V8 in the Mustang, on the other hand, is more like an instantaneous kick in the posterior. It’s ferocious.

Healthy competition

The ZR1 is brutally fast, yes, but it’s still remarkably composed, and it feels every bit as usable and refined as any of the other flavors of modern Corvette. The GTD, on the other hand, is a completely different breed than the base Mustang, every bit the purpose-built racer you’d expect from a race shop like Multimatic.

Chevrolet did the ZR1 and ZR1X development in-house. Cattell said that is a huge point of pride for the team. So, too, is setting those ZR1 and ZR1X lap times using General Motors’ development engineers. Ford turned to a pro race driver for its laps.

A racing driver stands in front his car as mechanics and engineers celebrate in the background. — Ford factory racing driver Dirk Muller was responsible for setting the GTD’s time at the ‘ring. Credit: Giles Jenkyn Photography LTD/Ford

An engineer in a fire suit stands next to a yellow Corvette, parked on the Nurburgring. — GM vehicle dynamics engineer Drew Cattell set the ZR1X’s Nordschleife time. Credit: Chevrolet

That, though, was as close to a barb as I could get out of any engineer on either side of this new Nürburgring. Both teams were extremely complimentary of each other.

“We’re pretty proud of that record. And I don’t say this in a snarky way, but we were first, and you can’t ever take away first,” Ford’s Goodall said. “Congratulations to them. We know better than anybody how hard of an accomplishment or how big of an accomplishment it is and how much effort goes into it.”

But he quickly added that Ford isn’t done. “You’re not a racer if you’re just going to take that lying down. So it took us approximately 30 seconds to align that we were ready to go back and do something about it,” he said.

In other words, this Nürburgring war is just beginning.

ZR1, GTD, and America’s new Nürburgring war Read More »

LG’s $1,800 TV for seniors makes misguided assumptions

lg, Tech, TVs / Tim Belzer / September 27, 2025

LG is looking to create a new market: TVs for senior citizens. However, I can’t help thinking that the answer for a TV that truly prioritizes the needs of older people is much simpler—and cheaper.

On Thursday, LG announced the Easy TV in South Korea, aiming it at the “senior TV market,” according to a Google translation of the press release. One of the features that LG has included in attempts to appeal to this demographic is a remote control with numbers. Many remotes for smart TVs, streaming sticks, and boxes don’t have numbered buttons, with much of the controller’s real estate dedicated to other inputs.

The Easy TV's remote. — The Easy TV’s remote. Credit: LG

LG released a new version of its Magic Remote in January with a particularly limited button selection that is likely to confuse or frustrate newcomers. In addition to not having keys for individual numbers, there are no buttons for switching inputs, play/pause, or fast forward/rewind.

LG AI remote — LG’s 2025 Magic Remote. Credit: Tom’s Guide/YouTube

The Easy TV’s remote has all of those buttons, plus mute, zoom, and bigger labels. The translated press release also highlights a button that sounds like “back” and says that seniors can push it to quickly return to the previous broadcast. The company framed it as a way for users to return to what they were watching after something unexpected occurs, such as an app launching accidentally or a screen going dark after another device is plugged into the TV.

You’ll also find the same sort of buttons that you typically find with new smart TV remotes these days, including buttons for launching specific streaming services.

Beyond the remote, LG tweaked its operating system for TVs, webOS, to focus on “five senior-focused features and favorite apps” and use a larger font, the translated announcement said.

Some Easy TV features are similar to those available on LG’s other TVs, but tailored to use cases that LG believes seniors are interested in. For instance, LG says seniors can use a reminder feature for medication alerts, set up integrated video calling features to quickly connect with family members who can assist with TV problems or an emergency, and play built-in games aimed at brain health.

LG’s $1,800 TV for seniors makes misguided assumptions Read More »

Ebola outbreak in DR Congo rages, with 61% death rate and funding running dry

Congo, ebola, health, USAID, WHO / Tim Belzer / September 26, 2025

Jeopardized efforts

This week, the IFRC requested $25 million to contain the outbreak, but it has only $2.2 million in emergency funds for its outbreak response so far. The WHO likewise estimated the cost of responding to the outbreak over the next three months to be $20 million. But WHO spokesperson Tarik Jasarevic told the AP on Thursday that it only had $4.3 million in funding to draw from—a $2 million emergency fund and $2.3 million in funding from the United Kingdom, Germany, and the Gavi vaccine alliance.

“Without immediate support, gaps in operations will persist, jeopardizing efforts to contain the outbreak and protect vulnerable communities,” Jasarevic said.

In the past, the US Agency for International Development, USAID, has provided critical support to respond to such outbreaks. But, with funding cuts and a dismantling of the agency by the Trump administration, the US is notably absent, and health officials fear it will be difficult to compensate for the loss.

Mathias Mossoko, the Ebola Response Coordinator in Bulape, told the AP that the US has provided “some small support” but declined to elaborate.

Amitié Bukidi, chief medical officer of the Mweka health zone—another health zone in the Kasai province—told the outlet that there was still much work to do to contain the outbreak. “The need is still very great,” he said. “If USAID were to be involved, that would be good.”

Ebola outbreak in DR Congo rages, with 61% death rate and funding running dry Read More »

50+ scientific societies sign letter objecting to Trump executive order

Grant funding, grant review process, Policy, Science, Science policy / Tim Belzer / September 26, 2025

Last month, the Trump administration issued an executive order asserting political control over grant funding, including all federally supported research. In general, the executive order inserts a layer of political control over both the announcement of new funding opportunities and the approval of individual grants. Now, a coalition of more than 50 scientific and medical organizations is firing back, issuing a letter to the US Congress expressing grave concerns over the order’s provisions and urging Congress to protect the integrity of what has long been an independent, merit-based, peer-review system for awarding federal grants.

As we previously reported, the order requires that any announcement of funding opportunities be reviewed by the head of the agency or someone they designate, which means a political appointee will have the ultimate say over what areas of science the US funds. Individual grants will also require clearance from a political appointee and “must, where applicable, demonstrably advance the President’s policy priorities.”

The order also instructs agencies to formalize the ability to cancel previously awarded grants at any time if they’re considered “no longer advance agency priorities.” Until a system is in place to enforce the new rules, agencies are forbidden from starting new funding programs.

In short, the new rules would mean that all federal science research would need to be approved by a political appointee who may have no expertise in the relevant areas, and the research can be canceled at any time if the political winds change. It would mark the end of a system that has enabled US scientific leadership for roughly 70 years.

50+ scientific societies sign letter objecting to Trump executive order Read More »

Rocket Report: Keeping up with Kuiper; New Glenn’s second flight slips

ariane 6, artemis II, astra, blue origin, Firefly Aerospace, moon, NASA, new glenn, rocket report, Security, Space, space launch system, spacex, starship, VIPER / Tim Belzer / September 26, 2025

Amazon plans to conduct two launches of Kuiper broadband satellites just days apart.

An unarmed Trident II D5 Life Extension (D5LE) missile launches from an Ohio-class ballistic missile submarine off the coast of Florida. Credit: US Navy

Welcome to Edition 8.12 of the Rocket Report! We often hear from satellite operators—from the military to venture-backed startups—about their appetite for more launch capacity. With so many rocket launches happening around the world, some might want to dismiss these statements as a corporate plea for more competition, and therefore lower prices. SpaceX is on pace to launch more than 150 times this year. China could end the year with more than 70 orbital launches. These are staggering numbers compared to global launch rates just a few years ago. But I’m convinced there’s room for more alternatives for reliable (and reusable) rockets. All of the world’s planned mega-constellations will need immense launch capacity just to get off the ground, and if successful, they’ll go into regular replacement and replenishment cycles. Throw in the still-undefined Golden Dome missile shield and many nations’ desire for a sovereign launch capability, and it’s easy to see the demand curve going up.

As always, we welcome reader submissions. If you don’t want to miss an issue, please subscribe using the box below (the form will not appear on AMP-enabled versions of the site). Each report will include information on small-, medium-, and heavy-lift rockets, as well as a quick look ahead at the next three launches on the calendar.

Sharp words from Astra’s Chris Kemp. Chris Kemp, the chief executive officer of Astra, apparently didn’t get the memo about playing nice with his competitors in the launch business. Kemp made some spicy remarks at the Berkeley Space Symposium 2025 earlier this month, billed as the largest undergraduate aerospace event at the university (see video of the talk). During the speech, Kemp periodically deviated from building up Astra to hurling insults at several of his competitors in the launch industry, Ars reports. To be fair to Kemp, some of his criticisms are not without a kernel of truth. But they are uncharacteristically rough all the same, especially given Astra’s uneven-at-best launch record and financial solvency to date.

Wait, what?! … Kemp is generally laudatory in his comments about SpaceX, but his most crass statement took aim at the quality of life of SpaceX employees at Starbase, Texas. He said life at Astra is “more fun than SpaceX because we’re not on the border of Mexico where they’ll chop your head off if you accidentally take a left turn.” For the record, no SpaceX employees have been beheaded. “And you don’t have to live in a trailer. And we don’t make you work six and a half days a week, 12 hours a day.” Kemp also accused Firefly Aerospace of sending Astra “garbage” rocket engines as part of the companies’ partnership on propulsion for Astra’s next-generation rocket.

The easiest way to keep up with Eric Berger’s and Stephen Clark’s reporting on all things space is to sign up for our newsletter. We’ll collect their stories and deliver them straight to your inbox.

Sign Me Up!

A step forward for Europe’s reusable rocket program. No one could accuse the European Space Agency and its various contractors of moving swiftly when it comes to the development of reusable rockets. However, it appears that Europe is finally making some credible progress, Ars reports. Last week, the France-based ArianeGroup aerospace company announced that it completed the integration of the Themis vehicle, a prototype rocket that will test various landing technologies, on a launch pad in Sweden. Low-altitude hop tests, a precursor for developing a rocket’s first stage that can vertically land after an orbital launch, could start late this year or early next.

Hopping into the future … “This milestone marks the beginning of the ‘combined tests,’ during which the interface between Themis and the launch pad’s mechanical, electrical, and fluid systems will be thoroughly trialed, with the aim of completing a test under cryogenic conditions,” ArianeGroup said. This particular rocket will likely undergo only short hops, initially about 100 meters. A follow-up vehicle, Themis T1E, is intended to fly medium-altitude tests at a later date. Some of the learnings from these prototypes will feed into a smaller, reusable rocket intended to lift 500 kilograms to low-Earth orbit. This is under development by MaiaSpace, a subsidiary of ArianeGroup. Eventually, the European Space Agency would like to use technology developed as part of Themis to develop a new line of reusable rockets that will succeed the Ariane 6 rocket.

Navy conducts Trident missile drills. The US Navy carried out four scheduled missile tests of a nuclear-capable weapons system off the coast of Florida within the last week, Defense News reports. The service’s Strategic Systems Programs conducted flights of unarmed Trident II D5 Life Extension missiles from a submerged Ohio-class ballistic missile submarine from September 17 to September 21 as part of an ongoing scheduled event meant to test the reliability of the system. “The missile tests were not conducted in response to any ongoing world events,” a Navy release said.

Secret with high visibility … The Navy periodically performs these Trident missile tests off the coasts of Florida and California, taking advantage of support infrastructure and range support from the two busiest US spaceports. The military doesn’t announce the exact timing of the tests, but warnings issued for pilots to stay out of the area give a general idea of when they might occur. One of the launch events Sunday was visible from Puerto Rico, illuminating the night sky in photos published on social media. The missiles fell in the Atlantic Ocean as intended, the Navy said. The Trident II D5 missiles were developed in the 1980s and are expected to remain in service on the Navy’s ballistic missile submarines into the 2040s. The Trident system is one leg of the US military’s nuclear triad, alongside land-based Minuteman ballistic missiles and nuclear-capable strategic bombers. (submitted by EllPeaTea)

Firefly plans for Alpha’s return to flight. Firefly Aerospace expects to resume Alpha launches in the “coming weeks,” with two flights planned before the end of the year, Space News reports. These will be the first flights of Firefly’s one-ton-class Alpha rocket since a failure in April destroyed a Lockheed Martin tech demo satellite after liftoff from California. In a quarterly earnings call, Firefly shared a photo showing its next two Alpha rockets awaiting shipment from the company’s Texas factory.

Righting the ship … These next two launches really need to go well for Firefly. The Alpha rocket has, at best, a mixed record with only two fully successful flights in six attempts. Two other missions put their payloads into off-target orbits, and two Alpha launches failed to reach orbit at all. Firefly went public on the NASDAQ stock exchange last month, raising nearly $900 million in the initial public offering to help fund the company’s future programs, namely the medium-lift Eclipse rocket developed in partnership with Northrop Grumman. There’s a lot to like about Firefly. The company achieved the first fully successful landing of a commercial spacecraft on the Moon in March. NASA has selected Firefly for three more commercial landings on the Moon, and Firefly reported this week it has an agreement with an unnamed commercial customer for an additional dedicated mission. But the Alpha program hasn’t had the same level of success. We’ll see if Firefly can get the rocket on track soon. (submitted by EllPeaTea)

Avio wins contract to launch “extra-European” mission. Italian rocket builder Avio has signed a launch services agreement with US-based launch aggregator SpaceLaunch for a Vega C launch carrying an Earth observation satellite for an “extra-European institutional customer” in 2027, European Spaceflight reports. Avio announced that it had secured the launch contract on September 18. According to the company, the contract was awarded through an open international competition, with Vega C chosen for its “versatility and cost-effectiveness.” While Avio did not reveal the identity of the “extra-European” customer, it said that it would do so later this year.

Plenty of peculiarities … There are several questions to unpack here, and Andrew Parsonson of European Spaceflight goes through them all. Presumably, extra-European means the customer is based outside of Europe. Avio’s statement suggests we’ll find out the answer to that question soon. Details about the US-based launch broker SpaceLaunch are harder to find. SpaceLaunch appears to have been founded in January 2025 by two former Firefly Aerospace employees with a combined 40 years of experience in the industry. On its website, the company claims to provide end-to-end satellite launch integration, mission management, and launch procurement services with a “portfolio of launch vehicle capacity around the globe.” SpaceLaunch boasts it has supported the launch of more than 150 satellites on 12 different launch vehicles. However, according to public records, it does not appear that the company itself has supported a single launch. Instead, the claim seems to credit SpaceLaunch with launches that were actually carried out during the two founders’ previous tenures at Spaceflight, Firefly Aerospace, Northrop Grumman, and the US Air Force. (submitted by EllPeaTea)

Falcon 9 launches three missions for NASA and NOAA. Scientists loaded three missions worth nearly $1.6 billion on a SpaceX Falcon 9 rocket for launch Wednesday, toward an orbit nearly a million miles from Earth, to measure the supersonic stream of charged particles emanating from the Sun, Ars reports. One of the missions, from the National Oceanic and Atmospheric Administration (NOAA), will beam back real-time observations of the solar wind to provide advance warning of geomagnetic storms that could affect power grids, radio communications, GPS navigation, air travel, and satellite operations. The other two missions come from NASA, with research objectives that include studying the boundary between the Solar System and interstellar space and observing the rarely seen outermost layer of our own planet’s atmosphere.

Immense value … All three spacecraft will operate in orbit around the L1 Lagrange point, a gravitational balance point located more than 900,000 miles (1.5 million kilometers) from Earth. Bundling these three missions onto the same rocket saved at least tens of millions of dollars in launch costs. Normally, they would have needed three different rockets. Rideshare missions to low-Earth orbit are becoming more common, but spacecraft departing for more distant destinations like the L1 Lagrange point are rare. Getting all three missions on the same launch required extensive planning, a stroke of luck, and fortuitous timing. “This is the ultimate cosmic carpool,” said Joe Westlake, director of NASA’s heliophysics division. “These three missions heading out to the Sun-Earth L1 point riding along together provide immense value for the American taxpayer.”

US officials concerned about China mastering reusable launch. SpaceX’s dominance in reusable rocketry is one of the most important advantages the United States has over China as competition between the two nations extends into space, US Space Force officials said Monday. But several Chinese companies are getting close to fielding their own reusable rockets, Ars reports. “It’s concerning how fast they’re going,” said Brig. Gen. Brian Sidari, the Space Force’s deputy chief of space operations for intelligence. “I’m concerned about when the Chinese figure out how to do reusable lift that allows them to put more capability on orbit at a quicker cadence than currently exists.”

By the numbers … China has used 14 different types of rockets on its 56 orbital-class missions this year, and none have flown more than 11 times. Eight US rocket types have cumulatively flown 145 times, with 122 of those using SpaceX’s workhorse Falcon 9. Without a reusable rocket, China must maintain more rocket companies to sustain a launch rate of just one-third to one-half that of the United States. This contrasts with the situation just four years ago, when China outpaced the United States in orbital rocket launches. The growth in US launches has been a direct result of SpaceX’s improvements to launch at a higher rate, an achievement primarily driven by the recovery and reuse of Falcon 9 boosters and payload fairings.

Atlas V launches more Kuiper satellites. Roughly an hour past sunrise on Thursday, an Atlas V rocket from United Launch Alliance took flight from Cape Canaveral Space Force Station, Florida. Onboard the rocket, flying in its most powerful configuration, were the next 27 Project Kuiper broadband satellites from Amazon, Spaceflight Now reports. This is the third batch of production satellites launched by ULA and the fifth overall for the growing low-Earth orbit constellation. The Atlas V rocket released the 27 Kuiper satellites about 280 miles (450 kilometers) above Earth. The satellites will use onboard propulsion to boost themselves to their assigned orbit at 392 miles (630 kilometers).

Another Kuiper launch on tap … With this deployment, Amazon now has 129 satellites in orbit. This is a small fraction of the network’s planned total of 3,232 satellites, but Amazon has enjoyed a steep ramp-up in the Kuiper launch cadence as the company’s satellite assembly line in Kirkland, Washington, continues churning out spacecraft. Another 24 Kuiper satellites are slated to launch September 30 on a SpaceX Falcon 9 rocket, and Amazon has delivered enough satellites to Florida for an additional launch later this fall. (submitted by EllPeaTea)

German military will fly with Ariane 6. Airbus Defense and Space has awarded Arianespace a contract to launch a pair of SATCOMBw-3 communications satellites for the German Armed Forces, European Spaceflight reports. Airbus is the prime contractor for the nearly $2.5 billion (2.1 billion euro) SATCOMBw-3 program, which will take over from the two-satellite SATCOMBw-2 constellation currently providing secure communications for the German military. Arianespace announced Wednesday that it had been awarded the contract to launch the satellites aboard two Ariane 6 rockets. “By signing this new strategic contract for the German Armed Forces, Arianespace accomplishes its core mission of guaranteeing autonomous access to space for European sovereign satellites,” said Arianespace CEO David Cavaillolès.

Running home to Europe … The chief goal of the Ariane 6 program is to provide Europe with independent access to space, something many European governments see as a strategic requirement. Several European military, national security, and scientific satellites have launched on SpaceX Falcon 9 rockets in the last few years as officials waited for the debut of the Ariane 6 rocket. With three successful Ariane 6 flights now in the books, European customers seem to now have the confidence to commit to flying their satellites on Ariane 6. (submitted by EllPeaTea)

Artemis II launch targeted for February. NASA is pressing ahead with preparations for the first launch of humans beyond low-Earth orbit in more than five decades, and officials said Tuesday that the Artemis II mission could take flight early next year, Ars reports. Although work remains to be done, the space agency is now pushing toward a launch window that opens on February 5, 2026, officials said during a news conference on Tuesday at Johnson Space Center. The Artemis II mission represents a major step forward for NASA and seeks to send four astronauts—Reid Wiseman, Victor Glover, Christina Koch, and Jeremy Hansen—around the Moon and back. The 10-day mission will be the first time astronauts have left low-Earth orbit since the Apollo 17 mission in December 1972.

Orion named Integrity … The first astronauts set to fly to the Moon in more than 50 years will do so in Integrity, Ars reports. NASA’s Artemis II crew revealed Integrity as the name of their Orion spacecraft during a news conference on Wednesday at the Johnson Space Center in Houston. “We thought, as a crew, we need to name this spacecraft. We need to have a name for the Orion spacecraft that we’re going to ride this magical mission on,” said Wiseman, commander of the Artemis II mission.

FAA reveals new Starship trajectories. Sometime soon, perhaps next year, SpaceX will attempt to fly one of its enormous Starship rockets from low-Earth orbit back to its launch pad in South Texas. A successful return and catch at the launch tower would demonstrate a key capability underpinning Elon Musk’s hopes for a fully reusable rocket. For this to happen, SpaceX must overcome the tyranny of geography. A new document released by the Federal Aviation Administration shows the narrow corridors Starship will fly to space and back when SpaceX tries to recover them, Ars reports.

Flying over people … It was always evident that flying a Starship from low-Earth orbit back to Starbase would require the rocket to fly over Mexico and portions of South Texas. The rocket launches to the east over the Gulf of Mexico, so it must approach Starbase from the west when it comes in for a landing. The new maps show SpaceX will launch Starships to the southeast over the Gulf and the Caribbean Sea, and directly over Jamaica, or to the northeast over the Gulf and the Florida peninsula. On reentry, the ship will fly over Baja California and Mexico’s interior near the cities of Hermosillo and Chihuahua, each with a population of roughly a million people. The trajectory would bring Starship well north of the Monterrey metro area and its 5.3 million residents, then over the Rio Grande Valley near the Texas cities of McAllen and Brownsville.

New Glenn’s second flight at least a month away. The second launch of Blue Origin’s New Glenn rocket, carrying a NASA smallsat mission to Mars, is now expected in late October or early November, Space News reports. Tim Dunn, NASA’s senior launch director at Kennedy Space Center, provided an updated schedule for the second flight of New Glenn in comments after a NASA-sponsored launch on a Falcon 9 rocket on Wednesday. Previously, the official schedule from NASA showed the launch date as no earlier than September 29.

No surprise … It was already apparent that this launch wouldn’t happen on September 29. Blue Origin has test-fired the second stage for the upcoming flight of the New Glenn rocket but hasn’t rolled the first stage to the launch pad for its static fire. Seeing the rocket emerge from Blue’s factory in Florida will be an indication that the launch date is finally near. Blue Origin will launch NASA’s ESCAPADE mission, a pair of small satellites to study how the solar wind interacts with the Martian upper atmosphere.

Blue Origin will launch a NASA rover to the Moon. NASA has awarded Blue Origin a task order worth up to $190 million to deliver its Volatiles Investigating Polar Exploration Rover (VIPER) to the Moon’s surface, Aviation Week & Space Technology reports. Blue Origin, one of 13 currently active Commercial Lunar Payload Services (CLPS) providers, submitted the only bid to carry VIPER to the Moon after NASA requested offers from industry last month. NASA canceled the VIPER mission last year, citing cost overruns with the rover and delays in its planned ride to the Moon aboard a lander provided by Astrobotic. But engineers had already completed assembly of the rover, and scientists protested NASA’s decision to terminate the mission.

Some caveats … Blue Origin will deliver VIPER to a location near the Moon’s south pole in late 2027 using a robotic Blue Moon MK1 lander, a massive craft larger than the Apollo lunar landing module. The company’s first Blue Moon MK1 lander is scheduled to fly to the Moon next year. NASA’s contract for the VIPER delivery calls for Blue Origin to design accommodations for the rover on the Blue Moon lander. The agency said it will decide whether to proceed with the actual launch on a New Glenn rocket and delivery of VIPER to the Moon based partially on the outcome of the first Blue Moon test flight next year.

Next three launches

Sept. 26: Long March 4C | Unknown Payload | Jiuquan Satellite Launch Center, China | 19: 20 UTC

Sept. 27: Long March 6A | Unknown Payload | Taiyuan Satellite Launch Center, China | 12: 39 UTC

Sept. 28: Falcon 9 | Starlink 11-20 | Vandenberg Space Force Base, California | 23: 32 UTC

Stephen Clark is a space reporter at Ars Technica, covering private space companies and the world’s space agencies. Stephen writes about the nexus of technology, science, policy, and business on and off the planet.

Rocket Report: Keeping up with Kuiper; New Glenn’s second flight slips Read More »

Fiji’s ants might be the canary in the coal mine for the insect apocalypse

ants, Biology, evolution, Fiji, insects, Science, speciation / Tim Belzer / September 26, 2025

A new genetic technique lets museum samples track population dynamics.

In late 2017, a study by Krefeld Entomological Society looked at protected areas across Germany and discovered that two-thirds of the insect populations living in there had vanished over the last 25 years. The results spurred the media to declare we’re living through an “insect apocalypse,” but the reasons behind their absence were unclear. Now, a joint team of Japanese and Australian scientists have completed a new, multi-year study designed to get us some answers.

Insect microcosm

“In our work, we focused on ants because we have systematic ways for collecting them,” says Alexander Mikheyev, an evolutionary biologist at the Australian National University. “They are also a group with the right level of diversity, where you have enough species to do comparative studies.” Choosing the right location, he explained, was just as important. “We did it in Fiji, because Fiji had the right balance between isolation—which gave us a discrete group of animals to study—but at the same time was diverse enough to make comparisons,” Mikheyev adds.

Thus, the Fijian archipelago, with its 330 islands, became the model the team used to get some insights into insect population dynamics. A key difference from the earlier study was that Mikheyev and his colleagues could look at those populations across thousands of years, not just the last 25.

“Most of the previous studies looked at actual observational data—things we could come in and measure,” Mikheyev explains. The issue with those studies was that they could only account for the last hundred years or so, because that’s how long we have been systematically collecting insect samples. “We really wanted to understand what happened in the longer time frame,” Mikheyev says.

To do this, his team focused on community genomics—studying the collective genetic material of entire groups of organisms. The challenge is that this would normally require collecting thousands of ants belonging to hundreds of species across the entire Fijian archipelago. Given that only a little over 100 out of 330 islands in Fiji are permanently inhabited, this seemed like an insurmountable challenge.

To go around it, the team figured they could run its tests on ants already collected in Fijian museums. But that came with its own set of difficulties.

DNA pieces

Unfortunately, the quality of DNA that could be obtained from museum collections was really bad. From the perspective of DNA preservation, the ants were obtained and stored in horrific conditions, since the idea was to showcase them for visitors, not run genetic studies. “People were catching them in malaise traps,” Mikheyev says. “A malaise trap is basically a bottle of alcohol that sits somewhere in Fiji for a month. Those samples had horribly fragmented, degraded DNA.”

To work with this degraded genetic material, the team employed a technique they called high-throughput museumomics, a relatively new technique that looks at genetic differences across a genome without sequencing the whole thing. DNA sampled from multiple individuals was cut and marked with unique tags at the same repeated locations, a bit like using bookmarks to pinpoint the same page or passage in different issues of the same book. Then, the team sequenced short DNA fragments following the tag to look for differences between them, allowing them to evaluate the genetic diversity within a population. “We developed a series of methods that actually allowed us to harness these museum-grade specimens for population genetics,” Mikheyev explains.

But the trouble didn’t end there. Differences among Fijian ant taxa are based on their appearance, not genetic analysis. For years, researchers were collecting various ants and determining their species by looking at them. This led to 144 species belonging to 40 genera. For Mikheyev’s team, the first step was to look at the genomes in the samples and see if these species divisions were right. It turned out that they were mostly correct, but some species had to be split, while others were lumped together. At the end, the team confirmed that 127 species were represented among their samples.

Overall, the team analyzed more than 4,000 specimens of ants collected over the past decade or so. And gradually, a turbulent history of Fijian ants started to emerge from the data.

The first colonists

The art of reconstructing the history of entire populations from individual genetic sequences relies on comparing them to each other thoroughly and running a whole lot of computer simulations. “We had multiple individuals per population,” Mikheyev explains. “Let’s say we look at this population and find it has essentially no diversity. It suggests that it very recently descended from a small number of individuals.” When the contrary was true and the diversity was high, the team assumed it indicated the population had been stable for a long time.

With the DNA data in hand, the team simulated how populations of ants would evolve over thousands of years under various conditions, and picked scenarios that best matched the genetic diversity results it obtained from real ants. “We identified multiple instances of colonization—broadscale evolutionary events that gave rise to the Fijian fauna that happened in different timeframes,” Mikheyev says. There was a total of at least 65 colonization events.

The first ants, according to Mikheyev, arrived at Fiji millions of years ago and gave rise to 88 endemic Fijian ant species we have today. These ants most likely evolved from a single ancestor and then diverged from their mainland relatives. Then, a further 23 colonization events introduced ants that were native to a broader Pacific region. These ants, the team found, were a mixture of species that colonized Fiji naturally and ones that were brought by the first human settlers, the Lapita people, who arrived around 3,000 years ago.

The arrival of humans also matched the first declines in endemic Fijian ant species.

Slash and burn

“In retrospect, these declines are not really surprising,” Mikheyev says. The first Fijian human colonists didn’t have the same population density as we have now, but they did practice things like slash-and-burn agriculture, where forests were cut down, left to dry, and burned to make space for farms and fertilize the soil. “And you know, not every ant likes to live in a field, especially the ones that evolved to live in a forest,” Mikheyev adds. But the declines in Fijian endemic ant species really accelerated after the first contact with the Europeans.

The first explorers in the 17th and 18th centuries, like Abel Tasman and James Cook, charted some of the Fijian islands but did not land there. The real apocalypse for Fijian ants began in the 19th century, when European sandalwood traders started visiting the archipelago on a regular basis and ultimately connected it to the global trade networks.

Besides the firearms they often traded for sandalwood with local chiefs, the traders also brought fire ants. “Fire ants are native to Latin America, and it’s a common invasive species extremely well adapted to habitats we create: lawns or clear-cut fields,” Mikheyev says. Over the past couple of centuries, his team saw a massive increase in fire ant populations, combined with accelerating declines in 79 percent of endemic Fijian ant species.

Signs of apocalypse

To Mikheyev, Fiji was just a proving ground to test the methods of working with museum-grade samples. “Now we know this approach works and we can start leveraging collections found in museums around the world—all of them can tell us stories about places where they were collected,” Mikheyev says. His ultimate goal is to look for the signs of the insect apocalypse, or any other apocalypse of a similar kind, worldwide.

But the question is whether what’s happening is really that bad? After all, not all ants seem to be in decline. Perhaps what we see is just a case of a better-adapted species taking over—natural selection happening before our eyes?

“Sure, we can just live with fire ants all along without worrying about the kind of beautiful biodiversity that evolution has created on Fiji,” Mikheyev says. “But I feel like if we just go with that philosophy, we’re really going to be irreparably losing important and interesting parts of our ecology.” If the current trends persist, he argues, we might lose endemic Fijian ants forever. “And this would make our world worse, in many ways,” Mikheyev says.

Science, 2025. DOI: 10.1126/science.ads3004

Jacek Krywko is a freelance science and technology writer who covers space exploration, artificial intelligence research, computer science, and all sorts of engineering wizardry.

Fiji’s ants might be the canary in the coal mine for the insect apocalypse Read More »

Economics Roundup #6

Economics / Tim Belzer / September 26, 2025

I obviously cover many economical things in the ordinary course of business, but I try to reserve the sufficiently out of place or in the weeds stuff that is not time sensitive for updates like this one.

We love trade now, so maybe it’ll all turn out great?

John Burn-Murdoch: Negative partisanship is a helluva drug:

Up until a few months ago, liberal and conservative Americans held pretty much the same views on free trade.

Now, not so much…

Yet another explanation that says ‘the China shock killed particular jobs but had large diffuse benefits and left us all much better off.’

In other trade is good news, Argentina under the crazy libertarian Melei is now growing at 5.8%, faster than China.

Alex Recouso: The recent capital gains tax increase in the UK was expected to bring additional tax revenue.

Instead, high-net-worth individuals and families are leaving the country leading to an 18% fall in net capital gains tax revenue. A £2.7b loss.

Welcome to the Laffer curve, suckers.

Here’s what looks at first like a wild paper, claiming that surge pricing is great overall and fantastic for riders increasing their surplus by 3.57%, but that it decreases driver surplus by 0.98% and the platform’s current profits by 0.5% of gross revenue.

At first that made no sense, obviously raising prices will be good for drivers, and Uber wouldn’t do it if it lowered revenue.

This result only makes sense once you realize that the paper is not holding non-surge pricing constant. It assumes without surge pricing, Uber would raise their baseline rates substantially. That’s also why this is bad for workers with long hours at off-peak times, as their revenue declines. Uber could raise more revenue now with higher off-peak hours, but it prefers to focus on the long term, which helps riders and hurts drivers.

That makes sense, but it also raises the question of why Uber is keeping prices so low at this point. Early on, sure, you’re growing the market and fighting for market share. But now, the market is mature, and has settled into a duopoly. Is Uber that afraid of competition? Is it simply corporate culture and inertia? I mean, Uber, never change (in this particular way), but it doesn’t seem optimal from your perspective.

Story mostly checks out in theory, as the practice is commonly used, with some notes. If tips are 90%+ a function of how much you tip in general, and vary almost none based on service, at equilibrium they’re mostly a tax on tippers paid to non-tippers.

Gabriel: it’s actually hilarious how tipping is just redistribution of capital from people pleasers to non people pleasers

tipping never increases salaries in the long run because free markets, so our entire tip becomes savings of the rude people that don’t tip ironically.

say waiters started earning 30% more from tips, then everyone wants to become a waiter, and now businesses can decrease salaries to match market value. more restaurants will be started from tipping, not an increase in salary

tipping would have served a great purpose if it was socially acceptable to not tip, cause people doing a great job and who are nice would be paid more than the not nice people

i always tip cause i feel bad if i wouldn’t, but in theory your tip makes no difference and markets would adjust accordingly in the long run (but slightly inefficiently since you’d be spending less until menu prices actually increase)

Nitish Panesar: It’s wild how “voluntary” generosity just ends up subsidizing the less generous Makes you wonder if the system is rewarding the exact opposite of what we hope.

It’s not only a distribution from pleasers to non-pleasers, it is also from truth tellers to liars, because being able to say that you tip, and tip generously, is a lot of what people are really buying via their tips.

There are some additional effects.

The menu price illusion effect raises willingness to pay, people are silly.
Tax policy (e.g. ‘no tax on tips’ or not enforcing one) can be impacted.
Tips can create an effective floor on server pay if a place has some mix of high prices and general good tippers, assuming the tips don’t get confiscated. That floor could easily bind.
If you tip generously as a regular, some places do give you better service in various ways that can be worthwhile. And if you tip zero or very low systematically enough that people remember this, there will be a response in the other direction.
Note that the network of at least the Resy-level high end restaurants keep and share notes on their customers. So if you tip well or poorly at that level, word will likely get out, even if you go to different places.

It is also likely that you can effectively do acausal trade a la Parfit’s Hitchhiker (Parfit’s Diner?) as I bet waiters are mostly pretty good at figuring out who is and is not going to tip well.

The other factor is, even if tipping as currently implemented doesn’t work in theory, it seems suspiciously like it works in practice – in the sense that American service in restaurants in particular is in general vastly better than in other places. It’s not obvious whether that would have been true anyway, but if it works, even if it in theory shouldn’t? Then it is totally worth all the weird effects.

The tax benefits of rewarding people via orgies, as a form of non-wage compensation. That is on top of the other benefits, as certain things cannot be purchased, purchased on behalf of others or even asked for without degrading their value. The perfect gift!

Even though we are taking longer retirements, the sheer amount of work being done is staying roughly flat over time:

Noah Smith explains that the divergence between mean productivity and median wages is mostly about unequal compensation, and when you adjust the classic divergence chart in several ways, the gap between wages and productivity declines dramatically, although not entirely:

That also says that real median wages are up 16.3% over that period. One can also consider that this is all part of the story of being forced to purchase more and better goods and services, without getting the additional funds to do that. As I’ve said, I do think that life as the median worker did get net harder over that period, but things are a lot less crazy than people think.

Matt Bruenig attempts a definitive response to the controversy around ‘how many Americans live paycheck to paycheck?’ Man, it’s weird.

For starters, what the hell does that actually mean? When advocates cite it, they clearly mean that such people are ‘one expense away from disaster’ but the studies they cite mostly mean something else, which is a low savings rate.

It seems all the major surveys people cite are actually asking about low savings rate.

The go-to LendingClub claim that it’s 60% of Americans asks if someone’s savings rate is zero or less. We also have BankRate at 34%, asking essentially the same question, that’s a huge gap. Bank of America got 50% to either agree or strongly agree.

Bank of America also looked at people’s expenses in detail to see who used 95% or more of household income on ‘necessity spending’ and got 26% but that whole process seems super weird and wonky, I wouldn’t rely on it for anything.

Whereas the median American net worth is $192,700 with a liquid net worth of $7,850. Which after adjustments is arguably only a month of income. But 54% of Americans, when asked in a survey, said they had 3 months of savings available. But then 24% of people who said that then said they ‘couldn’t afford’ an emergency expense of $2k, what? So it’s weird. $8k is still very different from the ‘most Americans can’t pay a few thousand in expenses’ claims we often see, and this is before you start using credit cards or other forms of credit.

So, yeah, it’s all very weird. People aren’t answering in ways that are logically consistent or that make much sense.

What is clear is that when people cite how many Americans are ‘living paycheck to paycheck,’ almost always they are presenting a highly false impression. The map you would take away from that claim does not match the territory.

In addition to the ‘half of Americans live paycheck to paycheck’ claims being well-known to be objectively false, there’s another mystery behind them that Matt Yglesias points out. What policies would prevent this from being true? The more of a social safety net you provide, the less costly it is to live ‘paycheck to paycheck’ and the more people will do it. If you want people to have savings that aren’t in effect provided by their safety nets, then take away their safety net, watch what happens.

Unrealized capital gains create strange incentives. How much do the rich practice ‘buy, borrow, die’? The answer is some, but not all that much, with such borrowing being on average only 1-2% of economic income. Mostly they ‘buy, save, die,’ as their liquid incomes usually exceed consumption. The loophole should still be closed, especially as other changes could cause it to matter far more, but what matters is the cost basis step-up on death. There should at minimum be a cap on that.

Capitalists wisely have a norm that you are very not allowed to lie about revenue, to a far greater extent than you are not allowed to lie about other things. Revenue is a sacred trust. If you are counting secondary effects you absolutely cannot say the amount is ‘in revenue.’

Patrick McKenzie goes insanely deep on the seemingly unbelievable story about a woman withdrawing $50k in cash at the bank with very little questioning, and then losing it to a scammer. After an absurd amount of investigation, including tracking down The Room Where It Happened, it turned out the story all made sense. The bank let her withdraw the money because, if you take into account the equity in her house, she was actually wealthy, and the bank knew that so $50k did not raise so many alarm bells.

Patrick McKenzie pushes back against the idea that interchange fees and credit card rewards, taken together, are regressive redistribution. I find his argument convincing.

RIP to the US Bank Smartly Card that offered unlimited 4% cash back. They have no idea how there could have been such a big adverse selection problem.

Refund bonuses on crowdsourced projects give good incentives and signals all around, so they are highly effective at attracting more funding for privately created public goods. The problem is that if you do fail to fund or fail to deliver, and need to give out the bonus, that is a rather big disaster. So you win big on one level, and lose big on another. If I’m actively poorer when I fail, I’m going to have a high threshold for starting. Good selection, but you lose a lot of projects, some of which you want.

Federal Reserve Board removes ‘reputational risk’ component from bank exams, replacing them with more specific discussions of financial risk. This is plausibly a good change but there is a core function here we want to be preserving.

When you hire a real estate agent to help you buy a house, your agent’s direct financial incentive is to make you pay a higher price, not a lower one.

Darth Powell: LMFAO.

Austen Allred: I just realized the buyer’s agent in a real estate transaction has no incentive to help their customers get a better price

“But the agent isn’t going to rip you off, if you save $50k they only lose $1500.”

They’re just not going to scratch and claw to save every penny the same way I would if it were my money. Example: Very easy to say, “It’s competitive you should just put an offer in at asking.”

“But they want referrals.” Sure, but they have all of the information about what is reasonable and not.

“But they have a code of ethics.” Lol. Lmfao.

Given the math it’s very obvious that it’s a volume game. A real estate agent gets paid not by saving or making money, but by pushing deals through. It’s very clear incentive misalignment.

How does a buyer know if a deal is competitive, or if they need to move quickly? If the buying agent tells them it is. How does the buying agent know if a deal is competitive? If the selling agent tells them so. I’m sure that’s never been abused ever.

Steven Sinofsky: Anyone who has ever bought or sold a house comes to realize you’re mostly negotiating with your own agent, while your agent is mostly colluding with the other agent against both buyer and seller offering back incomplete information—an entirely unaligned transaction experience.

Sam Pullara: it is extremely difficult to align agents

Austen Allred: OpenAI has a team for that.

Astrus Dastan: And they’re failing miserably.

Yep, if you get 3% of the sale price you’re not going to fight for the lowest (or even highest, for the seller, if it means risking or dragging out the sale) possible price on that basis. It’s a volume business. But also it’s a reputation and networking business.

What this is missing is that I got my real estate agent because my best friend referred her to me, and then she actively warned us off of buying a few places until we found the right one and walked me through all the steps, and then because she did a great job fighting for me I recommended her to other clients and she got two additional commissions. The incentives are there.

As for the quoted transaction, well, yes, they were forced to leave money on the table because their agent negotiated with a dishonest and risky strategy, got their bluff called and it backfired. Tough. I’m guessing there won’t be a recommendation there.

People also think businesses try to follow the Ferengi Rules of Acquisition, basically?

Rashida Tlaib: Families are struggling to put food on the table. I sent a letter to @Kroger about their decision to roll out surge pricing using facial recognition technology. Facial recognition technology is often discriminatory and shouldn’t be used in grocery stores to price gouge residents.

Leon: >walk into kroger with diarrhea

>try to make my face seem normal

>grab $5 diarrhea medicine

>sharp pain hits my gut

>diarrhea medicine now costs $100

Something tells me that if they do that, then next time, if you have any choice whatsoever (and usually you do have such a choice) you’re not going to Kroger, and that’s the best case scenario for your reaction here. That’s the whole point of competition and capitalism.

What would actually happen if Kroger had the ability to do perfect price discrimination based on facial recognition? Basic economics says there would be higher volumes and less deadweight loss. Assuming there was competition from other stores such that profits stayed roughly the same, consumers would massively benefit.

The actual danger is the ‘try to make my face seem normal’ step. If you can plausibly spend resources to fool the system, then that spends everyone’s time and effort on games that do not produce anything. That’s the part to avoid. We’ve been doing this for a long time, with coupons and sales and other gimmicks that do price discrimination largely on the basis of willingness to do extra work. If anything basing on facial recognition seems better at that, and dynamic pricing should be better at managing inventory as well.

According to Victor Shih, China essentially says, no profits for you. The state won’t look kindly upon you if you try to turn too much of a profit.

Dwarkesh Patel: I asked Victor Shih this question – why has the Chinese stock market been flat for so long despite the economy growing so fast?

This puzzle is explained via China’s system of financial repression.

If you save money in China, banks are not giving you the true competitive interest rate. Rather, they’ll give you the government capped 1.3% (lower than inflation, meaning you’re earning a negative return).

The net interest (which is basically a tax on all Chinese savers) is shoveled into politically favored state owned enterprises that survive only on subsidized credit.

But here’s what I didn’t understand at first: Why don’t companies just raise equity capital and operate profitably for shareholders?

The answer apparently is that there’s no ‘outside’ the system.

The state doesn’t just control credit – it controls land, permits, market access, even board seats through Party committees. Companies that prioritize profits over market share lose these privileges. Those that play along get subsidized loans, regulatory favors, and government contracts.

Regular savers, founders, and investors are all turned into unwitting servants of China’s industrial policy.

The obvious follow-up question is why is there not epic capital flight by every dollar that isn’t under capital controls? Who would ever invest in a Chinese company if they had a choice (other than me, a fool whose portfolio includes IEMG)? Certainly not anyone outside China, and those inside China would only do it if they couldn’t buy outside assets, even treasuries or outside savings accounts. No reason to stick around while they drink your milkshake.

This falls under the category of ‘things that if America contemplated doing even 10% of what China does, various people would say this will instantly cause us to ‘Lose to China’’. I very much have zero desire to do this one, but perhaps saying that phrase a lot should be a hint that something else is going on?

It’s also fun to see the cope, that this must be all that free market competition.

Amjad Msad: This doesn’t make much sense. China’s market is hyper competitive. In other words, it’s the opposite of socialist. That’s why you see thinner margins and more overall dynamism than US markets.

Yes, it’s hyper competitive, and the ways in which it is not socialist are vital to its ability to function, but that hyper competition is, as I understand it, ‘not natural,’ and very much not due to the invisible hand, but rather a different highly visible one.

We couldn’t pull it off and shouldn’t try. The PCR’s strategy is a package deal, the same way America’s strategy is a package deal, and our strategy has been the most successful in world history until and except where we started shooting ourselves in the foot. They are using their advantages and we must use ours. If we try to play by their rules, especially on top of our rules, they will win.

John Arnold: CA raised min wage for fast food workers 25% ($16 -> $20) and employment in the sector fell 3.2% in the first year. While I hate sectoral specific min wage laws, this is less than I’d have thought. That said, the real risk is tech substitutes for labor over long term, not year 1.

It’s interesting to see people’s biases as they respond to this paper.

Claude estimates 70% of employees were previously near minimum wage and only maybe 10% were previously over $20. It estimates the average wage shock at around 14%, although this is probably an underestimate due to spillover effects (as in, you have to adjust higher wages so that rank orders are preserved). If this was a more than 14% raise and employment only fell 3.2%, then on its face that is a huge win.

You then have to account for effects on things like hours, overtime and other work conditions, and for long term effects being larger than short term effects, since a lot of investments are locked in and adjustments take time. But I do think that this is a win for ‘minimum wage increases are not as bad for employment as one would naively think and can be welfare enhancing for the workers,’ of course it makes things worse for employers and customers.

A new report from the Ludwig Institute for Shared Economic Prosperity (uh huh) claims 60% of Americans cannot afford a ‘minimal quality of life.’

This is the correct reaction:

Zac Hill: This is just obviously false, though.

Daniel Eth: Okay, so… this “analysis” is obvious bullshit.

I mean, obviously. Are you saying the median American household can’t afford a ‘minimal quality of life’? That’s Obvious Nonsense. Here’s a few more details, I guess:

Megan Cerullo (CBS): LISEP tracks costs associated with what the research firm calls a “basket of American dream essentials.”

…

The Ludwig Institute also says that the nation’s official unemployment rate of 4.2% greatly understates the level of economic distress around the U.S. Factoring in workers who are stuck in poverty-wage jobs and people who are unable to find full-time employment, the U.S. jobless rate now tops 24%, according to LISEP, which defines these groups as “functionally unemployed.”

Claiming the ‘real unemployment rate’ is 24% is kind of a giveaway. So is saying that your ‘minimal quality of life’ costs $120,302 for a family of four, sorry what in hell? Looking at their methodology table of contents tells you a lot about what are they even measuring.

Luca Dellanna: The study considered necessary for “Minimal Quality of Life”, I kid you not, attendance of two MLB games per year per person.

Do these charlatans hope no one reads their studies?

That’s not quite fair, the MLB games and six movies a year are a proxy for ‘basic leisure,’ but that kind of thing is happening throughout.

I love this: Businesses ‘setting traps’ for private equity by taking down their website, so Private Equity Guy says ‘oh I can get a big win by giving them a website’ and purchases your business, but the benefits of your (previously real) website are already priced in. You don’t even have to fool them, they’ll fool themselves for you:

Lauren Balik: One of the biggest hacks for small business owners is removing your website in order to sell your company at a premium.

For example, I used to have a website, then I took it down and all of a sudden I was getting legitimate, fat offers to buy my business.

See, private equity people are lazy as hell. They get data from databases showing revenue proxies, run rate estimates, all kinds of ZoomInfo crap, etc. and they are willing to pay large premiums for easy, quick wins.

What’s the easiest, quickest win? Making a website for a business that has no website.

“Wow, this business is doing $1.5M a year and $500k EBITDA with no website. Imagine if we made a website! We could get this to $3M gross and $1.5M EBITDA overnight!”

Because private equity people are narcissistic, they don’t even consider that a small business owner may have outfoxed them and purposely taken down their website to set a trap.

You should be doing less, not more, and baiting snares for PE.

Hunter: Maybe a few fake bad reviews about the owners aren’t properly leveraging technology.

Mark Le Dain (from another thread): If you are planning to sell a plumbing company to PE make sure you get rid of the website before selling it They love to say “and I can’t even imagine what it will be like once we add a website and a CRM”

Lauren Balik: It even happens at scale. Subway pulled this on Roark lmfao.

People think I make stuff up. All throughout late 2023 and early 2024 Subway started breaking their own website and making it unusable for customers as Subway was trying to put pressure on Ayn Rand-inspired PE firm Roark Capital to close the acquisition.

Every time the website lost sales or went down it put more pressure on Roark to close the deal, which was finally completed in April 2024.

Should GDP include defense spending? The argument is it is there to enable other goods and services, it is not useful per se. To which I say, tons of other production is also there to enable other goods and services. Even if we’re talking purely about physical security, should we not count locksmiths or smoke alarms or firefighting? Should we not count bike helmets? Should we not count advertising? Should we not count goods that are not useful, or are positional or zero-sum? Lawyers? Accountants? All investments? Come on.

It is fine to say ‘there is a measure of non-defense production and it is more meaningful as a measure of living standards,’ sure, but that is not GDP. But if we are measuring living standards, a much bigger issue is that cost is very different from consumer surplus, especially regarding the internet and soon also AI.

Roon notices that as Taleb told us there is almost never a shortage that is not followed by a glut, and wonders why we ever need to panic about lack of domestic production if others want to subsidize our consumption. The answer is mostly mumble mumble politics, of course, except for certain strategically vital things we might lose access to (e.g. semiconductors) or where it’s the government that’s stopping us from producing.

Discussion about this post

Economics Roundup #6 Read More »

Experts urge caution about using ChatGPT to pick stocks

AI, AI assistants, AI in finance, AI risks, Biz & IT, chatbots, chatgpt, eToro, financial advice, Gemini, Google, Investment, machine learning, openai, retail investors, robo-advisory, Stock Market, UBS / Tim Belzer / September 25, 2025

“AI models can be brilliant,” Dan Moczulski, UK managing director at eToro, told Reuters. “The risk comes when people treat generic models like ChatGPT or Gemini as crystal balls.” He noted that general AI models “can misquote figures and dates, lean too hard on a pre-established narrative, and overly rely on past price action to attempt to predict the future.”

The hazards of AI stock picking

Using AI to trade stocks at home feels like it might be the next step in a long series of technological advances that have democratized individual retail investing, for better or for worse. Computer-based stock trading for individuals dates back to 1984, when Charles Schwab introduced electronic trading services for dial-up customers. E-Trade launched in 1992, and by the late 1990s, online brokerages had transformed retail investing, dropping commission fees from hundreds of dollars per trade to under $10.

The first “robo-advisors” appeared after the 2008 financial crisis, which began the rise of automated online services that use algorithms to manage and rebalance portfolios based on a client’s goals. Services like Betterment launched in 2010, and Wealthfront followed in 2011, using algorithms to automatically rebalance portfolios. By the end of 2015, robo-advisors from nearly 100 companies globally were managing $60 billion in client assets.

The arrival of ChatGPT in November 2022 arguably marked a new phase where retail investors could directly query an AI model for stock picks rather than relying on pre-programmed algorithms. But Leung acknowledged that ChatGPT cannot access data behind paywalls, potentially missing crucial analyses available through professional services. To get better results, he creates specific prompts like “assume you’re a short analyst, what is the short thesis for this stock?” or “use only credible sources, such as SEC filings.”

Beyond chatbots, reliance on financial algorithms is growing. The “robo-advisory” market, which includes all companies providing automated, algorithm-driven financial advice from fintech startups to established banks, is forecast to grow roughly 600 percent by 2029, according to data-analysis firm Research and Markets.

But as more retail investors turn to AI tools for investment decisions, it’s also potential trouble waiting to happen.

“If people get comfortable investing using AI and they’re making money, they may not be able to manage in a crisis or downturn,” Leung warned Reuters. The concern extends beyond individual losses to whether retail investors using AI tools understand risk management or have strategies for when markets turn bearish.

Experts urge caution about using ChatGPT to pick stocks Read More »

As many as 2 million Cisco devices affected by actively exploited 0-day

Biz & IT, Cisco, exploit, Security, zero-day, zero-day vulnerability / Tim Belzer / September 25, 2025

As many as 2 million Cisco devices are susceptible to an actively exploited zero-day that can remotely crash or execute code on vulnerable systems.

Cisco said Wednesday that the vulnerability, tracked as CVE-2025-20352, was present in all supported versions of Cisco IOS and Cisco IOS XE, the operating system that powers a wide variety of the company’s networking devices. The vulnerability can be exploited by low-privileged users to create a denial-of-service attack or by higher-privileged users to execute code that runs with unfettered root privileges. It carries a severity rating of 7.7 out of a possible 10.

Exposing SNMP to the Internet? Yep

“The Cisco Product Security Incident Response Team (PSIRT) became aware of successful exploitation of this vulnerability in the wild after local Administrator credentials were compromised,” Wednesday’s advisory stated. “Cisco strongly recommends that customers upgrade to a fixed software release to remediate this vulnerability.”

The vulnerability is the result of a stack overflow bug in the IOS component that handles SNMP (simple network management protocol), which routers and other devices use to collect and handle information about devices inside a network. The vulnerability is exploited by sending crafted SNMP packets.

To execute malicious code, the remote attacker must have possession of read-only community string, an SNMP-specific form of authentication for accessing managed devices. Frequently, such strings ship with devices. Even when modified by an administrator, read-only community strings are often widely known inside an organization. The attacker would also require privileges on the vulnerable systems. With that, the attacker can obtain RCE (remote code execution) capabilities that run as root.

As many as 2 million Cisco devices affected by actively exploited 0-day Read More »

A “cosmic carpool” is traveling to a distant space weather observation post

heliophysics, heliosphere, IMAP, NASA, NOAA, Science, solar wind, Space, space weather, Sun / Tim Belzer / September 24, 2025

“It’s like a bus. You wait for one and then three come at the same time.”

NASA’s IMAP spacecraft (top), the Carruthers Geocorona Observatory (left), and NOAA’s first operational space weather satellite (right) shared a ride to space on a Falcon 9 rocket Wednesday. Credit: SpaceX

Scientists loaded three missions worth nearly $1.6 billion on a SpaceX Falcon 9 rocket for launch Wednesday, toward an orbit nearly a million miles from Earth, to measure the supersonic stream of charged particles emanating from the Sun.

One of the missions, from the National Oceanic and Atmospheric Administration (NOAA), will beam back real-time observations of the solar wind to provide advance warning of geomagnetic storms that could affect power grids, radio communications, GPS navigation, air travel, and satellite operations.

The other two missions come from NASA, with research objectives that include studying the boundary between the Solar System and interstellar space and observing the rarely seen outermost layer of our own planet’s atmosphere.

All three spacecraft were mounted to the top of a Falcon 9 rocket for liftoff at 7: 30 am EDT (11: 30 UTC) on Wednesday from NASA’s Kennedy Space Center in Florida. The rocket arced on a trajectory heading east from Florida’s Space Coast, shed its reusable first stage booster for a landing offshore, then fired its upper stage engine twice to propel the trio of missions into deep space.

A few minutes later, each of the spacecraft separated from the Falcon 9 to begin a multi-month journey toward their observing locations in halo orbits around the L1 Lagrange point, a gravitational balance point roughly 900,000 miles (1.5 million kilometers) from Earth toward the Sun. The combined pull from the Earth and Sun at this location provides a stable region for satellites to operate in, and a good location for instruments designed for solar science.

Liftoff of IMAP and its two co-passengers on a Falcon 9 rocket. Credit: SpaceX

Seeing the big picture

The primary mission launched on Wednesday is called the Interstellar Mapping and Acceleration Probe (IMAP). The spin-stabilized IMAP spacecraft is shaped like a donut, with a diameter of about 8 feet (2.4 meters) and 10 science instruments looking inward toward the Sun and outward toward the edge of the heliosphere, the teardrop-shaped magnetic bubble blown outward by the solar wind.

At the edge of the heliosphere, the solar wind runs up against the interstellar medium, the gas, dust, and radiation in the space between the stars. This boundary remains a poorly understood frontier in space science, but it’s important because the heliosphere protects the Solar System from damaging galactic cosmic rays.

“IMAP is a mission of firsts,” said Nicky Fox, associate administrator of NASA’s science mission directorate. “It’ll be the first spacecraft dedicated to mapping the heliosphere’s outer boundary, a key piece in the heliophysics puzzle about the Sun’s influence on our Solar System. To do this, IMAP will spin every 15 seconds to measure the invisible using a very comprehensive suite of revolutionary instruments.”

During each rotation, IMAP’s sensors will scoop up all sorts of stuff: ions traveling 1 million miles per hour in the solar wind, interstellar dust particles, and energetic neutral atoms kicked back into the Solar System from the edge of the heliosphere.

“These energetic neutral atoms act as cosmic messengers,” said David McComas, IMAP’s principal investigator from Princeton University. “They’re unaffected by magnetic fields so they can propagate all the way in from the boundaries to Earth’s orbit and be measured by IMAP.”

Tracking these energetic neutral atoms will allow scientists to map the boundary of the heliosphere and what shapes it. The Sun’s movement through the Milky Way galaxy forms a shock wave on the front side of the heliosphere, similar to the wave created by the bow of a ship moving through water.

Artist’s illustration of the IMAP spacecraft in orbit. Credit: NASA

“We ended up with this fabulous observatory that measures everything,” McComas said. “The particles coming out from the Sun are moving out in the solar wind to get to the outer heliosphere. Some fraction of them become neutralized and come right back, and we observe them a few years later as ENAs (energetic neutral atoms). So, we’re really observing the entire life cycle of this particle energization and how it interacts at the boundaries of the heliosphere.”

IMAP follows a much smaller mission, named IBEX, that carried just two instruments to begin probing the edge of the heliosphere in 2008. IBEX discovered an unexpected ribbon-like pattern of energetic neutral emissions coming from the front of the heliosphere. Scientists have developed several theories to explain the ribbon signature. One of the theories postulates that the ribbon represents a group of particles that somehow leaked from the heliosphere and bounced around interstellar space before returning to the Solar System.

“It was found that interstellar matter, particles, and neutrals streaming in from outside the Solar System, actually… have a significant effect in how the entire heliosphere behaves,” said Shri Kanekal, IMAP’s mission scientist at NASA’s Goddard Space Flight Center.

IBEX’s discoveries fueled enthusiasm among space scientists for a more sophisticated follow-up mission like IMAP. NASA selected IMAP for development in 2018, and the $782 million mission will spend at least two years conducting scientific observations. The spacecraft was built at the Johns Hopkins University Applied Physics Laboratory in Laurel, Maryland.

The ribbon remains one of IBEX’s biggest discoveries. It refers to a vast, diagonal swath of energetic neutrals, painted across the front of the heliosphere. Credit: NASA/IBEX

“Immense value”

Two years after NASA approved IMAP for development, the agency’s heliophysics division selected another mission to head for the L1 Lagrange point. This smaller spacecraft, called the Carruthers Geocorona Observatory, hitched a ride to space with IMAP on Wednesday.

The $97 million Carruthers mission carries two co-aligned ultraviolet imagers designed for simultaneous observations of Earth’s exosphere, a tenuous cloud of hydrogen gas that fades into the airless void of outer space about halfway to the Moon. The hydrogen atoms in the exosphere generate a faint glow called the geocorona, which is only detectable in ultraviolet light at great distances. Images of the entire geocorona can’t be collected from a satellite in Earth orbit.

The mission is named for George Carruthers, an engineer and solar physicist who developed an ultraviolet camera placed on the Moon by the Apollo 16 astronauts in 1972. This camera captured the first view of the geocorona, a term coined by Carruthers himself.

The 531-pound (241-kilogram) Carruthers observatory was built by BAE Systems, with instruments provided by the University of California Berkeley’s Space Sciences Lab.

There’s a lot for scientists to learn from the Carruthers mission, because they know little about the exosphere or geocorona.

“We actually don’t know exactly how big it is,” said Lara Waldrop, the mission’s principal investigator from the University of Illinois Urbana-Champaign. “We don’t know whether it’s spherical or oval, how much it changes over time or even the density of its constituent hydrogen atoms.”

What scientists do know is that the exosphere plays an important role in shaping how solar storms affect the Earth. The exosphere is also the path by which the Earth is (very) slowly losing atomic hydrogen from water vapor lofted high into the atmosphere. “This process is extremely slow at Earth, and I’m talking billions of years. It is certainly nothing to worry about,” Waldrop ensures.

This image illustrates the location of the Sun-Earth L1 Lagrange point, where IMAP, Carruthers, and SWFO-L1 will operate. Credit: NOAA

The final spacecraft aboard Wednesday’s launch is the world’s first operational satellite dedicated to monitoring space weather. This $692 million mission is called the Space Weather Follow On-L1, or SWFO-L1, and serves as an “early warning beacon” for the potentially devastating effects of geomagnetic storms, said Irene Parker, deputy assistant administrator for systems at NOAA’s National Environmental Satellite, Data, and Information Service.

NOAA’s previous satellites peer down at Earth from low-Earth orbit or geosynchronous orbit, gathering data for numerical weather models and tracking the real-time movement of hurricanes and severe storms. Until now, NOAA has relied upon a hodgepodge of research satellites to monitor the solar wind upstream from Earth. SWFO-L1, also built by BAE Systems, is the first mission designed from the start for real-time, around-the-clock solar wind observations.

“We’ll use SWFO-L1 to buy power grid, airline, and satellite operators precious time to act before billion-dollar storms strike,” said Clinton Wallace, director of NOAA’s Space Weather Prediction Center.

Once on station around the L1 Lagrange point, the satellite will be renamed SOLAR-1 before NOAA declares it operational in mid-2026. The platform hosts four instruments, one of which is a coronagraph to detect the massive eruptions from the Sun that spark geomagnetic storms. The other instruments will sample solar particles as they pass over the spacecraft about a half-hour before they reach our planet.

These instruments are akin to weather satellites that detect a hurricane’s formation over the remote ocean and hurricane hunters that take direct measurements of the storm to assess its intensity before landfall, NOAA said.

Bundling IMAP, Carruthers, and SWFO-L1 onto the same rocket saved at least tens of millions of dollars in launch costs. Normally, they would have needed three different rockets.

Rideshare missions to low-Earth orbit are becoming more common, but spacecraft departing for more distant destinations like the L1 Lagrange point are rare. Getting all three missions on the same launch required extensive planning, a stroke of luck, and fortuitous timing.

“This is the ultimate cosmic carpool,” said Joe Westlake, director of NASA’s heliophysics division. “These three missions heading out to the Sun-Earth L1 point riding along together provide immense value for the American taxpayer.”

“It’s like a bus,” Fox said. “You wait for one and then three come at the same time.”

A “cosmic carpool” is traveling to a distant space weather observation post Read More »

Author name: Tim Belzer

Transparency instead of testing

Discussion about this post

The Green Hell as a must-visit for manufacturers

Mustang, deconstructed

More Corvette

’Ring battles

Behind the wheel

Healthy competition

Jeopardized efforts

Next three launches

Insect microcosm

DNA pieces

The first colonists

Slash and burn

Signs of apocalypse

Discussion about this post

The hazards of AI stock picking

Exposing SNMP to the Internet? Yep

Seeing the big picture

“Immense value”