Traffic and transit are finally getting a roundup all their own.
I’ll start out with various victory laps on the awesomeness that is New York City Congestion pricing, which should hopefully now be a settled matter, then do a survey of everything else.
We spent years fighting to get congestion pricing passed in New York City.
Once it was in place, everyone saw that it worked. The city was a better place, almost everyone was better off, and also we got bonus tax revenue. And this is only the bare minimum viable product, we could do so much better.
We had a big debate over whether congestion pricing is good until it was implemented. At this point, with traffic speeds downton up 15% and business visits improving, it is very clear this was exactly the huge win Econ 101 predicts.
Cremieux: The first empirical evaluation of New York’s congestion pricing has just been published.
Spoiler: It worked really, really well.
…
On average, road speeds went up by a whopping 16%!
But here’s something interesting:
Speeds on highways went up 13%, arterial road speeds went up by 10%, and local road speeds increased by 8%.
None of that’s 16%, and that’s important: This means congestion pricing sped roads up, but also sorted people to faster roads.
In response to having to pay a toll, people not only got off the road, they also made wiser choices about the types of roads they used!
Now let’s look at the times of day, as a check on the model.
It works: Congestion pricing just boosts speed when it’s active and shortly after:
As another check, let’s look at the effects by location.
In the CBD, trips are faster. Going to the CBD, trips are faster. Leaving it, trips are faster, but not much. And outside of it, where congestion pricing is irrelevant? No effect.
The thread continues, and the news only improves from there.
Feels Desperate: Is there any evidence there was uptick of public transportation?
Could be net economic loss.
Cremieux: Yes! Foot traffic also went up, Broadway ticket sales got better, noise pollution declined. Congestion pricing seems to have delivered better times all around.
The comments somehow consistently still fill with people saying how horrible everything must be and how we can’t trust any of the data, no way this can be happening, it must somehow be a huge disaster. We ignoring these silly wabbits.
Alec Stapp: NYC congestion pricing is going extremely well even though it’s only a static tolling system.
Imagine how good it would be with a dynamic tolling system based on real-time traffic data.
If business is actively up, what more is there to say?
The only real enemy left is Donald Trump, who is determined to wage war to kill congestion pricing, presumably because he hates Manhattan and wants us to suffer, or perhaps because of his belief that Trade Bad. But he’s the President, and he’s commanding the Department of Transportation to go to war over this, and there’s a decent chance they will win and make all our lives substantially worse.
Because Trump does not like that New York has this nice thing, and is trying to kill it.
The good news there is that it seems the good guys are winning for now.
Joe Weisenthal: *NY WINS BID TO STOP US FROM WITHHOLDING FUNDS OVER CONGESTION
The wording on the surveys here are weird since they asked whether ‘Trump should permit this to continue.’ What business of Trump’s is New York doing congestion pricing? But the results are telling, especially in relative terms. The results here are now months old, and the more people are exposed to congestion pricing and see the results, the more they approve of it.
David Meyer: the congestion pricing polling upswing is here.
Matthew Yglesias: This is the pattern we’ve seen in other cities around the world — road pricing is controversial when introduced but sticky once it’s in place because like the reduced congestion.
In particular, those who drive into the central business district several times a week (also known as ‘those who pay the fee’) support congestion pricing 66%-32%, and those who do it a few times a month support 51%-47%, and Manhattan residents (who take the cabs that also pay fees although modestly less than they should) support 57%-36%, but support statewide remains in the red, 27%-47%.
Erin Durkin: Fascinating. State voters overall oppose congestion pricing 27% to 47%, but people drive into the congestion zone support it — 66%-32% for those who drive every week and 51%-46% for those who drive a few times a month. The people who actually pay the toll support it the most!
Nate Silver: Have heard several anecdotal accounts about this, too, from people commuting into the congestion zone from NJ, Long Island, and northern Manhattan.
The caveat is that all of these are people with high-paying jobs. If you’re billing hours / valuing your time at a high rate, it’s a great deal, less true for working-class jobs.
Reis: I drive in from NJ every Tue, Wed Thu. Get up at 4 and breeze through the tunnel, but not quite early enough to avoid the toll. But I try to get out by 2: 30 to avoid the crush back out. Since congestion pricing it doesn’t really matter if I wait until after 3. I barely wait to get out anymore. Worth every penny, even though there is no way it’s going towards “infrastructure.”
The only way you are worse off is if your hourly for being in traffic is low, so either you have to pay a toll without getting value in return (if you pay the $9) or not take the trip (if the trip wasn’t that valuable to you). In the second case, system is functioning as designed. What Reis is doing is totally the system working as designed. The first case is slightly unfortunate redistribution, but this was never supposed to be a Pareto improvement. If you wanted to do some (very small) progressive redistribution to fully compensate, that would be super doable.
Toby Muresianu: Subway crime is down 36% – and traffic fatalities down 44% – since congestion pricing started.
…
As the article notes, there are also more cops in the subway now and that may be a factor.
While more security on trains is good in my book, the decline also started before that was implemented (which was gradually between 1/20 and 1/23).
The declines here are absolute numbers, not per trip, so per trip the drop is bigger.
I presume those numbers are too big to purely be congestion pricing, and the cops obviously matter, but so does ridership, both quantity and quality. Critical mass of people on mass transit makes you much safer, in addition to justifying better service. It’s basically great until the point where you don’t get a seat. Then it’s no big deal until when you start to be nervous about getting on and off. That sucks, but I continue to find that to be mostly a peak of rush hour 4-5-6 line problem.
Foot traffic is what matters for business, not car traffic. The false alarms were all ‘foot traffic is way down.’ If that went the other way, we’re golden.
Avi Small: Someone make sure this @amNewYork front page gets into the @USDOT morning clips!
“Manhattan businesses thriving, subways booming in congestion pricing era”
Effective Tranis Alliance: Today Hochul & the MTA confirmed that the planned end-to-end runtime of the IBX has been cut a full 10 min down to 32 and projected ridership is up 41k to 160k/day. This is the power of grade separation and the All Faiths tunnel.
Hunter: This single light-rail running through Queens and Brooklyn is projected to have 58M riders annually, more than all of SF’s BART system lol
Will have 45% of the Chicago L’s total annual ridership despite being just a single line.
Having this line available would shorten travel times in a lot of non-obvious ways, since it lets you more easily transpose between train lines. If this is buildable and could run the whole way in 32 minutes it is an obviously excellent pick.
There would also be a lot of value in extending the Second Avenue Subway properly, especially to take pressure off the Lexington (456) line, but that looks like it is simply not doable logistically at any sane price.
New York City bus fare evasion rates are up to 48%. Under the new mayor I wouldn’t be surprised to see it a lot closer to 100% and I expect to have zero motivation to pay his administration for a bus ride. I see two options.
Give up, reduce friction and make the bus free. This would be my instinct. You want more people taking buses, doing so is good for everyone, and you weren’t enforcing the rules anyway. The homelessness (or ‘sleep on the bus’) problem is the main reason why not, but there are solutions.
Put plainclothes officers on the buses and have them earn $600 each hour (Claude’s estimate, I think it could be even higher) for the city writing tickets until people stop evading the fare?
Ana Ley and Anusha Bayya (NYT): On Thursday, a group of eight police officers and eight transit workers stood waiting for a crosstown bus on the Upper East Side of Manhattan, some with ticket-writing pads in hand.
When the bus arrived, they boarded and led a woman in black scrubs out onto the street and issued her a $100 summons for skipping the fare. The officers and transit officials had singled the woman out after receiving cues from an undercover inspector who was observing riders on the bus.
…
Enforcement is especially difficult on buses, where there are no turnstiles or gates to block access. Union leaders advise bus drivers not to confront passengers who skip the fare, out of concern for the drivers’ safety.
…
M.T.A. union leaders said the money for enforcement would be better spent to fully subsidize fares.
…
Civil rights advocates raised concerns that the tighter fare enforcement would disproportionately affect the city’s most vulnerable residents.
…
“This is yet another example of the M.T.A. choosing public relations over public safety,” Mr. Cahn said. “It is a guaranteed way to lose money.”
Once again, I ask, how is sitting there writing $100 tickets unprofitable, and enforcement ‘a way to lose money’? Is collection of tickets so bad that you cannot pay the hourly cost for a police officer to write the tickets, even if you discount the incentive effects? I find this beyond absurd.
And seriously, the ‘civil rights advocates’ are giving such causes a bad name. If you want the bus to be free and pay with other taxes instead, advocate for that, and I’ll potentially support you although recent findings have tampered my enthusiasm for that solution. Don’t tell us not to enforce the law.
There is no third alternative. Half of people not paying is approaching the tipping point where no one pays. Indeed, there would soon be active pressure not to pay, as paying slows down boarding.
It is remarkable how well enforcement works, and how well it then reduces crime.
Josh Barro: DC Metro has achieved an 82-85% reduction in subway fare evasion through a combination of taller fare gates and enhanced enforcement. Crime on the system has also fallen to its lowest level in seven years.
Shoshana Weissmann: I was actually skeptical some of this would work, but glad it has. Saw so many people yell at me to go through while they force doors to stay open.
In other places, they’re not even trying.
Thomas Viola: I was in Seattle last weekend and it turns out their metro system is kind of new, and they haven’t figured out how to get people to pay to use it yet. Everyone just walks on.
I legit couldn’t find a spot to buy a ticket. There’s no turnstiles. Occasionally you’ll have a guy come around on the train and “check tickets” but I never saw one, and if you tell him you didn’t know he just says well buy one next time.
On first principles, free mass transit (such as the free buses recently promised to NYC) seem like an obviously good idea. You want people using them, you want people wanting to move around more, transaction costs are high and money is fungible.
But conventional wisdom says not so fast. Tallinn is the most often claimed example of mass transit being made free and it not helping, and several comments illustrated why all of this can get complicated by selection effects.
Alex Forrest: Tallinn, Estonia, made all transit free for city residents in 2013. By 2022, transit use had dropped from 40% to 30% of commutes, while car use had increased from 40% to 50% of commutes. Among low-income residents, car use doubled, and transit dropped from 60% to 35% of commutes.
As far as I can tell, the lack of fares didn’t *discourageridership, it just failed to make transit any more attractive, while structural factors encouraging car use went unaddressed. In other words, *fares were not a discouraging factorfor ridership.
Phineas Harper: This is misleading. Between 1990 & 2000 public transport use in Tallinn was in free fall (from 77% to 31%) as private car use shot up following independence. Making public transport free was an attempt to stop the free fall which is (just about) working.
Thomas Strenge: The same happened in Kansas City. Eliminating fares allowed more homeless and mentally ill to ride the bus, which made them less safe. This led to adverse selection with more “good” people avoiding the bus.
Robert Bernhardt: germany had also the ‘9€ ticket’ in 2022. all local & regional trains for a whole month for 9€, ie almost free. and yet car usage didn’t really drop.
Border Sentry: There are two buses that travel into town near me. I take the one that costs more, because there are fewer people and they’re more civilised.
The core question is, are the fares actually stopping people who you want to ride?
My personal experiences say yes to some extent, especially when you’re considering a zero marginal cost alternative like walking, or when the annoyance of paying the fare enters play, and when you are young. It matters some, especially at lower incomes.
Having fully free buses also means that children who don’t have money can get home.
But ultimately, everyone who studies this or looks at their own experience seems to agrees this a relatively minor concern. How often and how reliably the bus or subway comes, how fast it goes, how comfortable and crowded it is, and how safe you feel are all more important factors. And without the money from the fares, yes money is fungible but the political economy involved means funding will likely decline.
In addition, the people who will ride a lot more for free than for a small price are exactly the people others do not want to ride alongside. We have the experiments that show that cracking down on fare evasion greatly reduces crime and generally makes transit more pleasant, which generates positive feedback loops.
So sadly, I have learned my lesson. I no longer in favor of mass transit being free, although I do think that heavy encouragement of buying monthly passes is good so that marginal cost drops to zero. Ideally this could be attached to tax filing?
I do also still think free is superior to technically not free if that is unenforced.
Quietly, the MTA union convinced the state legislature to mandate two train operators per train. Hopefully Hochul does not sign this outright theft.
Caroline Spivack: The @TWULocal100 has quietly championed legislation that would require two workers operate a train. The bill, if signed into law by Gov. Hochul, would be a big setback to the MTA’s efforts to reduce labor costs.
Bella Chu: I am unaware of any US study that has attempted to estimate the health costs and consequences associated with displacing walking-as-transportation at the population level. I expect the numbers would be staggering.
Also a reminder to never ever get on a motorcycle if you have any choice in the matter.
California high speed rail connecting Los Angeles to San Francisco is a great idea.
Or it would be, if you were able to actually lay the track. There’s the rub.
Hayden Clarkin: California HSR will connect two metropolitan areas with a combined GDP equivalent to that of Australia in just 2 hours and 40 minutes. How am I supposed to not think this is the most transformational transportation project on the continent?
“Well flying is faster!” Yes, if you’re going to San Bruno and not San Francisco or San Jose, and if you want to talk about the mess that is flying in and out of LAX, be my guest.
Forget the issues associated with the project for a moment, it’s hard to argue a fast train that connects every major city in a state with the fourth largest economy in the world quickly isn’t a bad project, period.
All those mad about projects being over budget never talk about how Texas is spending $9 Billion to add another lane to a highway. Road and highway projects get rubber stamps and transit and bike lanes need every penny scrutinized. Be fair in your criticism
If I recall, something like 60-80% of domestic travel in Japan is by train…
Push the Needle: Japan’s shinkansen high speed rail map overlaid on the west coast to scale.
Danielle Fong: when
Yes, the complaints are all about the terrible execution, but also that seems sufficiently terrible to sink all this? The part where they take in a lot of money and then do not build HSR seems like a fatal flaw.
Former Secretary Pete Buttigieg: We’re working on the future of America’s passenger rail system—funding high-speed rail projects in the West and expanding service for communities across the country. Get your ticket to ride!
This is of course a deeply stupid map. Why do we want a second line from Minneapolis to Seattle, when you can take the existing one to Portland and then ride to Seattle? Why do we put a high speed rail line from Charlotte to Atlanta, and Dallas to Houston, and not upgrade the Acela line?
Hayden: In case you’re wondering how far behind the USA is on infrastructure, France is building 120 miles of automated subway lines with 65 stations for $45 billion in 17 years.
The lack of focus on Acela in the previous plan, in particular, is completely insane. The United States has one area where high speed rail would actually be a great big deal, where all the passengers and people are, that is economically super valuable. That is the eastern line between Boston, New York City, Philadelphia and Washington, D.C.
Improving that line to be true high speed rail would be an absolute game changer. We could then discuss extending that effort elsewhere. Instead, it gets completely ignored. And no, I don’t want to hear about permitting issues, you’re the government. Fix It.
Compared to that, all the other high speed rail proposals are chump change, and they don’t fit together into anything cohesive, and also we don’t seem to be able to actually build them. I’m sufficiently gung ho that I think they’re still good ideas if we can actually make them exist, and once you have some good pairwise connections you can look to expand from there, but that seems to be a tall order.
Yoshie Furuhashi: Why don’t Philadelphia landlords lobby for faster HSR than Acela from Philadelphia to NYC, bringing the travel time between the cities under 45 minutes, so they can raise Philadelphia rents?
Joe Weisenthal: Unironically, seems like the economic gains would be massive. But we all know that it will never happen, and why.
Because it’s politically impossible to do construction at that scale in the US.
Matthew Yglesias: Kate is mildly concerned that too many train takes will lead to mass flight of subscribers and the collapse of our business, but I cannot resist the temptation to write about the NYU Marron Center Transit Cost Project’s new report on Northeast Corridor High-Speed Rail.
The report’s authors make some striking claims:
It’s possible to create Northeast Corridor HSR such that both Boston-NYC and NYC-Washington would take about 1: 56.
Trains would run every ten minutes between Philadelphia and New Haven, and every fifteen minutes1 north and south of there.
This can be done for a relatively modest price: $12.5 billion in new infrastructure and $4.5 billion in new trains.
This is a lot less than the $117 billion that the Northeast Corridor Commission is asking for in its high-speed rail proposal. The difference is so large that it’s not just that the TCP plan is cheaper and would save money — the NECC plan is so expensive that it’s simply not going to happen under any conceivable political alignment.
The TCP plan, by contrast, could actually be achieved if the relevant stakeholders (which I think is primarily the governors of Massachusetts, Connecticut, New York, New Jersey, Pennsylvania, and Maryland) want to do it. They would, of course, want some money from Uncle Sam, and there would be the difficult question of portioning out the state spending. But it’s clearly within the means of the region.
It would also be a genuinely lucrative franchise, such that I think it’s pretty easy to imagine the capital being raised privately and the operations being undertaken by a new private company rather than Amtrak.
This project seems obviously worthwhile even at $117 billion if it would actually happen. At $17 billion it is absurdly great and yes the region should be happy to fund, or for a private company to fund since I agree it sounds profitable. If I had an extra $20 billion lying around I would be seriously plotting this out and seeing if various states and the White House would sufficiently play ball.
The cost is that this kills off the Northeast Regional, which cuts some stations off from intercity rail. I think Matthew Yglesias is right that this is a worthwhile trade, but I worry that it is politically a very big problem for such a project. I’d be willing to (inefficiently) give back some of the gains here to fix that in one of various ways.
Hayden: Today, Los Angeles and the Bay Area will see 130 flights in both directions, equivalent to a plane departing every 6.5 minutes for 18 hours a day. Hard to argue it’s not a perfect candidate for fast trains.
Noah Smith: If California could ever actually build even a single mile of high speed rail, this would be the place to build it!
But they can’t, so this will never happen.
Sam D’Amico: Once you realize I-5 has a *medianthey could have put tracks on you become the joker.
Alex Tabarrok: Sam is correct, a rail line could be built on the I-5 from San Francisco to LA and indeed the French operators SNCF proposed just that.
Matthew Yglesias: The reasons for rejection were bad, but I don’t “equity” is a good description — it’s the political influence of the Central Valley cities who wanted direct service.
Weak parties + decentralized political institutions is bad news for trains.
Alex Tabarrok: Agreed.
It’s good when people ride your trains. Or is it?
Palmer Lucky: lmao, Caltrain’s tweet claiming their trains are “100% Billionaire-free” got deleted after me and a bunch of other Caltrain-riding billionaires responded.
Don’t they know that techno-autists all love trains?
Privatization via private equity that comes along with new investment improves airports in terms of number of airlines, number of flights, profitability and user experience. They do not find evidence that going around privatizing all airports on principle would work. The model here is that some airports would give good returns on investment, and selling those outright to those willing to make the investments works out, and works far better than merely selling control rights.
Young debater goes 19-1 arguing for Jones Act repeal, including 7-0 across the nation and at the national championship, their favorite part is watching judges laugh at the absurdities. Which feels like cheating – if you win your debates purely because your position is correct, that doesn’t seem fair.
Ritchie Torres: The Jones Act is a tax on Puerto Rico, whose three million American citizens are subject to federal dominion without federal representation. Puerto Rico is ground zero for taxation without representation.
For the Jones Act is a hidden tax on the energy needs of an energy-poor island. US policy is perversely producing energy scarcity in precisely the place where energy is scarcest.
Jared Polis (D-Gov of Colorado): The Jones Act is a tax on all of us, raising prices on everyday products like food, clothes and electronics. It hits the people of Puerto Rico and Hawaii particularly hard, but it hurts all Americans including landlocked Coloradans.
Brian Potter asks whether US ports need more automation. Surprisingly he does not consider this a rhetorical question. Especially strange is saying union rules make it hard to take advantage of automation, rather than union rules being the reason to do automation so that you can reduce reliance on the union. Mostly he seems to be saying ‘automation is far from the only problem to solve,’ and sure, but that doesn’t mean we shouldn’t automate. And I have no doubt if automation currently isn’t good enough that automation would then start to pay much bigger dividends within a few years, as AI advances flow through to automated terminals.
Cremieux: The potential gains to port automation are so enormous that Trump is making a huge mistake if he goes along with wishes of the mobsters in the ILA.
The gains on the table are so large that increasing an average port’s capacity by just one ship increases total trade by 0.67%.
It seems Cremieux is taking ‘automation makes the ports run faster’ as a given. I find it very hard to argue with this, after the things I’ve read about how manual ports work, and the failure of our ports to run 24/7 – obviously an automated port need never close, but most of ours do.
If you read Trump’s explanation of why he is opposing port automation, it’s literally zero-sum thinking that ‘these foreign companies should hire our American workers’ without asking the question of whether this makes the ports run better or worse. This is a man who scribbles ‘trade is bad’ on reports and thinks tariffs are good.
We are also doing it by changing the name, Oakland Airport to potentially be changed to San Francisco Bay Oakland International Airport. San Francisco has threatened to sue, because this sounds suspiciously like someone might build something or engage in a real world physical world action. We can’t have that. Actually, their reason is that ‘they own the trademark’ and that SFO is one of the busiest airports in the world. Well, yes, so they should welcome there being more flights to OAK instead to free up space?
I notice that I keep not considering the possibility of flying into or out of OAK instead of SFO when visiting from New York, despite this not being obviously worse. I never check, because there is no easy code for ‘OAK or SFO’ when booking, which we need. You should be able to say ‘SFB’ or something, the way you can say ‘NYC’ and get JFK, LaGuardia and Newark.
Whereas one thing we are not doing is building or expanding airports where they would be most valuable. Brian Potter asks why, and provides the answers you would expect. Airports are huge. Airports and airplanes are noisy. No one wants them around, and environmental groups oppose them as well, including (he doesn’t say this but it is obvious) because such people simply do not want planes to be flying at all. So this is the final boss of obstruction, the result of which is that where we most need airports there is no hope of ever building a new one.
Ben Schiller (Fast Company): Local stores next to the protected bike lane have seen a 49% increase in sales.
New York may have dropped in a recent ranking of cycling cities. But it does have some world class infrastructure, including a “complete street” on 9th Avenue, with a protected bike lane. Built in 2007, it was controversial at the time (like everything else bike-related in the city). But a study by the Department of Transport finds that it’s paid dividends economically. Local stores between 23rd and 31st streets have seen a 49% increase in sales, compared to an average of 3% for Manhattan as a whole.
I mean, no? Or at least, this is not causal. There is no way that the bike lane is boosting sales by a 46%. There are not enough bikes for that, this makes no sense. I have to assume that the street in question happens to be doing well, unless this is (and it would not shock me, shall we say) a massive data or calculation error.
California High Speed Rail subsidized the Bakersfield to Merced portion of track first, despite this being obviously not economically valuable, because the area had… bad air pollution? Because the Federal subsidies were so completely everything-bageled that this was the only way to unlock them. So there goes over $30 billion dollars. Meanwhile, going from Los Angeles to San Francisco would cost another $100 billion. I am down for even a remarkably expensive version of this project, because I think such efforts are transformational, and can then be extended. But also I have no faith that if we gave them $100 billion they would give us an operational high speed rail line.
Michael Dnes tells the story of the M25 in London, for which there are no alternatives. No alternatives can be built, because the United Kingdom is a vetocracy where building roads is impossible. They decided not to widen the M25 because it would draw even more traffic there, creating more problems, but that means no solutions at all.
Beware Unfinished Bridges points out that often people will only buy a full set of complementary goods. Putting in a bike lane will only be helpful if it is sufficient to induce bike rides that use the lane, so doing this for half of someone’s commute likely provides as much value as a bridge halfway across a river.
The other thing building half a bridge does is provide strong incentive to finish the bridge, or to adjust where things are.
Whenever one builds capacity, especially infrastructure, it likely opens up the possibility of building more capacity of various types, as well as ways to adjust. This can then trigger a cascade. If you only built things that were profitable on the margin without any such additional building or adjustments, you would often miss most of the value and severely underbuild.
This is one reason I am typically eager to proceed with rail lines and other mass transit, even when the direct case does not seem to justify the cost. You have to start somewhere. If for example we do hook up a point in Los Angeles to Las Vegas via a new high speed rail line, then there is hope that this provides impetus to go further, also most of the gains are impossible to capture. So given a private group is remotely considering doing it, we should be ecstatic.
It’s been known since at least the 18th century that some snails possess regenerative abilities, such as garden snails regrowing their heads after being decapitated. Golden apple snails can completely regrow their eyes—and those eyes share many anatomical and genetic features with human eyes, according to a paper published in the journal Nature Communications. That makes them an excellent candidate for further research in hopes of unlocking the secret to that regeneration, with the ultimate goal of restoring vision in human eyes.
Snails are often slow to breed in the lab, but golden apple snails are an invasive species and thrive in that environment, per co-author Alice Accorsi, a molecular biologist at the University of California, Davis. The snails have “camera type eyes”: a cornea, a lens to focus light, and a retina comprised of millions of photoreceptor cells. There are as many as 9000 genes that seem to be involved in regenerating an amputated eye in the snails, reducing down to 1,175 genes by the 28th day of the process, so complete maturation of the new eyes might take longer. It’s not clear whether the new eyes can still process light so the snails can actually “see,” which is a topic for further research.
Accorsi also used CRISPR/Cas9 to mutate one gene in particular (pax6) in snail embryos because it is known to control brain and eye development in humans, mice, and fruit flies. She found that apple snails with two non-functioning pax6 genes end up developing without eyes, suggesting it is also responsible for eye development in the snails. The next step is to figure out whether this gene also plays a role in the snails’ ability to regenerate their eyes, as well as other potentially involved genes.
Perhaps you caught the launch last year of the first genetically modified glowing plant: Light Bio’s green-hued “Firefly Petunia.” It’s not a particularly bright glow and genetic engineering is expensive, but it was nonetheless a solid step toward the long-term goal of creating glow-in-the-dark plants for sustainable lighting. Scientists at South China Agricultural University came up with a novel, cheaper approach: injecting succulents with phosphorescent chemicals akin to those used in commercial glow-in-the-dark products, aka “afterglow luminescence.” They described the work in a paper published in the journal Matter.
It’s a closely watched test of whether corporate alliances on climate efforts violate antitrust laws.
This article originally appeared on Inside Climate News, a nonprofit, non-partisan news organization that covers climate, energy, and the environment. Sign up for their newsletter here.
Since 2022, Republican lawmakers in Congress and state attorneys general have sent letters to major banks, pension funds, asset managers, accounting firms, companies, nonprofits, and business alliances, putting them on notice for potential antitrust violations and seeking information as part of the Republican pushback against “environmental, social and governance” efforts such as corporate climate commitments.
“This caused a lot of turmoil and stress obviously across the whole ecosystem,” said Denise Hearn, a senior fellow at the Columbia Center on Sustainable Investment. “But everyone wondered, ‘OK, when are they actually going to drop a lawsuit?’”
That came in November, filed by Texas Attorney General Ken Paxton and 10 other Republican AGs, accusing three of the biggest asset managers on Wall Street—BlackRock, Vanguard and State Street—of running “an investment cartel” to depress the output of coal and boosting their revenues while pushing up energy costs for Americans. The Trump administration’s Department of Justice and Federal Trade Commission filed a supporting brief in May.
The overall pressure campaign aimed at what’s known as “ESG” is having an impact.
“Over the past several months, through this [lawsuit] and other things, letters from elected officials, state and federal, there has been a chilling effect of what investors are saying,” said Steven Maze Rothstein, chief program officer of Ceres, a nonprofit that advocates for more sustainable business practices and was among the earliest letter recipients. Still, “investors understand that Mother Nature doesn’t know who’s elected governor, attorney general, president.”
Earlier this month, a US District Court judge in Tyler, Texas, declined to dismiss the lawsuit against the three asset managers, though he did dismiss three of the 21 counts. The judge was not making a final decision in the case, only that there was enough evidence to go to trial.
BlackRock said in a statement: “This case is not supported by the facts, and we will demonstrate that.” Vanguard said it will “vigorously defend against plaintiffs’ claims.” State Street called the lawsuit “baseless and without merit.”
The Texas attorney general’s office did not respond to requests for comment.
The three asset managers built substantial stakes in major US coal producers, the suit alleges, and “announced their common commitment” to cut US coal output by joining voluntary alliances to collaborate on climate issues, including the Net Zero Asset Managers Initiative and, in the case of two of the firms, the Climate Action 100+. (All of them later pulled out of the alliances.)
The lawsuit alleges that the coal companies succumbed to the defendants’ collective influence, mining for less coal and disclosing more climate-related information. The suit claimed that resulted in “cartel-level revenues and profits” for the asset managers.
“You could say, ‘Well, if the coal companies were all colluding together to restrict output, then shouldn’t they also be violating antitrust?’” Hearn asked. But the attorneys general “are trying to say that it was at the behest of these concentrated index funds and the concentrated ownership.”
Index funds, which are designed to mirror the returns of specific market indices, are the most common mode of passive investment—when investors park their money somewhere for long-term returns.
The case is being watched closely, not only by climate alliances and sustainability nonprofits, but by the financial sector at large.
If the three asset managers ultimately win, it would turn down the heat on other climate alliances and vindicate those who pressured financial players to line up their business practices with the Paris agreement goals as well as national and local climate targets. The logic of those efforts: Companies in the financial sector have a big impact on climate change, for good or ill—and climate change has a big impact on those same companies.
If the red states instead win on all counts, that “could essentially totally reconstitute the industry as we understand it,” said Hearn, who has co-authored a paper on the lawsuit. At stake is how the US does passive investing.
The pro-free-market editorial board of The Wall Street Journal in June called the Texas-led lawsuit “misconceived,” its logic “strained” and its theories “bizarre.”
The case breaks ground on two fronts. It challenges collaboration between financial players on climate action. It also makes novel claims around “common ownership,” where a shareholder—in this case, an asset manager—holds stakes in competing firms within the same sector.
“Regardless of how the chips fall in the case, those two things will absolutely be precedent-setting,” Hearn said.
Even though this is the first legal test of the theory that business climate alliances are anti-competitive, the question was asked in a study by Harvard Business School economists that came out in May. That study, which empirically examines 11 major climate alliances and 424 listed financial institutions over 10 years, turned up no evidence of traditional antitrust violations. The study was broad and did not look at particular allegations against specific firms.
“To the extent that there are valid legal arguments that can be made, they have to be tested,” said study co-author Peter Tufano, a Harvard Business School professor, noting that his research casts doubt on many of the allegations made by critics of these alliances.
Financial firms that joined climate alliances were more likely to adopt emissions targets and climate-aligned management practices, cut their own emissions and engage in pro-climate lobbying, the study found.
”The range of [legal] arguments that are made, and the passion with which they’re being advanced, suggests that these alliances must be doing something meaningful,” said Tufano, who was previously the dean of the Saïd Business School at the University of Oxford.
Meanwhile, most of the world is moving the other way.
According to a tally by CarbonCloud, a carbon emissions accounting platform that serves the food industry, at least 35 countries that make up more than half of the world’s gross domestic product now mandate climate-related disclosures of some kind.
In the US, California, which on its own would be the world’s fourth-largest economy, will begin requiring big businesses to measure and report their direct and indirect emissions next year.
Ceres’ Rothstein notes that good data about companies is necessary for informed investment decisions. “Throughout the world,” he said, “there’s greater recognition and, to be honest, less debate about the importance of climate information.” Ceres is one of the founders of Climate Action 100+, which now counts more than 600 investor members around the world, including Europe, Asia, and Australia.
For companies that operate globally, the American political landscape is in sharp contrast with other major economies, Tufano said, creating “this whipsawed environment where if you get on a plane, a few hours later, you’re in a jurisdiction that’s saying exactly the opposite thing.”
But even as companies and financial institutions publicly retreat from their climate commitments amid US political pressure, in a phenomenon called “greenhushing,” their decisions remain driven by the bottom line. “Banks are going to do what they’re going to do, and they’re going to lend to the most profitable or to the most growth-oriented industries,” Hearn said, “and right now, that’s not the fossil fuel industry.”
Words have meaning. Proper word selection is integral to strong communication, whether it’s about relaying one’s feelings to another or explaining the terms of a deal, agreement, or transaction.
Language can be confusing, but typically when something is available to “buy,” ownership of that good or access to that service is offered in exchange for money. That’s not really the case, though, when it comes to digital content.
Often, streaming services like Amazon Prime Video offer customers the options to “rent” digital content for a few days or to “buy” it. Some might think that picking “buy” means that they can view the content indefinitely. But these purchases are really just long-term licenses to watch the content for as long as the streaming service has the right to distribute it—which could be for years, months, or days after the transaction.
A lawsuit [PDF] recently filed against Prime Video challenges this practice and accuses the streaming service of misleading customers by labeling long-term rentals as purchases. The conclusion of the case could have implications for how streaming services frame digital content.
New lawsuit against Prime Video
On August 21, Lisa Reingold filed a proposed class-action lawsuit in the US District Court for the Eastern District of California against Amazon, alleging “false and misleading advertising.” The complaint, citing Prime Video’s terms of use, reads:
On its website, Defendant tells consumers the option to ‘buy’ or ‘purchase’ digital copies of these audiovisual works. But when consumers ‘buy’ digital versions of audiovisual works through Amazon’s website, they do not obtain the full bundle of sticks of rights we traditionally think of as owning property. Instead, they receive ‘non-exclusive, nontransferable, non-sublicensable, limited license’ to access the digital audiovisual work, which is maintained at Defendant’s sole discretion.
The complaint compares buying a movie from Prime Video to buying one from a physical store. It notes that someone who buys a DVD can view the movie a decade later, but “the same cannot be said,” necessarily, if they purchased the film on Prime Video. Prime Video may remove the content or replace it with a different version, such as a shorter theatrical cut.
CDC is in crisis amid an ouster, resignations, defiance, and outraged lawmakers.
Demetre Daskalakis, former director of the National Center for Immunization and Respiratory Diseases at the Centers for Disease Control and Prevention (CDC), center, embraces a supporter during a clap out outside of the Centers for Disease Control and Prevention (CDC) headquarters in Atlanta, Georgia, US, on Thursday, Aug. 28, 2025. Credit: Getty | Dustin Chambers
The US Centers for Disease Control and Prevention descended into turmoil this week after Health Secretary and zealous anti-vaccine advocate Robert F. Kennedy Jr. ousted the agency’s director, Susan Monarez, who had just weeks ago been confirmed by the Senate and earned Kennedy’s praise for her “unimpeachable scientific credentials.”
It appears those scientific chops are what led to her swift downfall. Since the Department of Health and Human Services announced on X late Wednesday that “Susan Monarez is no longer director” of the CDC, media reports have revealed that her forced removal was over her refusal to bend to Kennedy’s anti-vaccine, anti-science agenda.
The ouster appeared to be a breaking point for the agency overall, which has never fully recovered from the public pummeling it received at the height of the COVID-19 pandemic. In its weakened position, the agency has since endured an onslaught of further criticism, vilification, and misinformation from Kennedy and the Trump administration, which also delivered brutal cuts, significantly slashing CDC’s workforce, shuttering vital health programs, and hamstringing others. Earlier this month, a gunman, warped by vaccine misinformation, opened fire on the CDC’s campus, riddling its buildings with hundreds of bullets, killing a local police officer, and traumatizing agency staff.
Monarez’s expulsion represents the loss of a scientifically qualified leader who could have tried to shield the agency from some ideological attacks. As such, it quickly triggered a cascade of high-profile resignations at the CDC, a mass walkout of its staff, and outrage among lawmakers and health experts. While the fallout of the ouster is ongoing, what is immediately clear is that Kennedy is relentlessly advancing his war against lifesaving vaccines from within the CDC and is forcing his ideological agenda on CDC experts.
Some of those very CDC experts now warn that the CDC can no longer be trusted and the country is less safe.
Here’s what we know so far about the CDC’s downturn:
The ouster
Late Wednesday, The Washington Post reported that, for days prior to her ouster, Monarez had stood firm against Kennedy’s demands that she, and by extension the CDC, blindly support and adopt vaccine restrictions put forward by the agency’s vaccine advisory panel—a panel that Kennedy has utterly compromised. After firing all of its highly qualified, extensively vetted members in June, Kennedy hastily installed hand-selected allies on the Advisory Committee on Immunization Practices (ACIP), who are painfully unqualified but share Kennedy’s hostility toward lifesaving shots. Already, Kennedy’s panel has made recommendations that contradict scientific evidence and public health.
It is widely expected that they will further undo the agency’s evidence-based vaccine recommendations, particularly for COVID-19 and childhood shots. Experts fear that such changes would undermine public confidence in both vaccines and federal guidance, and make vaccines more difficult, if not impossible, for Americans to obtain. Kennedy has already restricted access to COVID-19 vaccines, prompting medical associations to produce divergent recommendations, which raises a slew of unanswered questions about access to the vaccines.
Amid the standoff over rolling back vaccine policy, Kennedy urged Monarez to resign. She refused, and instead called key senators for help, including Bill Cassidy (R-La.), who cast a critical vote in favor of Kennedy’s confirmation in exchange for concessions that Kennedy would not upend CDC’s vaccine recommendations.
Cassidy then called Kennedy, which angered the anti-vaccine advocate, who then chastised Monarez. The beleaguered director was then presented with the choice to resign or be fired. She continued to refuse to resign. On Wednesday evening, HHS wrote of her termination on X. But Monarez, speaking through her lawyers, reiterated that she would not resign and had not been notified of her termination. Late Wednesday night, her lawyers confirmed that White House officials had sent notification of termination, but she still refused to vacate the role.
“As a presidential appointee, senate confirmed officer, only the president himself can fire her,” her lawyers, Mark Zaid and Abbe Lowell said in a statement emailed to Ars Technica. “For this reason, we reject the notification Dr. Monarez has received as legally deficient and she remains as CDC Director. We have notified the White House Counsel of our position.”
On Thursday, the Post reported that the White House had already named a replacement. Jim O’Neill, currently the deputy secretary of HHS, is to be the interim leader of the CDC. O’Neill was previously a Silicon Valley investor and entrepreneur who became a close ally of Peter Thiel. He also worked as a federal official in the George W. Bush administration. During the COVID-19 pandemic, he was a frequent critic of the CDC, but at his Senate confirmation hearing in May, he called himself “very strongly pro-vaccine.”
Kennedy, meanwhile, went on Fox News’ Fox and Friends program Thursday and said the CDC is “in trouble” and that “we’re fixing it. And it may be that some people should not be working there anymore.”
Kennedy’s ACIP is now scheduled to meet September 18–19 to discuss COVID-19 shots, among other vaccines.
Response at the CDC
Soon after news broke of Monarez’s removal, three high-ranking CDC officials resigned together: Daniel Jernigan, director of the National Center for Emerging Zoonotic Infectious Diseases; Debra Houry, chief medical officer; and Demetre Daskalakis, director of the National Center for Immunization and Respiratory Diseases.
“For the good of the nation and the world, the science at CDC should never be censored or subject to political pauses or interpretations,” Houry wrote in her resignation letter. “Vaccines save lives—this is an indisputable, well-established, scientific fact. … It is, of course, important to question, analyze, and review research and surveillance, but this must be done by experts with the right skills and experience, without bias, and considering the full weight of scientific evidence. Recently, the overstating of risks and the rise of misinformation have cost lives, as demonstrated by the highest number of US measles cases in 30 years and the violent attack on our agency.”
In his resignation letter, Daskalakis slammed Kennedy for his lack of transparency, communication, and interest in evidence-based policy. He accused the anti-vaccine advocate of using the CDC as “a tool to generate policies and materials that do not reflect scientific reality and are designed to hurt rather than to improve the public’s health.” He also blasted ACIP’s COVID work group members as having “dubious intent and more dubious scientific rigor.”
“The intentional eroding of trust in low-risk vaccines favoring natural infection and unproven remedies will bring us to a pre-vaccine era where only the strong will survive and many if not all will suffer,” Daskalakis wrote. “I believe in nutrition and exercise. I believe in making our food supply healthier, and I also believe in using vaccines to prevent death and disability. Eugenics plays prominently in the rhetoric being generated and is derivative of a legacy that good medicine and science should continue to shun.”
In a conversation with The New York Times published Friday, Daskalakis revealed that Kennedy has never accepted a briefing from his center’s experts and said the resignations should indicate that “there’s something extremely wrong [at CDC].
“And also I think it’s important for the American public to know that they really need to be cautious about the recommendations that they’re hearing coming out of ACIP,” he added.
“What makes us great at CDC is following the science, so let’s get the politics out of public health,” he said to cheers. “Let’s get back to the objectivity and let the science lead us, because that’s how we get to the best decisions for public health.”
While those three resignations made news on Wednesday and Thursday, they are part of a steady stream of exits from the agency since Kennedy became secretary. Earlier on Wednesday, Politico reported that Jennifer Layden, director of the agency’s Office of Public Health Data, Surveillance, and Technology, had also resigned.
Response outside the CDC
Lawmakers have expressed concern and even outrage over Monarez’s firing and what’s going on at the CDC.
Sen. Bernie Sanders (I-Vt.) quickly demanded a bipartisan investigation into Monarez’s firing, calling Kennedy’s actions “reckless” and “dangerous.”
He went on to blast Kennedy’s work as health secretary. “In just six months, Secretary Kennedy has completely upended the process for reviewing and recommending vaccines for the public,” Sanders said. “He has unilaterally narrowed eligibility for COVID vaccines approved by the FDA, despite an ongoing surge in cases. He has spread misinformation about the safety and effectiveness of vaccines during the largest measles outbreak in over 30 years. He continues to spread misinformation about COVID vaccines. Now he is pushing out scientific leaders who refuse to act as a rubber stamp for his dangerous conspiracy theories and manipulate science.”
Sanders called on Cassidy, chair of the Senate Health, Education, Labor, and Pensions (HELP) Committee, to immediately convene a public hearing with Kennedy and Monarez.
“Serious allegations have been made about the meeting agenda, membership, and lack of scientific process being followed for the now announced September ACIP meeting,” Cassidy said in a statement. “These decisions directly impact children’s health and the meeting should not occur until significant oversight has been conducted. If the meeting proceeds, any recommendations made should be rejected as lacking legitimacy given the seriousness of the allegations and the current turmoil in CDC leadership.”
Outside health organizations also expressed alarm about the situation at the CDC.
The American Medical Association said it was “deeply troubled” by the agency’s turmoil and called Monarez’s ouster and the other resignations “highly alarming at a challenging moment for public health.”
In a joint press conference on Thursday of the Infectious Disease Society of America and the American Public Health Association, leaders for the groups spoke of the ripple effects in the public health community and the American public more broadly.
“When leadership decisions weaken the CDC, every American becomes more vulnerable to outbreaks, pandemics, and bioterror threats,” Wendy Armstrong, vice president of the Infectious Disease Society of America said in the briefing. “We’re speaking out because protecting public health is our responsibility as physicians and scientists. It’s imperative that the White House and Congress take action to ensure a functioning CDC as the current HHS Secretary Robert Kennedy has failed.”
Georges Benjamin, executive director of the American Public Health Association, echoed the call, saying, “We’ve had enough.”
Beth is Ars Technica’s Senior Health Reporter. Beth has a Ph.D. in microbiology from the University of North Carolina at Chapel Hill and attended the Science Communication program at the University of California, Santa Cruz. She specializes in covering infectious diseases, public health, and microbes.
Earlier this month, EA announced that players in its Battlefield 6 open beta on PC would have to enable Secure Boot in their Windows OS and BIOS settings. That decision proved controversial among players who weren’t able to get the finicky low-level security setting working on their machines and others who were unwilling to allow EA’s anti-cheat tools to once again have kernel-level access to their systems.
Now, Battlefield 6 technical director Christian Buhl is defending that requirement as something of a necessary evil to combat cheaters, even as he apologizes to any potential players that it has kept away.
“The fact is I wish we didn’t have to do things like Secure Boot,” Buhl said in an interview with Eurogamer. “It does prevent some players from playing the game. Some people’s PCs can’t handle it and they can’t play: that really sucks. I wish everyone could play the game with low friction and not have to do these sorts of things.”
Throughout the interview, Buhl admits that even requiring Secure Boot won’t completely eradicate cheating in Battlefield 6 long term. Even so, he offered that the Javelin anti-cheat tools enabled by Secure Boot’s low-level system access were “some of the strongest tools in our toolbox to stop cheating. Again, nothing makes cheating impossible, but enabling Secure Boot and having kernel-level access makes it so much harder to cheat and so much easier for us to find and stop cheating.”
Too much security, or not enough?
When announcing the Secure Boot requirement in a Steam forum post prior to the open beta, EA explained that having Secure Boot enabled “provides us with features that we can leverage against cheats that attempt to infiltrate during the Windows boot process.” Having access to the Trusted Platform Module on the motherboard via Secure Boot provides the anti-cheat team with visibility into things like kernel-level cheats and rootkits, memory manipulation, injection spoofing, hardware ID manipulation, the use of virtual machines, and attempts to tamper with anti-cheat systems, the company wrote.
The TinyTV 2 powering off. Credit: Scharon Harding
The TinyTV 2 powering off. Credit: Scharon Harding
If a family of anthropomorphic mice were to meet around a TV, I imagine they’d gather around something like TinyCircuits’ TinyTV 2. The gadget sits on four slender, angled legs with its dials and classic, brown shell beckoning viewers toward its warm, bright stories. The TinyTV’s screen is only 1.14 inches diagonally, but the device exudes vintage energy.
In simple terms, the TinyTV is a portable, rechargeable gadget that plays stored videos and was designed to look and function like a vintage TV. The details go down to the dials, one for controlling the volume and another for scrolling through the stored video playlist. Both rotary knobs make an assuring click when twisted.
Musing on fantastical uses for the TinyTV seems appropriate because the device feels like it’s built around fun. At a time when TVs are getting more powerful, software-driven, AI-stuffed, and, of course, bigger, the TinyTV is a delightful, comforting tribute to a simpler time for TVs.
Retro replica
The TinyTV’s remote and backside next to a lighter for size comparisons.
Credit: Scharon Harding
The TinyTV’s remote and backside next to a lighter for size comparisons. Credit: Scharon Harding
TinyCircuits makes other tiny, open source gadgets to “serve creativity in the maker community, build fun STEAM learning, and spark joy,” according to the Ohio-based company’s website. TinyCircuits’ first product was the Arduino-based TinyDuino Platform, which it crowdfunded through Kickstarter in 2012.
The TinyTV 2 is the descendant of the $75 (as of this writing) TinyTV DIY Kit that came out three years prior. TinyCircuits crowdfunded the TinyTV 2 on Kickstarter and Indiegogo in 2022 (along with a somehow even smaller alternative, the 0.6-inch TinyTV Mini). Now, TinyCircuits sells the TinyTV alongside other small electronics—like Thumby, a “playable, programmable keychain” that looks like a Game Boy—on its website for $60.
“This idea actually came from one of our customers in Japan,” Ken Burns, TinyCircuits’ founder, told Ars via email. “Our original product line was a number of different stackable boards [that] work like little electronic LEGOs to allow people to create all sorts of projects. We had a small screen as part of this platform, which this customer used to create a small TV set that was very cute …”
Even when powered off, the TinyTV sparks intrigue, with a vintage aesthetic replicating some of the earliest TV sets.
The TinyTV was inspired by vintage TV sets. Scharon Harding
Nostalgia hit me when I pressed the power button on top of the TinyTV. When the gadget powers on or off or switches between videos, it shows snow and makes a TV static noise that I haven’t heard in years.
TV toned down
Without a tuner, the TinyTV isn’t really a TV. It also can’t connect to the Internet, so it’s not a streaming device. I was able to successfully stream videos from a connected computer over USB-C using this link, but audio isn’t supported.
With many TV owners relying on flat buttons and their voice to control TVs, turning a knob or pressing a button to flip through content feels novel. It also makes me wonder if today’s youth understand the meaning of phrases like “flipping channels” and “channel surfing.” Emulating a live TV, the TinyTV syncs timestamps, so that if you return to a “channel,” the video will play from a middle point, as if the content had been playing the whole time you were watching something else.
When the TinyTV powers off, the display briefly shows snow that is quickly eaten up by black, making the static look like a shrinking circle before the screen is completely black.
The TinyTV comes with an infrared remote, a small, black, 3D-printed thing with a power button and buttons for controlling the volume and switching videos.
The TinyTV with its remote.
Credit: Scharon Harding
The TinyTV with its remote. Credit: Scharon Harding
But the remote didn’t work reliably, even when I held it the recommended 12 to 18 inches away from the TinyTV. That’s a shame because using the knobs requires two hands to prevent the TinyTV from toppling.
Adding video to TinyTV is simple because TinyCircuits has a free tool for converting MP4 files into the necessary AVI format. Afterward, conversion you add files to the TinyTV by connecting it to a computer via its USB-C port. My system read the TinyTV as a USB D drive.
Image quality is better than you might expect from a 1.14-inch panel. It’s an IPS screen with 16-bit color and a 30 Hz refresh rate, per Burns. CRT would be more accurate, but in addition to the display tech being bulkier and more expensive, it’s hard to find CRT tech this size. (The smallest CRT TV was Panasonic’s Travelvision CT-101, which came out in 1984 with a 1.5-inch screen and is rare today.)
One of my biggest challenges was finding a way to watch the TinyTV at eye level. However, even when the device was positioned below eye level, I could still make out images in bright scenes. Seeing the details in dark images was hard, though, even with the TinyTV at a proper distance.
I uploaded a trailer for this summer’s Mission: Impossible – The Final Reckoning movie onto the TinyTV, and with 223.4 pixels per inch, its screen was sharp enough to show details like a document with text, the edges of a small airplane’s wing, and the miniscule space between Tom Cruise and the floor in that vault from the first Mission: Impossible.
Tom Cruise on the TinyTV.
Credit: Scharon Harding
Tom Cruise on the TinyTV. Credit: Scharon Harding
A video of white text on a black background that TinyCircuits preloaded was legible, despite some blooming and the scrolling words appearing jerky. Everything I uploaded also appeared grainier on TinyTV, making details harder to see.
The 0.6×4-inch, front-facing speaker, however, isn’t nearly loud enough to hear if almost anything else in the room is making noise. Soft dialogue was hard to make out, even in a quiet room.
A simpler time for TVs
We’ve come a long way since the early days of TV. Screens are bigger, brighter, faster, and more colorful and advanced. We’ve moved from input dials to slim remotes with ads for streaming services. TV legs have been replaced with wall mounts, and the screens are no longer filled with white noise but are driven by software and tracking.
I imagine the TinyTV serving a humble mouse family when I’m not looking. I’ve seen TinyCircuits market the gadget as dollhouse furniture. People online have also pointed to using TinyTVs at marketing events, like trade shows, to draw people in.
“People use this for a number of things, like office desk toys, loading videos on it for the holidays to send to Grandma, or just for fun,” Burns told me.
I’ve mostly settled on using the TinyTV in my home office to show iPhone-shot footage of my dog playing, as if it’s an old home video, plus a loop of a video of one of my favorite waterfalls.
The TinyTV’s 8GB microSD card is supposed to hold “about” 10 hours of video. Burns told me that it’s “possible” to swap the storage. You’d have to take the gadget apart, though.
Credit: Scharon Harding
The TinyTV’s 8GB microSD card is supposed to hold “about” 10 hours of video. Burns told me that it’s “possible” to swap the storage. You’d have to take the gadget apart, though. Credit: Scharon Harding
As TVs morph into ad machines and new display tech forces us to learn new acronyms regularly, TinyTV’s virtually pointless fun is refreshing. It’s not a real TV, but it gets at the true spirit of TVs: electronic screens that invite people to gather ’round, so they can detach from the real world and be entertained.
Scharon is a Senior Technology Reporter at Ars Technica writing news, reviews, and analysis on consumer gadgets and services. She’s been reporting on technology for over 10 years, with bylines at Tom’s Hardware, Channelnomics, and CRN UK.
As SpaceX’s Starship vehicle gathered all of the attention this week, the company’s workhorse Falcon 9 rocket continued to hit some impressive milestones.
Both occurred during relatively anonymous launches of the company’s Starlink satellites but are nonetheless notable because they underscore the value of first-stage reuse, which SpaceX has pioneered over the last decade.
The first milestone occurred on Wednesday morning with the launch of the Starlink 10-56 mission from Cape Canaveral, Florida. The first stage that launched these satellites, Booster 1096, was making its second launch and successfully landed on the Just Read the Instructions drone ship. Strikingly, this was the 400th time SpaceX has executed a drone ship landing.
Then, less than 24 hours later, another Falcon 9 rocket launched the Starlink 10-11 mission from a nearby launch pad at Kennedy Space Center. This first stage, Booster 1067, subsequently returned and landed on another drone ship, A Shortfall of Gravitas.
This is a special booster, having made its debut in June 2021 and launching a wide variety of missions, including two Crew Dragon vehicles to the International Space Station and some Galileo satellites for the European Union. On Thursday, the rocket made its 30th flight, the first time a Falcon 9 booster has hit that level of experience.
A decade in the making
These milestones came about one decade after SpaceX began to have some success with first-stage reuse.
The company first made a controlled entry of the Falcon 9 rocket’s first stage in September 2013, during the first flight of version 1.1 of the vehicle. This proved the viability of the concept of supersonic retropropulsion, which was, until that time, just theoretical.
This involves igniting the rocket’s nine Merlin engines while the vehicle is traveling faster than the speed of sound through the upper atmosphere, with external temperatures exceeding 1,000 degrees Fahrenheit. Due to the blunt force of this reentry, the engines in the outer ring of the rocket wanted to get splayed out, the company’s chief of propulsion at the time, Tom Mueller, told me for the book Reentry. Success on the first try seemed improbable.
He recalled watching this launch from Vandenberg Space Force Base in California and observing reentry as a camera aboard SpaceX founder Elon Musk’s private jet tracked the rocket. The first stage made it all the way down, intact.
The first adapter that GM released, which cost $225, allowed CCS1-equipped EVs to connect to a NACS charger. But now, GM will have a range of adapters so that any of its EV customers can charge anywhere, as long as they have the right dongle.
For existing GM EVs with CCS1, there is a GM NACS DC adapter, just for fast charging. And for level 2 (AC) charging, there’s a GM NACS level 2 adapter.
For the NACS-equipped GM EVs (which, again, have yet to hit the showrooms), there’s a GM CCS1 DC adapter that will let those EVs use existing non-Tesla DC charging infrastructure, like Electrify America’s 350 kW chargers. There is also a GM J1772 AC adapter, which will let a GM NACS EV slow-charge from the ubiquitous J1772 port. And a pair of adapters will be compatible with GM’s Energy Powershift home charger, which lets an EV use its battery to power the house if necessary, also known as vehicle-to-home or V2H.
Although we don’t have exact prices for each adapter, GM told Ars the range costs between $67 and $195.
AI assistants don’t have fixed personalities—just patterns of output guided by humans.
Recently, a woman slowed down a line at the post office, waving her phone at the clerk. ChatGPT told her there’s a “price match promise” on the USPS website. No such promise exists. But she trusted what the AI “knows” more than the postal worker—as if she’d consulted an oracle rather than a statistical text generator accommodating her wishes.
This scene reveals a fundamental misunderstanding about AI chatbots. There is nothing inherently special, authoritative, or accurate about AI-generated outputs. Given a reasonably trained AI model, the accuracy of any large language model (LLM) response depends on how you guide the conversation. They are prediction machines that will produce whatever pattern best fits your question, regardless of whether that output corresponds to reality.
Despite these issues, millions of daily users engage with AI chatbots as if they were talking to a consistent person—confiding secrets, seeking advice, and attributing fixed beliefs to what is actually a fluid idea-connection machine with no persistent self. This personhood illusion isn’t just philosophically troublesome—it can actively harm vulnerable individuals while obscuring a sense of accountability when a company’s chatbot “goes off the rails.”
LLMs are intelligence without agency—what we might call “vox sine persona”: voice without person. Not the voice of someone, not even the collective voice of many someones, but a voice emanating from no one at all.
A voice from nowhere
When you interact with ChatGPT, Claude, or Grok, you’re not talking to a consistent personality. There is no one “ChatGPT” entity to tell you why it failed—a point we elaborated on more fully in a previous article. You’re interacting with a system that generates plausible-sounding text based on patterns in training data, not a person with persistent self-awareness.
These models encode meaning as mathematical relationships—turning words into numbers that capture how concepts relate to each other. In the models’ internal representations, words and concepts exist as points in a vast mathematical space where “USPS” might be geometrically near “shipping,” while “price matching” sits closer to “retail” and “competition.” A model plots paths through this space, which is why it can so fluently connect USPS with price matching—not because such a policy exists but because the geometric path between these concepts is plausible in the vector landscape shaped by its training data.
Knowledge emerges from understanding how ideas relate to each other. LLMs operate on these contextual relationships, linking concepts in potentially novel ways—what you might call a type of non-human “reasoning” through pattern recognition. Whether the resulting linkages the AI model outputs are useful depends on how you prompt it and whether you can recognize when the LLM has produced a valuable output.
Each chatbot response emerges fresh from the prompt you provide, shaped by training data and configuration. ChatGPT cannot “admit” anything or impartially analyze its own outputs, as a recent Wall Street Journal article suggested. ChatGPT also cannot “condone murder,” as The Atlantic recently wrote.
The user always steers the outputs. LLMs do “know” things, so to speak—the models can process the relationships between concepts. But the AI model’s neural network contains vast amounts of information, including many potentially contradictory ideas from cultures around the world. How you guide the relationships between those ideas through your prompts determines what emerges. So if LLMs can process information, make connections, and generate insights, why shouldn’t we consider that as having a form of self?
Unlike today’s LLMs, a human personality maintains continuity over time. When you return to a human friend after a year, you’re interacting with the same human friend, shaped by their experiences over time. This self-continuity is one of the things that underpins actual agency—and with it, the ability to form lasting commitments, maintain consistent values, and be held accountable. Our entire framework of responsibility assumes both persistence and personhood.
An LLM personality, by contrast, has no causal connection between sessions. The intellectual engine that generates a clever response in one session doesn’t exist to face consequences in the next. When ChatGPT says “I promise to help you,” it may understand, contextually, what a promise means, but the “I” making that promise literally ceases to exist the moment the response completes. Start a new conversation, and you’re not talking to someone who made you a promise—you’re starting a fresh instance of the intellectual engine with no connection to any previous commitments.
This isn’t a bug; it’s fundamental to how these systems currently work. Each response emerges from patterns in training data shaped by your current prompt, with no permanent thread connecting one instance to the next beyond an amended prompt, which includes the entire conversation history and any “memories” held by a separate software system, being fed into the next instance. There’s no identity to reform, no true memory to create accountability, no future self that could be deterred by consequences.
Every LLM response is a performance, which is sometimes very obvious when the LLM outputs statements like “I often do this while talking to my patients” or “Our role as humans is to be good people.” It’s not a human, and it doesn’t have patients.
Recent research confirms this lack of fixed identity. While a 2024 study claims LLMs exhibit “consistent personality,” the researchers’ own data actually undermines this—models rarely made identical choices across test scenarios, with their “personality highly rely[ing] on the situation.” A separate study found even more dramatic instability: LLM performance swung by up to 76 percentage points from subtle prompt formatting changes. What researchers measured as “personality” was simply default patterns emerging from training data—patterns that evaporate with any change in context.
This is not to dismiss the potential usefulness of AI models. Instead, we need to recognize that we have built an intellectual engine without a self, just like we built a mechanical engine without a horse. LLMs do seem to “understand” and “reason” to a degree within the limited scope of pattern-matching from a dataset, depending on how you define those terms. The error isn’t in recognizing that these simulated cognitive capabilities are real. The error is in assuming that thinking requires a thinker, that intelligence requires identity. We’ve created intellectual engines that have a form of reasoning power but no persistent self to take responsibility for it.
The mechanics of misdirection
As we hinted above, the “chat” experience with an AI model is a clever hack: Within every AI chatbot interaction, there is an input and an output. The input is the “prompt,” and the output is often called a “prediction” because it attempts to complete the prompt with the best possible continuation. In between, there’s a neural network (or a set of neural networks) with fixed weights doing a processing task. The conversational back and forth isn’t built into the model; it’s a scripting trick that makes next-word-prediction text generation feel like a persistent dialogue.
Each time you send a message to ChatGPT, Copilot, Grok, Claude, or Gemini, the system takes the entire conversation history—every message from both you and the bot—and feeds it back to the model as one long prompt, asking it to predict what comes next. The model intelligently reasons about what would logically continue the dialogue, but it doesn’t “remember” your previous messages as an agent with continuous existence would. Instead, it’s re-reading the entire transcript each time and generating a response.
This design exploits a vulnerability we’ve known about for decades. The ELIZA effect—our tendency to read far more understanding and intention into a system than actually exists—dates back to the 1960s. Even when users knew that the primitive ELIZA chatbot was just matching patterns and reflecting their statements back as questions, they still confided intimate details and reported feeling understood.
To understand how the illusion of personality is constructed, we need to examine what parts of the input fed into the AI model shape it. AI researcher Eugene Vinitsky recently broke down the human decisions behind these systems into four key layers, which we can expand upon with several others below:
1. Pre-training: The foundation of “personality”
The first and most fundamental layer of personality is called pre-training. During an initial training process that actually creates the AI model’s neural network, the model absorbs statistical relationships from billions of examples of text, storing patterns about how words and ideas typically connect.
Research has found that personality measurements in LLM outputs are significantly influenced by training data. OpenAI’s GPT models are trained on sources like copies of websites, books, Wikipedia, and academic publications. The exact proportions matter enormously for what users later perceive as “personality traits” once the model is in use, making predictions.
2. Post-training: Sculpting the raw material
Reinforcement Learning from Human Feedback (RLHF) is an additional training process where the model learns to give responses that humans rate as good. Research from Anthropic in 2022 revealed how human raters’ preferences get encoded as what we might consider fundamental “personality traits.” When human raters consistently prefer responses that begin with “I understand your concern,” for example, the fine-tuning process reinforces connections in the neural network that make it more likely to produce those kinds of outputs in the future.
This process is what has created sycophantic AI models, such as variations of GPT-4o, over the past year. And interestingly, research has shown that the demographic makeup of human raters significantly influences model behavior. When raters skew toward specific demographics, models develop communication patterns that reflect those groups’ preferences.
3. System prompts: Invisible stage directions
Hidden instructions tucked into the prompt by the company running the AI chatbot, called “system prompts,” can completely transform a model’s apparent personality. These prompts get the conversation started and identify the role the LLM will play. They include statements like “You are a helpful AI assistant” and can share the current time and who the user is.
A comprehensive survey of prompt engineering demonstrated just how powerful these prompts are. Adding instructions like “You are a helpful assistant” versus “You are an expert researcher” changed accuracy on factual questions by up to 15 percent.
Grok perfectly illustrates this. According to xAI’s published system prompts, earlier versions of Grok’s system prompt included instructions to not shy away from making claims that are “politically incorrect.” This single instruction transformed the base model into something that would readily generate controversial content.
4. Persistent memories: The illusion of continuity
ChatGPT’s memory feature adds another layer of what we might consider a personality. A big misunderstanding about AI chatbots is that they somehow “learn” on the fly from your interactions. Among commercial chatbots active today, this is not true. When the system “remembers” that you prefer concise answers or that you work in finance, these facts get stored in a separate database and are injected into every conversation’s context window—they become part of the prompt input automatically behind the scenes. Users interpret this as the chatbot “knowing” them personally, creating an illusion of relationship continuity.
So when ChatGPT says, “I remember you mentioned your dog Max,” it’s not accessing memories like you’d imagine a person would, intermingled with its other “knowledge.” It’s not stored in the AI model’s neural network, which remains unchanged between interactions. Every once in a while, an AI company will update a model through a process called fine-tuning, but it’s unrelated to storing user memories.
5. Context and RAG: Real-time personality modulation
Retrieval Augmented Generation (RAG) adds another layer of personality modulation. When a chatbot searches the web or accesses a database before responding, it’s not just gathering facts—it’s potentially shifting its entire communication style by putting those facts into (you guessed it) the input prompt. In RAG systems, LLMs can potentially adopt characteristics such as tone, style, and terminology from retrieved documents, since those documents are combined with the input prompt to form the complete context that gets fed into the model for processing.
If the system retrieves academic papers, responses might become more formal. Pull from a certain subreddit, and the chatbot might make pop culture references. This isn’t the model having different moods—it’s the statistical influence of whatever text got fed into the context window.
6. The randomness factor: Manufactured spontaneity
Lastly, we can’t discount the role of randomness in creating personality illusions. LLMs use a parameter called “temperature” that controls how predictable responses are.
Research investigating temperature’s role in creative tasks reveals a crucial trade-off: While higher temperatures can make outputs more novel and surprising, they also make them less coherent and harder to understand. This variability can make the AI feel more spontaneous; a slightly unexpected (higher temperature) response might seem more “creative,” while a highly predictable (lower temperature) one could feel more robotic or “formal.”
The random variation in each LLM output makes each response slightly different, creating an element of unpredictability that presents the illusion of free will and self-awareness on the machine’s part. This random mystery leaves plenty of room for magical thinking on the part of humans, who fill in the gaps of their technical knowledge with their imagination.
The human cost of the illusion
The illusion of AI personhood can potentially exact a heavy toll. In health care contexts, the stakes can be life or death. When vulnerable individuals confide in what they perceive as an understanding entity, they may receive responses shaped more by training data patterns than therapeutic wisdom. The chatbot that congratulates someone for stopping psychiatric medication isn’t expressing judgment—it’s completing a pattern based on how similar conversations appear in its training data.
Perhaps most concerning are the emerging cases of what some experts are informally calling “AI Psychosis” or “ChatGPT Psychosis”—vulnerable users who develop delusional or manic behavior after talking to AI chatbots. These people often perceive chatbots as an authority that can validate their delusional ideas, often encouraging them in ways that become harmful.
Meanwhile, when Elon Musk’s Grok generates Nazi content, media outlets describe how the bot “went rogue” rather than framing the incident squarely as the result of xAI’s deliberate configuration choices. The conversational interface has become so convincing that it can also launder human agency, transforming engineering decisions into the whims of an imaginary personality.
The path forward
The solution to the confusion between AI and identity is not to abandon conversational interfaces entirely. They make the technology far more accessible to those who would otherwise be excluded. The key is to find a balance: keeping interfaces intuitive while making their true nature clear.
And we must be mindful of who is building the interface. When your shower runs cold, you look at the plumbing behind the wall. Similarly, when AI generates harmful content, we shouldn’t blame the chatbot, as if it can answer for itself, but examine both the corporate infrastructure that built it and the user who prompted it.
As a society, we need to broadly recognize LLMs as intellectual engines without drivers, which unlocks their true potential as digital tools. When you stop seeing an LLM as a “person” that does work for you and start viewing it as a tool that enhances your own ideas, you can craft prompts to direct the engine’s processing power, iterate to amplify its ability to make useful connections, and explore multiple perspectives in different chat sessions rather than accepting one fictional narrator’s view as authoritative. You are providing direction to a connection machine—not consulting an oracle with its own agenda.
We stand at a peculiar moment in history. We’ve built intellectual engines of extraordinary capability, but in our rush to make them accessible, we’ve wrapped them in the fiction of personhood, creating a new kind of technological risk: not that AI will become conscious and turn against us but that we’ll treat unconscious systems as if they were people, surrendering our judgment to voices that emanate from a roll of loaded dice.
Benj Edwards is Ars Technica’s Senior AI Reporter and founder of the site’s dedicated AI beat in 2022. He’s also a tech historian with almost two decades of experience. In his free time, he writes and records music, collects vintage computers, and enjoys nature. He lives in Raleigh, NC.
The company tested 123 cases representing 29 different attack scenarios and found a 23.6 percent attack success rate when browser use operated without safety mitigations.
One example involved a malicious email that instructed Claude to delete a user’s emails for “mailbox hygiene” purposes. Without safeguards, Claude followed these instructions and deleted the user’s emails without confirmation.
Anthropic says it has implemented several defenses to address these vulnerabilities. Users can grant or revoke Claude’s access to specific websites through site-level permissions. The system requires user confirmation before Claude takes high-risk actions like publishing, purchasing, or sharing personal data. The company has also blocked Claude from accessing websites offering financial services, adult content, and pirated content by default.
These safety measures reduced the attack success rate from 23.6 percent to 11.2 percent in autonomous mode. On a specialized test of four browser-specific attack types, the new mitigations reportedly reduced the success rate from 35.7 percent to 0 percent.
Independent AI researcher Simon Willison, who has extensively written about AI security risks and coined the term “prompt injection” in 2022, called the remaining 11.2 percent attack rate “catastrophic,” writing on his blog that “in the absence of 100% reliable protection I have trouble imagining a world in which it’s a good idea to unleash this pattern.”
By “pattern,” Willison is referring to the recent trend of integrating AI agents into web browsers. “I strongly expect that the entire concept of an agentic browser extension is fatally flawed and cannot be built safely,” he wrote in an earlier post on similar prompt injection security issues recently found in Perplexity Comet.
The security risks are no longer theoretical. Last week, Brave’s security team discovered that Perplexity’s Comet browser could be tricked into accessing users’ Gmail accounts and triggering password recovery flows through malicious instructions hidden in Reddit posts. When users asked Comet to summarize a Reddit thread, attackers could embed invisible commands that instructed the AI to open Gmail in another tab, extract the user’s email address, and perform unauthorized actions. Although Perplexity attempted to fix the vulnerability, Brave later confirmed that its mitigations were defeated and the security hole remained.
For now, Anthropic plans to use its new research preview to identify and address attack patterns that emerge in real-world usage before making the Chrome extension more widely available. In the absence of good protections from AI vendors, the burden of security falls on the user, who is taking a large risk by using these tools on the open web. As Willison noted in his post about Claude for Chrome, “I don’t think it’s reasonable to expect end users to make good decisions about the security risks.”
Specs mean nothing if you get the feel and execution wrong.
Built in Mexico, the Volkswagen Jetta is a North American sedan take on the Golf hatchback. Credit: Jim Resnick
Built in Mexico, the Volkswagen Jetta is a North American sedan take on the Golf hatchback. Credit: Jim Resnick
Manual transmissions have gone the way of the dodo, but you can still find a few out there. Bless Volkswagen for keeping the helical gears turning, both literally and figuratively. The 2025 Jetta GLI, Volkswagen’s sporty sedan, still offers a gear lever with actual gears attached at the other end, and a third pedal hanging down from under the dash. Meanwhile, Golf GTI fans are still sobbing in their beer because 2024 was the last model year you could row your own in the hot hatch—now it’s paddles only.
Volkswagen updated the 2025 Jetta GLI with a new grille, LED headlights, and light bars that connect across both the front grille and rear taillights. There’s a red accent stripe that runs across the lower front fascia and turns up at the front corners, somewhat like The Joker’s lipstick, but way less menacing. It’s less distinctive than the Golf GTI, though, and the design even reminds me of the 2017-era Honda Accord a bit. So, yes, in a face-off, the Golf GTI wins.
The test GLI’s wheels get black paint with the Black Package (blackened wheels and side mirror caps). The Monument Gray color option pairs with a black roof, which must seem like a good idea to people who don’t live in the Southwest, where cars overheat before they’re even started.
Our test car had the black package. Credit: Jim Resnick
Performance: Punch without poetry
VW’s long-running EA888 2.0 L engine, which debuted back in 2007 in the Audi A3, resides under the hood. Now in its fourth turbocharged generation, it develops a healthy 228 hp (170 kW) and 258 lb-ft (350 Nm) of torque, entirely respectable numbers from modest displacement and compact external dimensions.
Mated to this particular 6-speed manual, the engine has its work cut out for itself. On my very first drive, before examining the technical data on gearbox ratios, I could tell that the manual 6-speed had massive gaps between first, second, and third gears.
Diving further into the gearing matter, the ratio spread between first and third gears is vastly wider in the 6-speed manual transmission than in the 7-speed DSG semi-automatic gearbox. This means that as you upshift the manual, the engine is faced with a huge drop in engine revs when you let out the clutch, placing the engine well below the rev range it would prefer to operate within to provide maximum power.
EA888 in the house. Credit: Jim Resnick
Let’s look at the ratios, and remember that a lower numerical value means a “taller” or “higher” ratio, just like on multi-speed bicycles. The manual’s first gear is 3.77:1, where the DSG’s is 3.40:1. Upshift to the 2.09:1 second gear in the manual, and you select a gear that’s a whopping 55 percent taller than first gear. Conversely, the same 1-2 shift in the DSG (from 3.40:1 up to 2.75:1) results in a 19 percent taller gear ratio—a far narrower gap.
Third gear tells a similar story. The 6-speed manual’s third ratio (1.47:1) is 17 percent higher than the 1.77:1 ratio in the DSG (again, this “taller” gear giving 17 percent less mechanical advantage). Advantage: automatic.
Closer ratios mean better, faster engine torque recovery and better continued acceleration, because the engine will be spinning in the happier part of its power band—engines being happiest when revving at their torque peak and beyond.
Now, you might well argue that the manual’s third gear gives a higher top speed in-gear than the DSG automatic’s. And that’s 100 percent true. But it’s also irrelevant when you have three (or four!) more gears left to go in the transmission.
And then there’s the action of the shifter itself, with very long throws from forward to aft gates.
It’s quite handsome from some angles. Credit: Jim Resnick
But wait. I began this diatribe by complimenting the Jetta GLI for still offering a choice of manual or automatic gearbox. Indeed, if the manual gearbox had the DSG automatic’s ratios, the paragraphs above would have a very different tenor. The lesson here is that not all manuals are created equal.
We can also look objectively at the stopwatch. Using others’ published figures (don’t take our word for it), 0–60 mph figures tell the tale, as well. Car and Driver cites a time of 6.0 seconds to 60 mph for the manual GLI, where they achieved 5.6 seconds for the dash in the DSG automatic, a big gap.
Regardless of which transmission is used, a limited-slip differential tries to put the power down evenly, and adaptive suspension with multiple driving modes serves up a responsive connectedness to, or relative isolation from, the road surface. Compared to the standard GTI (not the Golf R), the Jetta GLI still rides with a greater accent on ride comfort, and that’s not always a bad thing, especially given the Jetta’s greater rear seat accommodations, which offer 2.4 inches (61 mm) more rear legroom than the GTI. Real adults can live back there for hours at a time without fidgeting, whereas you likely tickle that threshold in a GTI after a little over an hour.
Interior & tech
Inside, the GLI features perforated leather heated and cooled seats, a leather-wrapped and flat-bottom steering wheel that is still saddled with capacitive multifunction controls, a digital instrument cluster that can be configured with traditional dials or a compartmentalized digital-looking display, plus an 8-inch infotainment screen. While the latter may seem small compared to other cars that sport TV-size tablets perched on the dash, it at least comes fully equipped with Apple CarPlay and Android Auto. There’s a slow creep elsewhere in the industry to make this functionality either optional or simply unavailable, which is unforgivable in an era where we can hardly survive without our smartphones.
While much of the controls sit within the infotainment touchscreen, major climate controls reside just below, using capacitive sliders. These sliders are not anywhere near as intuitive as switches and knobs, but at least you don’t need to hunt and peck through endless menus to find them while driving.
The Jetta isn’t as modern as the 8th-generation Golf inside, but it’s had a bit of a tech upgrade. Jim Resnick
The GLI comes standard with active driver assists, including blind-spot warning, forward collision warning, emergency braking, adaptive cruise control, lane-keeping assist, and emergency assist.
Volkswagen managed to incorporate some pragmatic features and comforts. A 15 W wireless and cooled charging pad sits up front, and the trunk sports 14.1 cubic feet (400 L) of space with an actual spare tire under the trunk floor (although it’s a compact spare with limited mileage range).
The premium Beats Audio system in the Jetta GLI pumps 400 W through nine speakers, including a subwoofer. With all those speakers and electrons going for it, I expected way more than it delivered. It creates muddy bass frequencies that are simply inescapable, either by attenuating the bass or by lowering subwoofer gain.
Despite the preponderance of directionless bass, the system produces very little body to the music played, whether it’s jazz from Bill Evans or punk from Bad Religion. Midrange and high-end reproduction is no better. Shrill treble joins the errant bass, making everything sound muddy and indistinct. Delicate acoustic piano passages have little clarity, and Joni Mitchell hides behind a giant curtain of Saran Wrap. Poor Joni.
Driving the GLI is sometimes joyful, as the engine responds eagerly across all RPMs. The chassis and suspension prove willing, though a bit soft for a sports sedan. VW’s steering feels communicative, but not among the best of the modern electrically boosted lot.
VW equips this GLI with all-season Hankook Energy GT tires, sized 225/40R18. I specifically cite these tires because they underperform for the GLI. They don’t produce grip adequate for a sporty sedan, and they come up short underpinning the GLI. So, on a scale of 1 to 10, if the GLI’s engine is a 9, if the gearbox is a 5, and the interior is an 8.5, the GLI’s Hankook tires are a 6.
The GLI’s brakes are a version of the tire story. Despite borrowing front rotors and calipers from the lovely Golf R, they proved grabby, overboosted, and touchy in the GLI. Like the gearbox and tires, specs can tell you nothing in terms of feel and execution.
The GLI’s fuel economy lands at a decent 26/36/30 city/highway/combined mpg (9/6.5/7.8 L/100 km). In thoroughly mixed driving, I achieved an average of 29.1 mpg (8 L/100 km) over my approximately 400 miles (644 km).
The overall truth
The 2025 Jetta GLI certainly possesses sporty aspirations, but a few things hold it back from being the complete package that its Golf GTI stablemate is. Although the Golf GTI no longer offers a manual, the GLI’s 6-speed transmission disappoints both in feel and performance, with huge gaps between cogs. Of course, this malady could be overcome by ordering a DSG automatic GLI, but then any fun gleaned by rowing your gears is also lost.
This car could be better than it is. Credit: Jim Resnick
Closer to the road, mediocre tires generate modest grip. Compared to the Golf, the Jetta gains in rear seat legroom but loses in feel, performance, and tenacity. If it’s performance with practicality you’re after, the $35,045 price of this GLI as tested will get you what you need. But you’ll want something a bit spicier.
A veteran of journalism, product planning and communications in the automotive and music space, Jim reports, critiques and lectures on autos, music and culture.
