Biz & IT

doing-dns-and-dhcp-for-your-lan-the-old-way—the-way-that-works

Doing DNS and DHCP for your LAN the old way—the way that works

BIND, bind9, Biz & IT, dhcpd, DNS, domain name system, Feature, Features, homelab, it projects, Projects, weekend projects / /
All shall tremble before your fully functional forward and reverse lookups!

Enlarge / All shall tremble before your fully functional forward and reverse lookups!

Aurich Lawson | Getty Images

Here’s a short summary of the next 7,000-ish words for folks who hate the thing recipe sites do where the authors babble about their personal lives for pages and pages before getting to the cooking: This article is about how to install bind and dhcpd and tie them together into a functional dynamic DNS setup for your LAN so that DHCP clients self-register with DNS, and you always have working forward and reverse DNS lookups. This article is intended to be part one of a two-part series, and in part two, we’ll combine our bind DNS instance with an ACME-enabled LAN certificate authority and set up LetsEncrypt-style auto-renewing certificates for LAN services.

If that sounds like a fun couple of weekend projects, you’re in the right place! If you want to fast-forward to where we start installing stuff, skip down a couple of subheds to the tutorial-y bits. Now, excuse me while I babble about my personal life.

My name is Lee, and I have a problem

(Hi, Lee.)

I am a tinkering homelab sysadmin forever chasing the enterprise dragon. My understanding of what “normal” means, in terms of the things I should be able to do in any minimally functioning networking environment, was formed in the days just before and just after 9/11, when I was a fledgling admin fresh out of college, working at an enormous company that made planes starting with the number “7.” I tutored at the knees of a whole bunch of different mentor sysadmins, who ranged on the graybeard scale from “fairly normal, just writes his own custom GURPS campaigns” to “lives in a Unabomber cabin in the woods and will only communicate via GPG.” If there was one consistent refrain throughout my formative years marinating in that enterprise IT soup, it was that forward and reverse DNS should always work. Why? Because just like a clean bathroom is generally a sign of a nice restaurant, having good, functional DNS (forward and reverse) is a sign that your IT team knows what it’s doing.

Just look at what the masses have to contend with outside of the datacenter, where madness reigns. Look at the state of the average user’s LAN—is there even a search domain configured? Do reverse queries on dynamic hosts work? Do forward queries on dynamic hosts even work? How can anyone live like this?!

I decided long ago that I didn’t have to, so I’ve maintained a linked bind and dhcpd setup on my LAN for more than ten years. Also, I have control issues, and I like my home LAN to function like the well-run enterprise LANs I used to spend my days administering. It’s kind of like how car people think: If you’re not driving a stick shift, you’re not really driving. I have the same kind of dumb hang-up, but for network services.

Honestly, though, running your LAN with bind and dhcpd isn’t even that much work—those two applications underpin a huge part of the modern Internet. The packaged versions that come with most modern Linux distros are ready to go out of the box. They certainly beat the pants off of the minimal DNS/DHCP services offered by most SOHO NAT routers. Once you have bind and dhcpd configured, they’re bulletproof. The only time I interact with my setup is if I need to add a new static DHCP mapping for a host I want to always grab the same IP address.

So, hey, if the idea of having perfect forward and reverse DNS lookups on your LAN sounds exciting—and, come on, who doesn’t want that?!—then pull up your terminal and strap in because we’re going make it happen.

(Note that I’m relying a bit on Past Lee and this old blog entry for some of the explanations in this piece, so if any of the three people who read my blog notice any similarities in some of the text, it’s because Past Lee wrote it first and I am absolutely stealing from him.)

But wait, there’s more!

This piece is intended to be part one of two. If the idea of having one’s own bind and dhcpd servers sounds a little silly (and it’s not—it’s awesome), it’s actually a prerequisite for an additional future project with serious practical implications: our own fully functioning local ACME-enabled certificate authority capable of answering DNS-01 challenges so we can issue our own certificates to LAN services and not have to deal with TLS warnings like plebes.

(“But Lee,” you say, “why not just use actual-for-real LetsEncrypt with a real domain on my LAN?” Because that’s considerably more complicated to implement if one does it the right way, and it means potentially dealing with split-horizon DNS and hairpinning if you also need to use that domain for any Internet-accessible stuff. Split-horizon DNS is handy and useful if you have requirements that demand it, but if you’re a home user, you probably don’t. We’ll keep this as simple as possible and use LAN-specific DNS zones rather than real public domain names.)

We’ll tackle all the certificate stuff in part two—because we have a ways to go before we can get there.

Doing DNS and DHCP for your LAN the old way—the way that works Read More »

openai-collapses-media-reality-with-sora,-a-photorealistic-ai-video-generator

OpenAI collapses media reality with Sora, a photorealistic AI video generator

AI, Ai video, Ai video generator, Biz & IT, chatgpt, chatgtp, cultural singularity, dall-e, DALL-E 3, deepfake, Deepfakes, image synthesis, machine learning, openai, OpenAI Sora, Sora, Stable Diffusion, video generator, video synthesis, will smith / /

Pics and it didn’t happen —

Hello, cultural singularity—soon, every video you see online could be completely fake.

–

Snapshots from three videos generated using OpenAI's Sora.

Enlarge / Snapshots from three videos generated using OpenAI’s Sora.

On Thursday, OpenAI announced Sora, a text-to-video AI model that can generate 60-second-long photorealistic HD video from written descriptions. While it’s only a research preview that we have not tested, it reportedly creates synthetic video (but not audio yet) at a fidelity and consistency greater than any text-to-video model available at the moment. It’s also freaking people out.

“It was nice knowing you all. Please tell your grandchildren about my videos and the lengths we went to to actually record them,” wrote Wall Street Journal tech reporter Joanna Stern on X.

“This could be the ‘holy shit’ moment of AI,” wrote Tom Warren of The Verge.

“Every single one of these videos is AI-generated, and if this doesn’t concern you at least a little bit, nothing will,” tweeted YouTube tech journalist Marques Brownlee.

For future reference—since this type of panic will some day appear ridiculous—there’s a generation of people who grew up believing that photorealistic video must be created by cameras. When video was faked (say, for Hollywood films), it took a lot of time, money, and effort to do so, and the results weren’t perfect. That gave people a baseline level of comfort that what they were seeing remotely was likely to be true, or at least representative of some kind of underlying truth. Even when the kid jumped over the lava, there was at least a kid and a room.

The prompt that generated the video above: “A movie trailer featuring the adventures of the 30 year old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colors.“

Technology like Sora pulls the rug out from under that kind of media frame of reference. Very soon, every photorealistic video you see online could be 100 percent false in every way. Moreover, every historical video you see could also be false. How we confront that as a society and work around it while maintaining trust in remote communications is far beyond the scope of this article, but I tried my hand at offering some solutions back in 2020, when all of the tech we’re seeing now seemed like a distant fantasy to most people.

In that piece, I called the moment that truth and fiction in media become indistinguishable the “cultural singularity.” It appears that OpenAI is on track to bring that prediction to pass a bit sooner than we expected.

Prompt: Reflections in the window of a train traveling through the Tokyo suburbs.

OpenAI has found that, like other AI models that use the transformer architecture, Sora scales with available compute. Given far more powerful computers behind the scenes, AI video fidelity could improve considerably over time. In other words, this is the “worst” AI-generated video is ever going to look. There’s no synchronized sound yet, but that might be solved in future models.

How (we think) they pulled it off

AI video synthesis has progressed by leaps and bounds over the past two years. We first covered text-to-video models in September 2022 with Meta’s Make-A-Video. A month later, Google showed off Imagen Video. And just 11 months ago, an AI-generated version of Will Smith eating spaghetti went viral. In May of last year, what was previously considered to be the front-runner in the text-to-video space, Runway Gen-2, helped craft a fake beer commercial full of twisted monstrosities, generated in two-second increments. In earlier video-generation models, people pop in and out of reality with ease, limbs flow together like pasta, and physics doesn’t seem to matter.

Sora (which means “sky” in Japanese) appears to be something altogether different. It’s high-resolution (1920×1080), can generate video with temporal consistency (maintaining the same subject over time) that lasts up to 60 seconds, and appears to follow text prompts with a great deal of fidelity. So, how did OpenAI pull it off?

OpenAI doesn’t usually share insider technical details with the press, so we’re left to speculate based on theories from experts and information given to the public.

OpenAI says that Sora is a diffusion model, much like DALL-E 3 and Stable Diffusion. It generates a video by starting off with noise and “gradually transforms it by removing the noise over many steps,” the company explains. It “recognizes” objects and concepts listed in the written prompt and pulls them out of the noise, so to speak, until a coherent series of video frames emerge.

Sora is capable of generating videos all at once from a text prompt, extending existing videos, or generating videos from still images. It achieves temporal consistency by giving the model “foresight” of many frames at once, as OpenAI calls it, solving the problem of ensuring a generated subject remains the same even if it falls out of view temporarily.

OpenAI represents video as collections of smaller groups of data called “patches,” which the company says are similar to tokens (fragments of a word) in GPT-4. “By unifying how we represent data, we can train diffusion transformers on a wider range of visual data than was possible before, spanning different durations, resolutions, and aspect ratios,” the company writes.

An important tool in OpenAI’s bag of tricks is that its use of AI models is compounding. Earlier models are helping to create more complex ones. Sora follows prompts well because, like DALL-E 3, it utilizes synthetic captions that describe scenes in the training data generated by another AI model like GPT-4V. And the company is not stopping here. “Sora serves as a foundation for models that can understand and simulate the real world,” OpenAI writes, “a capability we believe will be an important milestone for achieving AGI.”

One question on many people’s minds is what data OpenAI used to train Sora. OpenAI has not revealed its dataset, but based on what people are seeing in the results, it’s possible OpenAI is using synthetic video data generated in a video game engine in addition to sources of real video (say, scraped from YouTube or licensed from stock video libraries). Nvidia’s Dr. Jim Fan, who is a specialist in training AI with synthetic data, wrote on X, “I won’t be surprised if Sora is trained on lots of synthetic data using Unreal Engine 5. It has to be!” Until confirmed by OpenAI, however, that’s just speculation.

OpenAI collapses media reality with Sora, a photorealistic AI video generator Read More »

vmware-admits-sweeping-broadcom-changes-are-worrying-customers

VMware admits sweeping Broadcom changes are worrying customers

Biz & IT, Broadcom, vmware / /

Sorry, not sorry —

Broadcom has cut VMware products, perpetual licenses, and its partner program.

–

The logo of American cloud computing and virtualization technology company VMware is seen at the Mobile World Congress (MWC), the telecom industry's biggest annual gathering, in Barcelona on March 2, 2023.

Broadcom has made a lot of changes to VMware since closing its acquisition of the company in November. On Wednesday, VMware admitted that these changes are worrying customers. With customers mulling alternatives and partners complaining, VMware is trying to do damage control and convince people that change is good.

Not surprisingly, the plea comes from a VMware marketing executive: Prashanth Shenoy, VP of product and technical marketing for the Cloud, Infrastructure, Platforms, and Solutions group at VMware. In Wednesday’s announcement, Shenoy admitted that VMware “has been all about change” since being swooped up for $61 billion. This has resulted in “many questions and concerns” as customers “evaluate how to maximize value from” VMware products.

Among these changes is VMware ending perpetual license sales in favor of a subscription-based business model. VMware had a history of relying on perpetual licensing; VMware called the model its “most renowned” a year ago.

Shenoy’s blog sought to provide reasoning for the change, with the executive writing that “all major enterprise software providers are on [subscription models] today.”

However, the idea that ‘”everyone’s doing it” has done little to ameliorate impacted customers who prefer paying for something once and owning it indefinitely (while paying for associated support costs). Customers are also dealing with budget concerns with already paid-for licenses set to lose support and the only alternative being a monthly fee.

Shenoy’s blog, though, focused on license portability. “This means you will be able to deploy on-premises and then take your subscription at any time to a supported Hyperscaler or VMware Cloud Services Provider environment as desired. You retain your license subscription as you move,” Shenoy wrote, noting new Google Cloud VMware Engine license portability support for VMware Cloud Foundation.

Further, Shenoy claimed the discontinuation of VMware products so that Broadcom could focus on VMware Cloud Foundation and vSphere Foundation would be beneficial, because “offering a few offerings that are lower in price on the high end and are packed with more value for the same or less cost on the lower end makes business sense for customers, partners, and VMware.”

This week, Broadcom axed the free version of vSphere Hypervisor, ESXi. As reported by my colleague Andrew Cunningham, the offering was useful for enthusiasts “who wanted to run multipurpose home servers or to split a system’s time between Windows and one or more Linux distributions without the headaches of dual booting” or who wanted to familiarize themselves with vSphere Hypervisor without having to pay for licensing. The removal of ESXi could contribute to an eventual VMware skills gap, ServeTheHome suggested.

Broadcom addresses VMware partner changes

Broadcom has also announced that it’s ending the VMware partner program. Broadcom initially said it would invite a select number of VMware channel partners to the Broadcom partner program but didn’t say how many, causing concerns about how smaller businesses would get access to VMware products.

Broadcom said it ultimately invited 18,000 VMware resellers to its partner program and said this included “all active” partners, as defined by partners who had active contracts within the last two years. However, 18,000 is fewer than the 28,000 partners VMware told ChannelE2E it had in March 2023. Broadcom didn’t respond to CRN’s questions asking about the discrepancy in numbers and hasn’t responded to questions that Ars Technica previously sent about how it was deciding which VMware partners it would invite to its program.

There are still concerns that channel partners won’t be able to meet Broadcom’s new requirements for being a VMware reseller, meaning that smaller companies may have to consider notable infrastructure changes and moving off VMware. Broadcom’s layoffs of thousands of VMware employees has reportedly hurt communication and contributed to confusion, too.

VMware’s Wednesday post also addressed Broadcom taking VMware’s biggest customers direct, removing channel partners from the equation:

It makes business sense for Broadcom to have close relationships with its most strategic VMware customers to make sure VMware Cloud Foundation is being adopted, used, and providing customer value. However, we expect there will be a role change in accounts that will have to be worked through so that both Broadcom and our partners are providing the most value and greatest impact to strategic customers. And, partners will play a critical role in adding value beyond what Broadcom may be able.

But while taking over VMware’s biggest accounts (CRN estimated in January that this affects about 2,000 accounts) may make business sense for Broadcom, it’s hard to imagine how it would make business sense for the IT businesses managing those accounts previously.

While Broadcom has made headlines with its dramatic changes to VMware, Shenoy argued that “Broadcom identified things that needed to change and, as a responsible company, made the changes quickly and decisively.”

“The changes that have taken place over the past 60+ days were absolutely necessary,” he added.

The implications of these changes will continue to be debated over the coming months as the impact of Broadcom’s strategy is realized. But in the meantime, it looks like Broadcom is sticking to its guns, even with rivals looking to capitalize on related uncertainty.

VMware admits sweeping Broadcom changes are worrying customers Read More »

google-upstages-itself-with-gemini-15-ai-launch,-one-week-after-ultra-1.0

Google upstages itself with Gemini 1.5 AI launch, one week after Ultra 1.0

AI, Biz & IT, chatgpt, chatgtp, Google, google ai, Google Gemini, large language model, llm, Pro 1.5, Ultra 1.0 / /

Gemini’s Twin —

Google confusingly overshadows its own pro product a week after its last major AI launch.

–

The Gemini 1.5 logo

Enlarge / The Gemini 1.5 logo, released by Google.

Google

One week after its last major AI announcement, Google appears to have upstaged itself. Last Thursday, Google launched Gemini Ultra 1.0, which supposedly represented the best AI language model Google could muster—available as part of the renamed “Gemini” AI assistant (formerly Bard). Today, Google announced Gemini Pro 1.5, which it says “achieves comparable quality to 1.0 Ultra, while using less compute.”

Congratulations, Google, you’ve done it. You’ve undercut your own premiere AI product. While Ultra 1.0 is possibly still better than Pro 1.5 (what even are we saying here), Ultra was presented as a key selling point of its “Gemini Advanced” tier of its Google One subscription service. And now it’s looking a lot less advanced than seven days ago. All this is on top of the confusing name-shuffling Google has been doing recently. (Just to be clear—although it’s not really clarifying at all—the free version of Bard/Gemini currently uses the Pro 1.0 model. Got it?)

Google claims that Gemini 1.5 represents a new generation of LLMs that “delivers a breakthrough in long-context understanding,” and that it can process up to 1 million tokens, “achieving the longest context window of any large-scale foundation model yet.” Tokens are fragments of a word. The first part of the claim about “understanding” is contentious and subjective, but the second part is probably correct. OpenAI’s GPT-4 Turbo can reportedly handle 128,000 tokens in some circumstances, and 1 million is quite a bit more—about 700,000 words. A larger context window allows for processing longer documents and having longer conversations. (The Gemini 1.0 model family handles 32,000 tokens max.)

But any technical breakthroughs are almost beside the point. What should we make of a company that just trumpeted to the world about its AI supremacy last week, only to partially supersede that a week later? Is it a testament to the rapid rate of AI technical progress in Google’s labs, a sign that red tape was holding back Ultra 1.0 for too long, or merely a sign of poor coordination between research and marketing? We honestly don’t know.

So back to Gemini 1.5. What is it, really, and how will it be available? Google implies that like 1.0 (which had Nano, Pro, and Ultra flavors), it will be available in multiple sizes. Right now, Pro 1.5 is the only model Google is unveiling. Google says that 1.5 uses a new mixture-of-experts (MoE) architecture, which means the system selectively activates different “experts” or specialized sub-models within a larger neural network for specific tasks based on the input data.

Google says that Gemini 1.5 can perform “complex reasoning about vast amounts of information,” and gives an example of analyzing a 402-page transcript of Apollo 11’s mission to the Moon. It’s impressive to process documents that large, but the model, like every large language model, is highly likely to confabulate interpretations across large contexts. We wouldn’t trust it to soundly analyze 1 million tokens without mistakes, so that’s putting a lot of faith into poorly understood LLM hands.

For those interested in diving into technical details, Google has released a technical report on Gemini 1.5 that appears to show Gemini performing favorably versus GPT-4 Turbo on various tasks, but it’s also important to note that the selection and interpretation of those benchmarks can be subjective. The report does give some numbers on how much better 1.5 is compared to 1.0, saying it’s 28.9 percent better than 1.0 Pro at “Math, Science & Reasoning” and 5.2 percent better at those subjects than 1.0 Ultra.

A table from the Gemini 1.5 technical document showing comparisons to Gemini 1.0.

Enlarge / A table from the Gemini 1.5 technical document showing comparisons to Gemini 1.0.

Google

But for now, we’re still kind of shocked that Google would launch this particular model at this particular moment in time. Is it trying to get ahead of something that it knows might be just around the corner, like OpenAI’s unreleased GPT-5, for instance? We’ll keep digging and let you know what we find.

Google says that a limited preview of 1.5 Pro is available now for developers via AI Studio and Vertex AI with a 128,000 token context window, scaling up to 1 million tokens later. Gemini 1.5 apparently has not come to the Gemini chatbot (formerly Bard) yet.

Google upstages itself with Gemini 1.5 AI launch, one week after Ultra 1.0 Read More »

us-says-ai-models-can’t-hold-patents

US says AI models can’t hold patents

AI, AI ethics, Biz & IT, copyright, machine learning, patents, US Patent and Trademark Office, USPTO / /

Robot inventors dismayed —

Inventors must be human, but there’s still a condition where AI can officially help.

–

An illustrated concept of a digital brain, crossed out.

On Tuesday, the United States Patent and Trademark Office (USPTO) published guidance on inventorship for AI-assisted inventions, clarifying that while AI systems can play a role in the creative process, only natural persons (human beings) who make significant contributions to the conception of an invention can be named as inventors. It also rules out using AI models to churn out patent ideas without significant human input.

The USPTO says this position is supported by “the statutes, court decisions, and numerous policy considerations,” including the Executive Order on AI issued by President Biden. We’ve previously covered attempts, which have been repeatedly rejected by US courts, by Dr. Stephen Thaler to have an AI program called “DABUS” named as the inventor on a US patent (a process begun in 2019).

This guidance follows themes previously set by the US Copyright Office (and agreed upon by a judge) that an AI model cannot own a copyright for a piece of media and that substantial human contributions are required for copyright protection.

Even though an AI model itself cannot be named an inventor or joint inventor on a patent, using AI assistance to create an invention does not necessarily disqualify a human from holding a patent, as the USPTO explains:

“While AI systems and other non-natural persons cannot be listed as inventors on patent applications or patents, the use of an AI system by a natural person(s) does not preclude a natural person(s) from qualifying as an inventor (or joint inventors) if the natural person(s) significantly contributed to the claimed invention.”

However, the USPTO says that significant human input is required for an invention to be patentable: “Maintaining ‘intellectual domination’ over an AI system does not, on its own, make a person an inventor of any inventions created through the use of the AI system.” So a person simply overseeing an AI system isn’t suddenly an inventor. The person must make a significant contribution to the conception of the invention.

If someone does use an AI model to help create patents, the guidance describes how the application process would work. First, patent applications for AI-assisted inventions must name “the natural person(s) who significantly contributed to the invention as the inventor,” and additionally, applications must not list “any entity that is not a natural person as an inventor or joint inventor, even if an AI system may have been instrumental in the creation of the claimed invention.”

Reading between the lines, it seems the contributions made by AI systems are akin to contributions made by other tools that assist in the invention process. The document does not explicitly say that the use of AI is required to be disclosed during the application process.

Even with the published guidance, the USPTO is seeking public comment on the newly released guidelines and issues related to AI inventorship on its website.

US says AI models can’t hold patents Read More »

broadcom-owned-vmware-kills-the-free-version-of-esxi-virtualization-software

Broadcom-owned VMware kills the free version of ESXi virtualization software

Biz & IT, Broadcom, esxi, hypervisor, Tech, virtualization, VM, vmware / /

freesphere —

Software’s free version was a good fit for tinkerers and hobbyists.

–

Broadcom-owned VMware kills the free version of ESXi virtualization software

VMware

Since Broadcom’s $61 billion acquisition of VMware closed in November 2023, Broadcom has been charging ahead with major changes to the company’s personnel and products. In December, Broadcom began laying off thousands of employees and stopped selling perpetually licensed versions of VMware products, pushing its customers toward more stable and lucrative software subscriptions instead. In January, it ended its partner programs, potentially disrupting sales and service for many users of its products.

This week, Broadcom is making a change that is smaller in scale but possibly more relevant for home users of its products: The free version of VMware’s vSphere Hypervisor, also known as ESXi, is being discontinued.

ESXi is what is known as a “bare-metal hypervisor,” lightweight software that runs directly on hardware without requiring a separate operating system layer in between. ESXi allows you to split a PC’s physical resources (CPUs and CPU cores, RAM, storage, networking components, and so on) among multiple virtual machines. ESXi also supports passthrough for PCI, SATA, and USB accessories, allowing guest operating systems direct access to components like graphics cards and hard drives.

The free version of ESXi had limits compared to the full, paid enterprise versions—it could only support up to two physical CPUs, didn’t come with any software support, and lacked automated load-balancing and management features. But it was still useful for enthusiasts and home users who wanted to run multipurpose home servers or to split a system’s time between Windows and one or more Linux distributions without the headaches of dual booting. It was also a useful tool for people who used the enterprise versions of the vSphere Hypervisor but wanted to test the software or learn its ins and outs without dealing with paid licensing.

For the latter group, a 60-day trial of the VMware vSphere 8 software is still available. Tinkerers will be better off trying to migrate to an alternative product instead, like Proxmox, XCP-ng, or even the Hyper-V capabilities built into the Pro versions of Windows 10 and 11.

Broadcom-owned VMware kills the free version of ESXi virtualization software Read More »

openai-experiments-with-giving-chatgpt-a-long-term-conversation-memory

OpenAI experiments with giving ChatGPT a long-term conversation memory

AI, AI memory, Biz & IT, chatbot, chatbots, chatgpt, chatgtp, GPT-4, GPT-4-turbo, large language models, machine learning / /

“I remember…the Alamo” —

AI chatbot “memory” will recall facts from previous conversations when enabled.

–

A pixelated green illustration of a pair of hands looking through file records.

Enlarge / When ChatGPT looks things up, a pair of green pixelated hands look through paper records, much like this. Just kidding.

Benj Edwards / Getty Images

On Tuesday, OpenAI announced that it is experimenting with adding a form of long-term memory to ChatGPT that will allow it to remember details between conversations. You can ask ChatGPT to remember something, see what it remembers, and ask it to forget. Currently, it’s only available to a small number of ChatGPT users for testing.

So far, large language models have typically used two types of memory: one baked into the AI model during the training process (before deployment) and an in-context memory (the conversation history) that persists for the duration of your session. Usually, ChatGPT forgets what you have told it during a conversation once you start a new session.

Various projects have experimented with giving LLMs a memory that persists beyond a context window. (The context window is the hard limit on the number of tokens the LLM can process at once.) The techniques include dynamically managing context history, compressing previous history through summarization, links to vector databases that store information externally, or simply periodically injecting information into a system prompt (the instructions ChatGPT receives at the beginning of every chat).

A screenshot of ChatGPT memory controls provided by OpenAI.

Enlarge / A screenshot of ChatGPT memory controls provided by OpenAI.

OpenAI

OpenAI hasn’t explained which technique it uses here, but the implementation reminds us of Custom Instructions, a feature OpenAI introduced in July 2023 that lets users add custom additions to the ChatGPT system prompt to change its behavior.

Possible applications for the memory feature provided by OpenAI include explaining how you prefer your meeting notes to be formatted, telling it you run a coffee shop and having ChatGPT assume that’s what you’re talking about, keeping information about your toddler that loves jellyfish so it can generate relevant graphics, and remembering preferences for kindergarten lesson plan designs.

Also, OpenAI says that memories may help ChatGPT Enterprise and Team subscribers work together better since shared team memories could remember specific document formatting preferences or which programming frameworks your team uses. And OpenAI plans to bring memories to GPTs soon, with each GPT having its own siloed memory capabilities.

Memory control

Obviously, any tendency to remember information brings privacy implications. You should already know that sending information to OpenAI for processing on remote servers introduces the possibility of privacy leaks and that OpenAI trains AI models on user-provided information by default unless conversation history is disabled or you’re using an Enterprise or Team account.

Along those lines, OpenAI says that your saved memories are also subject to OpenAI training use unless you meet the criteria listed above. Still, the memory feature can be turned off completely. Additionally, the company says, “We’re taking steps to assess and mitigate biases, and steer ChatGPT away from proactively remembering sensitive information, like your health details—unless you explicitly ask it to.”

Users will also be able to control what ChatGPT remembers using a “Manage Memory” interface that lists memory items. “ChatGPT’s memories evolve with your interactions and aren’t linked to specific conversations,” OpenAI says. “Deleting a chat doesn’t erase its memories; you must delete the memory itself.”

ChatGPT’s memory features are not currently available to every ChatGPT account, so we have not experimented with it yet. Access during this testing period appears to be random among ChatGPT (free and paid) accounts for now. “We are rolling out to a small portion of ChatGPT free and Plus users this week to learn how useful it is,” OpenAI writes. “We will share plans for broader roll out soon.”

OpenAI experiments with giving ChatGPT a long-term conversation memory Read More »

the-super-bowl’s-best-and-wackiest-ai-commercials

The Super Bowl’s best and wackiest AI commercials

AI, american football, Anthropic, Biz & IT, Commercial, commercials, Football, Google, large language models, machine learning, microsoft, Super Bowl / /

Superb Owl News —

It’s nothing like “crypto bowl” in 2022, but AI made a notable splash during the big game.

–

A still image from BodyArmor's 2024

Enlarge / A still image from BodyArmor’s 2024 “Field of Fake” Super Bowl commercial.

BodyArmor

Heavily hyped tech products have a history of appearing in Super Bowl commercials during football’s biggest game—including the Apple Macintosh in 1984, dot-com companies in 2000, and cryptocurrency firms in 2022. In 2024, the hot tech in town is artificial intelligence, and several companies showed AI-related ads at Super Bowl LVIII. Here’s a rundown of notable appearances that range from serious to wacky.

Microsoft Copilot

Microsoft Game Day Commercial | Copilot: Your everyday AI companion.

It’s been a year since Microsoft launched the AI assistant Microsoft Copilot (as “Bing Chat“), and Microsoft is leaning heavily into its AI-assistant technology, which is powered by large language models from OpenAI. In Copilot’s first-ever Super Bowl commercial, we see scenes of various people with defiant text overlaid on the screen: “They say I will never open my own business or get my degree. They say I will never make my movie or build something. They say I’m too old to learn something new. Too young to change the world. But I say watch me.”

Then the commercial shows Copilot creating solutions to some of these problems, with prompts like, “Generate storyboard images for the dragon scene in my script,” “Write code for my 3d open world game,” “Quiz me in organic chemistry,” and “Design a sign for my classic truck repair garage Mike’s.”

Of course, since generative AI is an unfinished technology, many of these solutions are more aspirational than practical at the moment. On Bluesky, writer Ed Zitron put Microsoft’s truck repair logo to the test and saw results that weren’t nearly as polished as those seen in the commercial. On X, others have criticized and poked fun at the “3d open world game” generation prompt, which is a complex task that would take far more than a single, simple prompt to produce useful code.

Google Pixel 8 “Guided Frame” feature

Javier in Frame | Google Pixel SB Commercial 2024.

Instead of focusing on generative aspects of AI, Google’s commercial showed off a feature called “Guided Frame” on the Pixel 8 phone that uses machine vision technology and a computer voice to help people with blindness or low vision to take photos by centering the frame on a face or multiple faces. Guided Frame debuted in 2022 in conjunction with the Google Pixel 7.

The commercial tells the story of a person named Javier, who says, “For many people with blindness or low vision, there hasn’t always been an easy way to capture daily life.” We see a simulated blurry first-person view of Javier holding a smartphone and hear a computer-synthesized voice describing what the AI model sees, directing the person to center on a face to snap various photos and selfies.

Considering the controversies that generative AI currently generates (pun intended), it’s refreshing to see a positive application of AI technology used as an accessibility feature. Relatedly, an app called Be My Eyes (powered by OpenAI’s GPT-4V) also aims to help low-vision people interact with the world.

Despicable Me 4

Despicable Me 4 – Minion Intelligence (Big Game Spot).

So far, we’ve covered a couple attempts to show AI-powered products as positive features. Elsewhere in Super Bowl ads, companies weren’t as generous about the technology. In an ad for the film Despicable Me 4, we see two Minions creating a series of terribly disfigured AI-generated still images reminiscent of Stable Diffusion 1.4 from 2022. There’s three-legged people doing yoga, a painting of Steve Carell and Will Ferrell as Elizabethan gentlemen, a handshake with too many fingers, people eating spaghetti in a weird way, and a pair of people riding dachshunds in a race.

The images are paired with an earnest voiceover that says, “Artificial intelligence is changing the way we see the world, showing us what we never thought possible, transforming the way we do business, and bringing family and friends closer together. With artificial intelligence, the future is in good hands.” When the voiceover ends, the camera pans out to show hundreds of Minions generating similarly twisted images on computers.

Speaking of image synthesis at the Super Bowl, people mistook a Christian commercial created by He Gets Us, LLC as having been AI-generated, likely due to its gaudy technicolor visuals. With the benefit of a YouTube replay and the ability to look at details, the “He washed feet” commercial doesn’t appear AI-generated to us, but it goes to show how the concept of image synthesis has begun to cast doubt on human-made creations.

The Super Bowl’s best and wackiest AI commercials Read More »

canada-declares-flipper-zero-public-enemy-no.-1-in-car-theft-crackdown

Canada declares Flipper Zero public enemy No. 1 in car-theft crackdown

Biz & IT, Canada, car theft, Cars, flipper zero, hacking, Policy, radio, Security / /

FLIPPING YOUR LID —

How do you ban a device built with open source hardware and software anyway?

–

A Flipper Zero device

Enlarge / A Flipper Zero device

https://flipperzero.one/

Canadian Prime Minister Justin Trudeau has identified an unlikely public enemy No. 1 in his new crackdown on car theft: the Flipper Zero, a $200 piece of open source hardware used to capture, analyze and interact with simple radio communications.

On Thursday, the Innovation, Science and Economic Development Canada agency said it will “pursue all avenues to ban devices used to steal vehicles by copying the wireless signals for remote keyless entry, such as the Flipper Zero, which would allow for the removal of those devices from the Canadian marketplace through collaboration with law enforcement agencies.” A social media post by François-Philippe Champagne, the minister of that agency, said that as part of the push “we are banning the importation, sale and use of consumer hacking devices, like flippers, used to commit these crimes.”

In remarks made the same day, Trudeau said the push will target similar tools that he said can be used to defeat anti-theft protections built into virtually all new cars.

“In reality, it has become too easy for criminals to obtain sophisticated electronic devices that make their jobs easier,” he said. “For example, to copy car keys. It is unacceptable that it is possible to buy tools that help car theft on major online shopping platforms.”

Presumably, such tools subject to the ban would include HackRF One and LimeSDR, which have become crucial for analyzing and testing the security of all kinds of electronic devices to find vulnerabilities before they’re exploited. None of the government officials identified any of these tools, but in an email, a representative of the Canadian government reiterated the use of the phrase “pursuing all avenues to ban devices used to steal vehicles by copying the wireless signals for remote keyless entry.”

A humble hobbyist device

The push to ban any of these tools has been met with fierce criticism from hobbyists and security professionals. Their case has only been strengthened by Trudeau’s focus on Flipper Zero. This slim, lightweight device bearing the logo of an adorable dolphin acts as a Swiss Army knife for sending, receiving, and analyzing all kinds of wireless communications. It can interact with radio signals, including RFID, NFC, Bluetooth, Wi-Fi, or standard radio. People can use them to change the channels of a TV at a bar covertly, clone simple hotel key cards, read the RFID chip implanted in pets, open and close some garage doors, and, until Apple issued a patch, send iPhones into a never-ending DoS loop.

The price and ease of use make Flipper Zero ideal for beginners and hobbyists who want to understand how increasingly ubiquitous communications protocols such as NFC and Wi-Fi work. It bundles various open source hardware and software into a portable form factor that sells for an affordable price. Lost on the Canadian government, the device isn’t especially useful in stealing cars because it lacks the more advanced capabilities required to bypass anti-theft protections introduced in more than two decades.

One thing the Flipper Zero is exceedingly ill-equipped for is defeating modern antihack protections built into cars, smartcards, phones, and other electronic devices.

The most prevalent form of electronics-assisted car theft these days, for instance, uses what are known as signal amplification relay devices against keyless ignition and entry systems. This form of hack works by holding one device near a key fob and a second device near the vehicle the fob works with. In the most typical scenario, the fob is located on a shelf near a locked front door, and the car is several dozen feet away in a driveway. By placing one device near the front door and another one next to the car, the hack beams the radio signals necessary to unlock and start the device.

Canada declares Flipper Zero public enemy No. 1 in car-theft crackdown Read More »

london-underground-is-testing-real-time-ai-surveillance-tools-to-spot-crime

London Underground is testing real-time AI surveillance tools to spot crime

AI, AI surveillance, Biz & IT, Freedom of information act, london underground, Security, syndication / /

tube tracking —

Computer vision system tried to detect crime, weapons, people falling, and fare dodgers.

–

Commuters wait on the platform as a Central Line tube train arrives at Liverpool Street London Transport Tube Station in 2023.

Thousands of people using the London Underground had their movements, behavior, and body language watched by AI surveillance software designed to see if they were committing crimes or were in unsafe situations, new documents obtained by WIRED reveal. The machine-learning software was combined with live CCTV footage to try to detect aggressive behavior and guns or knives being brandished, as well as looking for people falling onto Tube tracks or dodging fares.

From October 2022 until the end of September 2023, Transport for London (TfL), which operates the city’s Tube and bus network, tested 11 algorithms to monitor people passing through Willesden Green Tube station, in the northwest of the city. The proof of concept trial is the first time the transport body has combined AI and live video footage to generate alerts that are sent to frontline staff. More than 44,000 alerts were issued during the test, with 19,000 being delivered to station staff in real time.

Documents sent to WIRED in response to a Freedom of Information Act request detail how TfL used a wide range of computer vision algorithms to track people’s behavior while they were at the station. It is the first time the full details of the trial have been reported, and it follows TfL saying, in December, that it will expand its use of AI to detect fare dodging to more stations across the British capital.

In the trial at Willesden Green—a station that had 25,000 visitors per day before the COVID-19 pandemic—the AI system was set up to detect potential safety incidents to allow staff to help people in need, but it also targeted criminal and antisocial behavior. Three documents provided to WIRED detail how AI models were used to detect wheelchairs, prams, vaping, people accessing unauthorized areas, or putting themselves in danger by getting close to the edge of the train platforms.

The documents, which are partially redacted, also show how the AI made errors during the trial, such as flagging children who were following their parents through ticket barriers as potential fare dodgers, or not being able to tell the difference between a folding bike and a non-folding bike. Police officers also assisted the trial by holding a machete and a gun in the view of CCTV cameras, while the station was closed, to help the system better detect weapons.

Privacy experts who reviewed the documents question the accuracy of object detection algorithms. They also say it is not clear how many people knew about the trial, and warn that such surveillance systems could easily be expanded in the future to include more sophisticated detection systems or face recognition software that attempts to identify specific individuals. “While this trial did not involve facial recognition, the use of AI in a public space to identify behaviors, analyze body language, and infer protected characteristics raises many of the same scientific, ethical, legal, and societal questions raised by facial recognition technologies,” says Michael Birtwistle, associate director at the independent research institute the Ada Lovelace Institute.

In response to WIRED’s Freedom of Information request, the TfL says it used existing CCTV images, AI algorithms, and “numerous detection models” to detect patterns of behavior. “By providing station staff with insights and notifications on customer movement and behaviour they will hopefully be able to respond to any situations more quickly,” the response says. It also says the trial has provided insight into fare evasion that will “assist us in our future approaches and interventions,” and the data gathered is in line with its data policies.

In a statement sent after publication of this article, Mandy McGregor, TfL’s head of policy and community safety, says the trial results are continuing to be analyzed and adds, “there was no evidence of bias” in the data collected from the trial. During the trial, McGregor says, there were no signs in place at the station that mentioned the tests of AI surveillance tools.

“We are currently considering the design and scope of a second phase of the trial. No other decisions have been taken about expanding the use of this technology, either to further stations or adding capability.” McGregor says. “Any wider roll out of the technology beyond a pilot would be dependent on a full consultation with local communities and other relevant stakeholders, including experts in the field.”

London Underground is testing real-time AI surveillance tools to spot crime Read More »

report:-sam-altman-seeking-trillions-for-ai-chip-fabrication-from-uae,-others

Report: Sam Altman seeking trillions for AI chip fabrication from UAE, others

AI, AI chips, Biz & IT, Google Gemini, large language models, machine learning, microsoft, Microsoft Copilot, openai, sam altman, semiconductor manufacturing, text synthesis, UAE / /

chips ahoy —

WSJ: Audacious $5-$7 trillion investment would aim to expand global AI chip supply.

–

WASHINGTON, DC - JANUARY 11: OpenAI Chief Executive Officer Sam Altman walks on the House side of the U.S. Capitol on January 11, 2024 in Washington, DC. Meanwhile, House Freedom Caucus members who left a meeting in the Speakers office say that they were talking to the Speaker about abandoning the spending agreement that Johnson announced earlier in the week. (Photo by Kent Nishimura/Getty Images)

Enlarge / OpenAI Chief Executive Officer Sam Altman walks on the House side of the US Capitol on January 11, 2024, in Washington, DC. (Photo by Kent Nishimura/Getty Images)

Getty Images

On Thursday, The Wall Street Journal reported that OpenAI CEO Sam Altman is in talks with investors to raise as much as $5 trillion to $7 trillion for AI chip manufacturing, according to people familiar with the matter. The funding seeks to address the scarcity of graphics processing units (GPUs) crucial for training and running large language models like those that power ChatGPT, Microsoft Copilot, and Google Gemini.

The high dollar amount reflects the huge amount of capital necessary to spin up new semiconductor manufacturing capability. “As part of the talks, Altman is pitching a partnership between OpenAI, various investors, chip makers and power providers, which together would put up money to build chip foundries that would then be run by existing chip makers,” writes the Wall Street Journal in its report. “OpenAI would agree to be a significant customer of the new factories.”

To hit these ambitious targets—which are larger than the entire semiconductor industry’s current $527 billion global sales combined—Altman has reportedly met with a range of potential investors worldwide, including sovereign wealth funds and government entities, notably the United Arab Emirates, SoftBank CEO Masayoshi Son, and representatives from Taiwan Semiconductor Manufacturing Co. (TSMC).

TSMC is the world’s largest dedicated independent semiconductor foundry. It’s a critical linchpin that companies such as Nvidia, Apple, Intel, and AMD rely on to fabricate SoCs, CPUs, and GPUs for various applications.

Altman reportedly seeks to expand the global capacity for semiconductor manufacturing significantly, funding the infrastructure necessary to support the growing demand for GPUs and other AI-specific chips. GPUs are excellent at parallel computation, which makes them ideal for running AI models that heavily rely on matrix multiplication to work. However, the technology sector currently faces a significant shortage of these important components, constraining the potential for AI advancements and applications.

In particular, the UAE’s involvement, led by Sheikh Tahnoun bin Zayed al Nahyan, a key security official and chair of numerous Abu Dhabi sovereign wealth vehicles, reflects global interest in AI’s potential and the strategic importance of semiconductor manufacturing. However, the prospect of substantial UAE investment in a key tech industry raises potential geopolitical concerns, particularly regarding the US government’s strategic priorities in semiconductor production and AI development.

The US has been cautious about allowing foreign control over the supply of microchips, given their importance to the digital economy and national security. Reflecting this, the Biden administration has undertaken efforts to bolster domestic chip manufacturing through subsidies and regulatory scrutiny of foreign investments in important technologies.

To put the $5 trillion to $7 trillion estimate in perspective, the White House just today announced a $5 billion investment in R&D to advance US-made semiconductor technologies. TSMC has already sunk $40 billion—one of the largest foreign investments in US history—into a US chip plant in Arizona. As of now, it’s unclear whether Altman has secured any commitments toward his fundraising goal.

Updated on February 9, 2024 at 8: 45 PM Eastern with a quote from the WSJ that clarifies the proposed relationship between OpenAI and partners in the talks.

Report: Sam Altman seeking trillions for AI chip fabrication from UAE, others Read More »