The Cerebras Situationship
Andrew CurranReally clears things up Mr A
Read discussion →Discourse
First-hand perspectives on technology, companies, and building what comes next.
Muse Gadgets
Ryan HooverThis is rad. Makes me want to build hardware
Read discussion →TypeSafe CEO says Jev is used by ~25% of Fortune 500
Zach DavidsonDiogo Almeida says they're serving a trillion tokens per day as of last week. The model is just 3 weeks old and is already spawning copycats. Curious to hear if anyone here is using Jev in production / what you're using it for?
Read discussion →Will transformers scale to AGI or do we need a different architecture?
LavanyaI’m curious if people think transformers will be the architecture AGI is born out of. People seem to fall into 3 buckets: 1. Scaling the existing recipe is enough and more parameters, data and compute will get us there. Scaling laws predict improvements, but are they enough for general intelligence? 2. The architecture is sufficient, but the learning recipe needs to evolve (more RL, interaction with environments, more inference time compute, multimodal learning cross text, images, video and audio). 3. We need a totally different architecture or representation of the world (e.g. changes to memory, learning mechanisms, or tokenizing in 4D/3D vs 2D might be needed for spatial intelligence). What do folks here think?
Read discussion →Making Discourse Better
Erik TorenbergWhat feedback for us? How can we make this site better -- both Discourse and Cosign?
Read discussion →When you evaluate AI at work, do you check what gets lost between the outputs?
Anna BelovaIn my customer discovery interviews for OpenWay AI, I’ve often asked what people expect AI to do for their business. One question I keep coming back to: are we using AI to help knowledge move through a company, or adding more places for it to get lost? We’d expect AI to help. Yet every summary and handoff creates another point where someone, or something, decides which details survive. Imagine a buyer finally says yes after sales explains the rollout. AI records “implementation discussed.” Marketing uses another AI tool to draft the next campaign from an existing brief, which still assumes price is the main objection.. We already had to manage what gets lost between people. Now we also have to manage what gets lost when AI compresses a conversation, decides what matters, or works from an incomplete brief. When you evaluate AI at work, do you check what gets lost between the outputs?
Read discussion →What Happens When Nothing Depends on Us?
Bilal KhanImagine AGI is achieved and alignment is solved. AI can now do almost everything. Only a handful of tasks still require humans, and very few people get to do them. What does everyone else do? Maybe you can play games, travel, climb mountains, or spend time on hobbies. But none of it is needed. Whether you succeed or fail changes very little. For most of history, people have had something to work toward because their effort mattered. They could build something, discover something, solve a problem, or become useful to others. What happens when there is almost nothing left for humans to contribute? What do you strive for when nothing important depends on you anymore?
Read discussion →Thoughts on Griffin
anthonyWould love to get thoughts on Griffin. Both cool and terrifying.
Read discussion →Looking for tips
Adam KillamAnyone have advice on how to build multi-step, agentic workflows using Claude Code or Codex? I'm a non-dev and I find coding harnesses great at building any single slice of a given piece of software but generally less good at building multi-step workflows where a piece of work output is created that then needs to be acted on by an agent or another part of the system and then handed off to the next agent or next part of the system. Hopefully that makes sense.
Read discussion →Compute buildout could run tens of millions if not billions of agents
From the blog: "AI chips shipped through 2027 could run tens to hundreds of millions of concurrent frontier-model agents. Running nonstop, these agents would supply as many weekly working hours as about 140–720 million full-time employees. More efficient models could potentially support billions of agents on the same hardware. Applying DeepSeek V4 Pro serving benchmarks to the projected hardware supply yields approximately 1.9 billion concurrent agents supplying as many weekly working hours as 8 billion people each working 40 hours."
Read discussion →Who is winning the personal AI assistant race?
Sharath KurugantyDots or instinct or town or grok bot or Muse or something else?
Read discussion →Anthropic's Frontier Safety Roadmap
Christopher WallaceAnthropic set a September 30th deadline, but we've yet to see any announcement of their provable inference prototype. I thought this was interesting because it's a very important part of the AI supply chain that folks aren't tracking, and it could establish responsible training standards. > We will develop a prototype by September 30, 2026 of provable inference, a technique for reliably, provably “signing” AI model outputs in a way that makes them attributable to a specific set of model weights. In the future, it’s possible that very sophisticated attackers will seek to infiltrate our systems and modify our models after we’ve trained them - whether to sabotage our work or co-opt our models into serving their own goals. If we could reliably and systematically verify that model outputs were coming from a specific set of model weights, we believe this threat would be significantly reduced.
Read discussion →on ai psychosis:
Harias i watch engineers fall deeper into the belief that an amalgamation of mathematical probabilities somehow understands their codebase better than they do, i find myself thinking back to a time when software wasn’t built for hypergrowth, but simply to do x without inventing y. ai seems almost fundamentally opposed to this philosophy. ask it to do x and it will eagerly invent y and z before it has even tried to understand x. the danger isn’t that ai writes worse code- it’s that it makes writing unnecessary code 100% free and the human condition is such that some of us will always prefer the fast, steep gains of ai, even when it does a bajillion unrelated things to accomplish something that could have been done without changing anything else. so, somewhat paradoxically, the quality of software may keep declining for as long as ai keeps getting better. the cheaper complexity becomes to create, the less incentive there is to understand or avoid it.
Read discussion →Part I: Securing Frontier Labs
s1r1usGoing to drop more work of ours on this platform. We secured openai by finding a complex exploit chain. Two bugs let us take over ChatGPT/Codex accounts of OpenAI employees (+some unaffiliated users) and reach connected services: Outlook, Slack, GitHub, etc. We reported and OpenAI fixed it in 14 hours. We will be publishing similar work soon.
Read discussion →How to be a compute capitalist - deciphering the revenue ($/MW) and risks at each layer of the stack - land, shell, racks, GPUs, inference and frontier models...
Anirudh ReddyI got to task one of the world’s most powerful telescopes for a night
Christian KeilWe found asteroids, nebula, and even the satellites I helped build and launch at Astranis. We also technically found a UFO, but it was probably another satellite
Read discussion →What does “college is important” actually mean now?
Gallup reports that 31% of Americans consider college “very important,” down from 70% in 2013, but it doesn’t tell us whether a degree pays off for a particular person or field. Jeffrey Tucker’s take is that college can amount to a long, subsidized detour for people who don’t face demanding post-college credentialing, and while I think that raises a useful question, it risks treating very different paths as one thing: some careers require a degree and further training while for others, the cost and time may be hard to justify. How should we judge college’s value? By career requirements? Earnings? Learning? Something else? I'm excited about HAA as an alternative: https://www.theacademysf.com/
Read discussion →Selling Company Data to the Frontier Labs (or equivalent)
Tod SacerdotiI receive an email or ad about once a week trying to get me to sell my company's data to Micro1, Handshake, Mercor, or some other middleman. I also get the same emails about my portfolio of startups. The pitch is always for millions of dollars of potential value. I am curious if any founders have actually sold their company data and what the process has been like. Would you recommend this for struggling companies that need resources or a broader set of businesses?
Read discussion →Trump expected to pick Jay Clayton as new AI czar
How do we feel about this pick?
Read discussion →Rust in Model Research and Development
AbinashHi everyone, I'm Abinash. For the past few weeks, I have been working on building a neural network in Rust to test the idea of using Rust in model development and research. Rust is being extensively used in LLM infrastructure and is growing in the inference space as we get official libraries to run CUDA kernels from Rust binaries. But if you see the model design, research or development space, Rust is non-existent. So I thought I'd give it a try. I tried to build a fairly easy neural network to solve the MNIST handwritten digit identification problem. I know it's not a huge milestone. But I tried to give it a try to test my idea how easy it is to develop neural networks in Rust. The neural network I'm developing has 2 hidden layers with 512 and 128 neurons, respectively. The goal is to achieve 95% accuracy on my local CPU-only system. I got the input system right; it can now take training and testing datasets and prepare the matrices for the training and testing stages. After that, I plan to start developing a real transformer-based LLM from scratch in Rust. I'm not sure how it will play out, but I will give it a try. I'd love to have your thoughts on it. If you're a senior ML engineer or MTS at a frontier lab, I'd love to have your thoughts on it.
Read discussion →European Commission creates *even worse* alternative to Microsoft Teams
Officials are describing it as "absolute shit" Can you think of anyone worse to build "Teams" than Microsoft? Yes: Europe 😂
Read discussion →Harness Engineering as UX Design
Siqi ChenI'm working on an article on everything we learned to make cfo.ai's Ari agent as capable as we can. What would you be most curious about?
Read discussion →Why Companies Should Own the Weights of Production
Christian CataliniAs AI makes execution increasingly cheap, the ability to properly steer and verify AI outputs better than your competitors will be the differentiator. World-class verification factories rely on two assets: unique ground truth and talent.
Read discussion →You vibe code enough software that you realize you should make a robot. kek.
//KalosAnyone felt that progression yet? -You make 10 products with astra and a couple API keys. - You're euphoric about 2-3 of the products. - You then realize once you put it out there anyone can also replicate it with astra. You have a number of feelings in between that somehow conclude in; You decide you need to instead build a robot in your garage.
Read discussion →What Does 90% Even Mean Anymore?
Serhan YilmazI increasingly don't care about the 2-point benchmark gap between frontier models. I care about their failure shape. Two models can both score 90% and feel completely different to build on. One gets something wrong and you catch it immediately. Another makes one incorrect assumption early, spends the next 20 steps building on top of it, and gives you something coherent enough that it passes a quick skim. Same score but very different effectiveness. This matters a lot more as we give models longer-running tasks. Does the model notice when it's off track? Does the mistake stay contained? Can it recover? And how expensive is it for me to figure out that something went wrong? Benchmark averages still tell us something. But at this point I want to know what the remaining 10% actually looks like.
Read discussion →Trillium Labs Launches Nonprofit for Open Frontier AI Science
They're starting w open post-training recipes and planning open infrastructure research into recursive self-improvement, reward hacking, and multi-agent systems. (inb4 this converts to for profit when successful?)
Read discussion →favorite dictation tool?
i've been using wispr flow mostly but have also trialed aqua / superwhisper / others curious if anyone has a strong opinion as to why they like one or another?
Read discussion →Back to Atari 8bit games dev after 42 years
Marek SpanelWhile we all are waiting four next Arma, here is a game Rio Grande 3D that spiritually started the journey for me, was attempted as first Bohemia game before Flashpoint and Arma, my brother and Bohemia Interactive in the 80s: River Raid. Let me present you a game I wished we could make bacl then but had no idea about he power of mathematics. Rio Grande 3D for Atari 800XL.
Read discussion →Opus vs Astra, write C++, win Starcraft
Daniel HunterAlex Duffy at Good Start Labs I think is still flying under the radar but is a remarkable talent. He lives at the intersection of AI and games. This is livestream on Twitch so sick.
Read discussion →SpaceX just did 3 launches in 13 hours
Trace CohenCrew-13 to the ISS, 130 sats on Transporter-18 (Google's Suncatcher TPU sat was on it), then Falcon Heavy's first ever NRO launch to close it out. One booster was on its 25th flight. Reusable rockets are just normal now
Read discussion →Underwrite the work - the tension between long horizon agents and selling the outcome
Kishen PatelAI personal assistants are free for anyone to use. But does everyone even need one?
Scott CattoA lot of tech advances take something that used to be a luxury and give it to everyone. Clothes cleaned for you used to mean paying someone to come scrub them. Washing machines. A personal driver ready for you at any hour of the day was very expensive. Uber. Instinct and Meta’s Muse are doing the same thing for personal assistants. The question is, do most of us not have a personal assistant because it’s too expensive, or because we just don’t need one? If you only have the occasional trip or boring online task, you probably won’t stay in the habit of using one. But now that the price is zero, maybe we’ll find a lot more use cases than we think?
Read discussion →spicy take of the day, these people have completely lost their minds
> I genuinely don't get Anthropic. It seems like every other day they're loudly puzzling over the ethics of a soul in a box - while building the box, claiming that it has a soul in it, and selling access to said box for $20 a pop. Like, dudes, these are incompatible positions. chris olah thinking he's gonna win a technical religious argument with a talmud scholar lol you bringing a fork to a gunfight dog
Read discussion →AI ‘godfather’ Yann LeCun has ‘zero concerns’ about human extinction, says Anthropic CEO Dario Amodei is ‘deluded’
I agree w Yann; most of these claims are entirely overblown
Read discussion →OpenRouter for tools
Liam HorneGoing to shamelessly share something we're working on at Tempo right now called Mercator The basic idea is aggregating 100+ services for agents (like web search, video generation, lead enrichment, bloomberg-like financial data, etc) into a single open marketplace of paid APIs where the best performing can compete for your prompt Not this: "use my elevenlabs API key and exa API key to produce a researched video on this topic" This: "use mercator to make a well researched video on this topic" Would love any feedback on the product and general idea space. We're trying to predict what the future will look like as agent queries become increasingly more specific and the internet molds around their needs!
Read discussion →Sholto predicts AI more capable than humans in next couple years
“We think that models which are as or more capable than all humans are very likely to occur in the next couple of years. So it's something that can do all of the things that a human could do on a computer, or once we get sufficiently advanced robotics, all of the things that a human could do in the physical world.”
Read discussion →The Compute Standard
Jay ScamblerFor the last few weeks my team has been building FundMyCompute, where we turn donations into AI inference credits. The hardest design question has been whether a credit should ever move from one person to another. Moves by some of the labs are starting to make tokens feel more like currency and this article wrestles with that.
Read discussion →DoW partners with In-Q-Tel
The partnership gives DoW’s research and engineering teams structured access to IQT’s technology scouting platform and portfolio of early-stage companies. The goal is to make it easier and faster to bring high-quality commercial solutions to warfighters.
Read discussion →Bad news for the pacing team
He's been clear this is a no go
Read discussion →The deployment layer for AI
Aaron LevieThere's a huge opportunity right now in being the deployment layer for AI into the economy. The amount of work it takes to change out workflows in enterprises tends to be far greater than anyone realizes or would prefer. Clearly this is what the applied layer of AI is going to look like in the form of software and agents, but also it opens up new services firms opportunities. Legacy systems need to be moved to the cloud, data organization and access needs to be updated, software needs to be connected to agents in new ways, workflows need to be reengineered for agents, HITL needs to be figured out for the process, evals need to be generated and maintained, and the entire system needs to be continually updated as new models get released and new capabilities emerge. And the full list may even be longer. AI is not the same as just deploying software. Software you generally did the implementation of an existing, well understood category of technology, then stepped back and the customer kept running. With AI agents, you're delivering actual work augmentation to the organization, which has a completely different set of complexities associated with it. You're no longer deploying tools that the company is merely enabled by, you're deploying work output in a process. Completely different implementation and enablement process. As a result, this is going to open up lots of new kinds of firms and plays for existing firms to diffuse AI into organizations. We're going to see approaches by industry, by size of company, and by problem inside of companies. Traditional SIs will modernize and adapt (some will clearly not adapt as well), and new entrants will also be founded in this period that take advantage of this window. Great time to be an FDE or FDE firm.
Read discussion →Do our eyes actually converge in VR?
Pieter NatanaelI’m curious about how our eyes behave when looking at virtual objects at different distances. For example, if a virtual object appears 50 cm away, do our eyes actually converge as if we’re looking at something 50 cm away, even though the display has a fixed optical focal distance? Does anyone know how this works?
Read discussion →dot = hardware device?
Dan Romeropretty useful name for a personal agent if you're planning to launch a hardware device
Read discussion →Launching Imbue Studio!
Kanjun 🐙Imbue Studio is a new kind of personal computer. Make durable personal tools and shape them just by telling Studio what you want. Share easily, use immediately across devices, and build together with others
Read discussion →my proposal for “pacing the frontier”
Elenamodel releases should be limited to 2x/year and coordinated as staggered seasonal collections, like fashion week. openai actually has the right idea with a september dev day but the other labs dont sufficiently respect the calendar
Read discussion →Reka AI Labs Releases Inverse Dynamics Model for Interactive World Models
Reka AI Labs released a new model designed to infer low-level actions from video and help train interactive world models using real-world footage without action labels. Seems like robotics is really having its moment !!
Read discussion →Advice is earned by understanding the work
A line from Matt Cynamon’s conversation with Fred Wilson and Nikhil reminds me of a useful standard for advising founders: writing a check doesn’t automatically qualify you to tell someone how to run their company. When Fred was 29, a founder challenged him to spend a week doing every job in the company before offering advice. Fred flew to Buffalo and did it. He quickly learned that advice lands differently when it comes from someone who has taken the time to understand the work and the people doing it. The conversation also touches on the other side of the job: building conviction when your partners disagree, while staying open to evidence that challenges your strongest beliefs. That balance seems hard to maintain. Confidence without curiosity turns into stubbornness; curiosity without conviction makes it difficult to act. I’m curious how founders and investors here think about this: What earns someone the right to give you advice? And what makes advice useful even when the person giving it hasn’t done your job?
Read discussion →America, going viral, funerals
Gaby GoldbergToday we sent out the latest edition of People Watching, a newsletter to help you discover great, under-the-radar people. Here are a few highlights: • @aijamayrock (835k+ on Instagram, 460k+ on TikTok) makes videos about fascinating topics, from technology and aging to Jews around the world. She spent two months in Japan as an Eisenhower Fellow and has written two bestselling books. • @SamRaus1 is the David Boaz Resident Writing Fellow at Young Voices. His writing on the past, present, and future of free society has appeared in USA Today, Newsweek, and The Hill. • @Simon__Grimm is an editor at @WorksInProgMag, where he covers AI and European progress—from the market for less capable models to liberal compute and talent sorting in Germany. • @being_on_line is a designer, artist, and cyberethnographer. In a recent conversation with @Inc, he explored the shift from monolithic to polylithic culture, and what “going viral” means when our feeds no longer give us shared references. • @kendallhtucker runs creative experiments at @tryramp, including debuting a Broadway musical and hosting a funeral for the penny. Outside work, Kendall has hosted a Hot Mitzvah, taught beer pong to Brits, and launched a Twitter lie detector. Special thanks to @knowerofmarkets for helping curate this week’s list. If you have feedback or recommendations, email or DM me. The full archive is at pplwatching.substack.com; subscribe via the original thread.
Read discussion →My favorite Jensen roadshow meetings comment from this week
Matt DratchRobotics “ChatGPT moment” is very close, “within a year” 👀 👀
Read discussion →What's your favorite Ai or agent?
I love claude code in the terminal, was jealous of GPT Astra when it came out but Opus 5.5 is amazing. Agent wise I've been using Muse and have 30B tokens :)
Read discussion →Rhun: a small and fast code editor written in Assembly
Someone wrote a full code editor in assembly. Vim mode, terminal, git diffs, even a Claude Code/Codex panel -- one static binary that draws every pixel itself. Solo project, MIT licensed. Respect
Read discussion →