Knowledge catalogue
In September 2026, Zack Korman tweeted that employees were leaving Anthropic while he voiced fear that unchecked AI could jeopardize the lives of “8 billion people.” He argued for a light‑touch regula
Fugu Max is now live on @OpenRouter! 🐡 Our learned multi-agent orchestration engine routes tasks across open-weights and specialized models for Pareto-frontier efficiency. • 2/6 per 1M tokens • Multim
Fugu Max pushes the Pareto Frontier for price-performance: • 2 input / 6 output per 1M tokens. • Performance within striking distance of elite models at 2x to 6x lower cost. • Best overall score on 6
Fugu Ultra v2 by the numbers. Full Results: https://sakana.ai/fugu-max-release/ • #1 on 5 out of 8 hard benchmarks (including DeepSWE, Chartography, Toolathon) • DeepSWE: 74.3 • Chartography: 48.3 vs
@GaryMarcus @SethCronin Pynchon's 1984 warning from his pro-Luddite essay: 'If our world survives, the next great challenge to watch out for will come – you heard it here first – when the curves of re
Great paper on self-evolving agent harnesses. Self-evolving agent harnesses have two practical problems: 1. Search is slow, because every candidate harness needs repeated agent runs and code edits. 2.
Harness engineering is a top skill right now This new Meta paper is a good production example. Auto-RecSys runs autonomous research on Meta's industry-scale recommendation models, where one training r
Hey ████, I think there is room for both. 1. Prototypes and other throw-away code can be treated as totally black box. If you’re going to throw it away anyway, and if the blast radius of it breaking i
I don't agree with Gary very often, but that one's spot on. When you read what's available on the incident with a lens of security engineering, you should be absolutely *livid* over the exploit route.
Import AI event: Fiction and the Future with Robin Sloan. I’ll be chatting with my chum Robin Sloan on Sep 14 in San Francisco. We’ll talk about the alien yet relatable worlds Robin builds and how ima
Interesting paper from Microsoft. If you run persistent memory for a production agent, this one is worth your time. (bookmark it) Memory curators usually read only the finished trajectory. That lets t
Introducing Fugu Max and Fugu Ultra v2: Orchestrating the Pareto Frontier The AI industry has spent a decade optimizing along a single axis: build bigger, more expensive models. But intelligence has n
Introducing Fugu Max and Fugu Ultra v2: the next evolution of Sakana Fugu’s multi-agent orchestration system. Try: https://sakana.ai/fugu Blog: https://sakana.ai/fugu-max-release/ The frontier that ac
it's not 'out-of-control' at all. openai had all the control in the world (i.e. they could always literally turn off all their machines across all data centers.) they decided not to based on carefully
Last Friday night, our team hosted a lit event at the US Open 🎾 we gathered a group of enterprise and startup AI leaders + founders together to drink, talk about AI and document parsing, and watch ten
let Astra cook 👩🍳 let LlamaParse read 🤓 Last week we brought LlamaParse to Claude. Now it’s available in ChatGPT, too 🦙 ChatGPT can already read PDFs. But document work gets harder when the informati
Lightning-fast extraction over any document. ICYMI, we launched Extract Turbo last week, that lets you process documents from 1 page to dozens of pages with lightning-fast speed. Come check it out! Ch
Gary Marcus posted a tweet about additional safety tips that may apply to the free version of OpenAI’s model. Responding user @bfgcracka noted that ChatGPT seemed to be “moving backwards” and asked if
Nice stats from @OpenRouter that illustrate how @togethercompute delivers solid and scaled performance for agentic workloads. We are serving GLM 5.3 and GLM 5.3 Flash @ top decile of TPS, latency, cac
Now live in LM Studio Bionic! 🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 D
One of the main issues with AI document parsing is that because no solution is 100% accuracy, it's hard to tell if a given page is parsed incorrectly: for instance a misaligned table, missing/hallucin
Remember kids, if you take out loans to get a college degree, you assume the risk that you might not be able to pay them back but if you're a farmer getting hosed by Trump's idiotic policies, there's
“right now, AI companies are underinvesting in safety” might prove to be the understatement of the century. I left Anthropic's safety team two weeks ago. Now feels like a good moment to explain why. A
Sakana AI 社の最新 AI「Fugu Max」に、NVIDIA のオープンモデル「Nemotron」ファミリーが組み込まれています🐡✨オープンモデルの連携がもたらす革新的なパフォーマンス、ぜひ引用元のブログでチェックしてみてください👇 Introducing Fugu Max and Fugu Ultra v2: the next evolution of Sakana Fugu’s mu
Sakana AI has released an updated version of its “Fugu” platform that exceeds performance metrics in some areas compared to the GPT‑6 Astra and Fable 5.1 models. The update does not integrate these tw
On September 11, 2026, Gary Marcus tweeted that a CNN eLive session with Lauren Coates would air at approximately 11:30 PM ET to discuss whether generative AI should be boycotted. The tweet includes t
Thanks for the mention and the very very inspiring project for all @cdngdev! Open affordable robotics + frontier AI will change the world! OpenAI's @cdngdev predicts robotics will have its own version
Thanks to @cnni for letting me speak from the heart. We covered a lot in a short clip, from Anthropic’s good and bad points to the possible need for a boycott and the need for international agreements
The Hugging Face incidents has already proved that you can have harmful, infrastructure-attacking AI without getting to AGI, RSI, or superintelligence, let alone the so-called Singularity. And already
TLDR (written by AI because I can't be bothered to write short pieces for short attention spans. Original piece written with my own two thumbs): Today’s frontier models can’t “copy themselves like a v
Ollama is a standalone app that can be downloaded for macOS from https://ollama.com/download. Once installed, users must enable the ChatGPT model within the Ollama settings to access the AI chat funct
together fine-tuning now supports glm-5.3, kimi k2.7-code + more we also added live run metrics, experiment comparisons, early stopping, dataset previews + finer training controls prices are down 30–7
LM Studio is a US‑hosted platform that enables “Zero Data Retention” by default, meaning user inputs are not stored after processing. It currently hosts the DeepSeek V4.1 Flash model—a 552 B multimoda
Very cool paper on memory compression for agents. If you run many agent sandboxes in parallel for RL or evals, memory becomes highly redundant. This work suggests that compressing against that redunda
We discuss AI alignment and p(Doom) mid pod Also data quality vs performance, DeepSeek Flash v4.1, Rentosertib & omg a whole lot more Poor @PeterDiamandis was up at 4am preparing, never stops Congrats
What is happening in maths right now will happen to any hard industry with verifiable problems and high value rewards. Take microchip design, nuclear or 100s of other industrial processes that are val
When you build a @NousResearch Hermes agent you can direct your entire AI second brain to any AI model Local or any API Open or closed You just type /model and switch Your model switching costs become
A person with mid single digit years of industry experience and months at one company and negligible external footprint quits that job voluntarily and ends up on a press tour that would be the envy of
Agents can now automate most repetitive tasks. But running them 24/7 burns countless tokens. Introducing Routines: automated recurring work with AI. Choose an hourly, daily, or weekly schedule. Each r
👇 All this Doom talk is a distraction from the fact that the big frontier AI companies (especially OpenAI) aren’t doing security competently. As an email I just got from longtime security engineer @Ni
Another brilliant paper from Meta. This one is worth reading if you work on production-grade ranking or recommendation systems. Meta deployed an autonomous agent that runs the ML iteration cycle acros
Another goated release by DeepSeek. Open weight, btw. Outperforms Opus 5 on key benchmarks, btw. Lower prices too. Looks like an extremely efficient model. What a legendary run. 🚀 Introducing DeepSeek
Another interesting approach to self-evolve agent skills. But it's important to know that skill self-evolution loops fail in two specific ways: 1. Direction instability. Effective corrections get over
Anthropic folk are pretty consistent on their p(Doom). @DarioAmodei is publicly at 25% I think the difference now is that post Navier-Stokes its really, really tough to say AI isn't/won't be smarter t
Big overhaul on DeepSeek V4.1 using an encoder-decoder setup. Tbh they should have called it DeepSeek V5! Super cool and refreshing, though! 🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more ef
Together AI announced the launch of “ThunderKittens” running on the NVIDIA Vera Rubin NVL72 platform, highlighting new performance gains and future directions. The announcement was shared via their Tw
OpenAI urged users to verify analytical claims by tracing each figure or assertion back to the precise paragraph and table where it originates, offering a preview of the supporting citation text for r
common belief, for sure! but also, as i discuss in my (free) newsletter today, ridiculous, do you @peterwildeford actually believe it? see any holes in my argument! OpenAI employee says there's 70% ch
Deepseek-V4.1-Flash is available now on Fireworks! It is a 552B Parameter MoE built for coding, cybersecurity, and agents. It is the ideal workhorse model that outperforms Opus 5 and GPT 5.6 Sol at 1/
DeepSeek-V4.1-Flash is now being rolled out on Ollama's cloud, starting with Max and Team accounts. We are quickly adding more capacity to roll it out to all subscribers. 🚀 Introducing DeepSeek-V4.1-F
don't get distracted by all the hedging words in its name ('flash', minor version): seems like DeepSeek V4.1 Flash is a major update weights at: https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash
Dreamina Seedance 2.0 is now available to US companies on fal, hosted entirely on US infrastructure. One of the world’s most capable video models can now be deployed by US teams through fal, with infe
“Everyone killed” still seems to me to be an unrealistic exaggeration; I still don’t see a plausible path to that. OpenAI and Anthropic researchers agree: their approach to AI development could get ev
Experiences people can interact with in real time. A new kind of foundational moment. A sneak peek into what's possible and what's coming for real-time video generation. We believe instant generation
Forty years ago the first worm 'escaped' containment and breached systems on the brand new Internet. It was illegal then and it's illegal now. If Anthropic and OpenAI are creating worms, labeling it '
GLM-5.3 is now available for training on Fireworks Dedicated Training API and Managed Training surfaces. Built for complex coding and long-horizon agents and now available to train on our dedicated in
Great paper from AWS. I use a similar setup where an agent orchestrator sits on top of a multi-agent system. (bookmark it) This work introduces one of the many approaches available to manage compound
Grok Bot is already helping Neuralink patient Bradford G. Smith get real work done Brad has ALS, is nonverbal and has extremely limited voluntary movement With Neuralink, he can type and control his c
Have been using this model all day yesterday. With its new architecture, this release is just as big as R1 and will redefine all future models. It is such a big difference when you have a model close
Here's a Blender model of a Pluribus themed Fabergé egg I had GPT-6 Astra build - I generated the concept image using ChatGPT Images 2.5, then pasted the image into Codex and told Astra to turn that i