Search: “claude”

Don't train the model, evolve the harness. I read a brilliant blog post from Hugging Face where they took a frozen open model scoring 0% on a hard legal agent benchmark, left its weights alone, and le

Some notes from @aiDotEngineer world fair: > the energy was incredible. it's magical to have a large group of smart, hungry, technical, and driven people learning from each other under one roof. > the

// AutoMem // I quite like this idea of metamemory. (bookmark it) This new research from Stanford treats agent's memory management as a trainable skill instead of a fixed module. The model decides wha

I just left the final day of the @aiDotEngineer World's Fair Conference in San Francisco. Kudos to @swyx for putting together a world-class lineup of speakers and workshops! It really was an invigorat

I’ll be actually-homeless soon (living out of my car) if I can’t land a job. Our house is being sold within two months if I don’t find work. I’ve been trying everywhere, but very few companies are ans

my fave question, talked about this coding agent Eval+Improvement loop infra + UX in my AIE talk yesterday! biased but LangSmith is the best spot to Eval + continuously improve your coding agents, and

The talk about Mythos and cybersecurity was not, in fact, hype. (As anyone using Fable to do autonomous work has probably recognized) AI appears to be finding software vulnerabilities at scale. In Jun

Fable, the indomitable model from Anthropic, is coming back. These last two weeks have been a great reminder that: models can be removed at any time, you should have fallback plans (including open sou

'For agents within an actual product, I have gravitated to using LangChain DeepAgents. I feel like their base harness is less heavy, and their abstraction for building agents that use skills and tools

good post on the value of knowing whats going on in your harness 'For agents within an actual product, I have gravitated to using LangChain DeepAgents' The frontier model lab harnesses are amazing - d

Have seen some questions about the updated classifiers and wanted to clarify. As with the original classifiers, a small fraction of routine coding and debugging tasks will be flagged and fall back to

Really confused by all the excitement I see in my timeline for a nerfed model. Never seen anything like it. So many will end up very disappointed. Time to rethink how to build around frontier and open

Every article saying 'THIS IS THE NEW JOB OF THE AI ERA' is one of two jobs. And the first is a lie imo. The first job, is usually something that an AI lab is hiring for and the news blows it complete

Fully solving document parsing includes covering every point on the Pareto curve of accuracy, cost, and latency: 1️⃣ High-accuracy parsing - requires 99%+ accuracy, price insensitive. Especially relev

Introducing SWE-Together: a multi-turn benchmark built from real user–agent coding sessions. Coding agents are often benchmarked like exam-takers: given the full spec up front, then graded on the fina

Sonnet 5 is here! This is going to support better long-running agents. Previous Sonnet models were unreliable, so it's great to see the improved version that can complete agentic tasks more reliably.

𝗚𝗟𝗠-𝟱.𝟮 (the latest open weights model) is having an Enterprise moment, and it is not an exaggeration.🚀 🔥 We have been impressed by how strongly GLM-5.2 is pushing long-horizon performance .. not just

Most popular model on @OpenRouter (10tr tokens) turns out to be a 1.6tr MoE by @Meituan_LongCat (superapp/DoorDash of China) Basically Gemini / Opus 4.6 level 35tr tokens trained entirely on 50k Chine

GLM 5.2 is the ultimate culmination of what i have argued here for the last three years. no technical moat and widely known formula -) convergence -) price wars -) small or negative margins. GLM-5.2 i

More reason why we’re excited about GLM-5.2 on Together 👇 Strong enough for serious coding work, cheap enough to change routing decisions, and easy to access through the tools developers already use.

I put together a new article on setting up local coding agents with open-weight models. Everything runs 100% locally. I thought it might be useful putting this together because many people asked me ab

The take that frontier models aren't ready for use in medicine is dead wrong. I've been using Opus 4.x in my clinical workflow every day for 6 months. I'm a deep sub-sub-specialist in dermatology - th

Fireworks AI is now live on EvoSkill v1.3.0! You can now use @FireworksAI_HQ directly with EvoSkill to run fast inference on open models as both the evolution harness backend and the LLM scorer. Along

📣📣 Meet Qwen-AgentWorld — a native language world model that simulates 7 agent environments (MCP, Search, Terminal, SWE, Web, OS, Android) within a single model. Environment modeling is the training o

A case study in why organizations should both incentivized their employees to explore AI uses that help them & have a Lab of dedicated AI builders Here, Cornell's finance & AI teams created a /treasur

Gray Swan: Red-Teaming after Mythos & the coming AI security crisis https://www.latent.space/p/gray-swan @GraySwanAI cofounders @zicokolter and Matt Fredrikson explain why AI security is fundamentally

Here is the prompt method behind this AR try-on app. The trick is not a magic prompt. It is the architecture of the prompt, and it works across GLM-5.2 and other frontier models. Full prompt: http://c

New listen: @GraySwanAI cofounders @zicokolter and Matt Fredrikson sat down with @swyx on @latentspacepod to unpack why AI security is its own discipline; not just 'cybersecurity with AI.' As agents l

The more you embrace AI, the more you need SaaS. This is not obvious to armchair market analysts who love disruption narratives, but it is obvious to people actually running companies. Levie now uses

'Verifiable results (True/False) combined with a solution space exploration technique' - very well put. Crux of neuro-symbolic #ai, what @GaryMarcus has been shouting about it for decades. Crux of the

I have some very big news... KernelBench-Hard with H100 and B200 (single gpu results) AND KernelBench-Mega tested on RTX PRO 6000, H100, B200 is finally out! Starting with Mega, each of models wrote a

it looks like AIE 2026 is mostly sold out, for anyone who couldn't get a ticket for this year, I love that AIE would livestream and have this archive of the conference and workshop. A few of my favori

Over the last two weeks, both the U.S. Government and Anthropic took significant actions that demonstrated their power to control access to AI by restricting what others can do with frontier models. T

POD UP! 🚨 Besties are back to discuss: -- SpaceX's record IPO, Cursor deal, and the first trillionaire -- The modern politburo and the new oligarchy (@friedberg cooks) -- Behind the scenes of the Anth

The model subsidies will eventually end and this workflow of “creating loops that will prompt your agents” will result in massive amounts of code that’s not well understood that you will have to pay l

As believers of open research, we are disappointed to see Anthropic silently degrading Fable 5 for AI development 'Any topic related to building pretraining pipelines, distributed training infrastruct

HISTORY LESSON: In 1968 the US, USSR, UK, France, and China signed the Nuclear Non-Proliferation Treaty, declaring nuclear weapons too dangerous for any more countries to build. All five already had t

This is a great article on how startups/frontier labs can coexist. Another way to look at this is task complexity - the number of bits of information needed to specify a task such that AI can solve th

👇@zerohedge confirms what I said here and in my May 29 Marcus on AI newsletter. Tokenmaxxing RIP. This morning @zerohedge out with a report on the death of Tokenmaxxing 'Microsoft’s AI Chief added to

If you thought AI progress was slowing down, well here's the immediate answer to that. Huge jump in capability across the board. This is going to deliver major improvement in agents across almost all

mythos will be bad ON PURPOSE on ai 'frontier llm research' tasks, this is very very sad for the research community also the fact that this is un purpose not visible to the user is crazy Introducing C

Oh my God! @METR_Evals’s coding benchmarks are saturated! 🤯 Mythos broke the METR graph 🤯 4 weeks later, out comes a new coding task, this time from @cognition: “FrontierCode Diamond remains unsaturat

Rubrics are even more flexible than /goal You can define a custom subagent for the grading, including custom tools, prompt, and iteration limits. Try it out and let us know what you think! we just shi

So excited to be opening up OpenEnv to the whole community. It will now be owned by @huggingface , Meta-PyTorch, @reflection_ai , @UnslothAI , @modal, @PrimeIntellect , @NVIDIAAI , @mercor_ai , and @f

we just shipped support for rubrics in deepagents ✅ give your agent a clear definition of what 'done' looks like, and force it to run in a loop until said goal is complete this is similar to /goal in

it’s a bad day when @scaling01 goes all Gary Marcus on the media 🤣 This is a good example of how AI news gets distorted. The example is from Tagesschau, the most watched news program in Germany. Anthr