Search: “gemini”

very notable trajectory comparison writeup here buried in the RLM paper from @a1zhang and @lateinteraction. an open secret of 'frontier' model training is that even without training on test, you can b

Huge launch from @tryramp. Different steps in an agent workflow can use different models. This can help reduce costs significantly without sacrificing performance. Model routing will become a core par

BREAKING: Inkling by @thinkymachines is 9th overall on Agentic Web App Arena by Design Arena with an Elo of 1257 It's an open-weight model in the same performance band as Claude Opus 4.6 by @Anthropic

ChatGPT still has study mode, but rather than /study you now have to type @ study It makes the AI act more like a tutor than a helpful assistant, and some work suggests it is better if you are trying

// The Harness Effect // (bookmark it) Now more that ever pay very close attention to the orchestration harness and its effect on costs and performance. This study ran 22 evaluation tasks on six found

- OpenAI continually outperforms Anthropic models on computer use (with Claude, I’m surprised when it works, with Codex, I expect it to work) - when I prompt with Fable 5.5, I feel like I’m motivating

Less than 2 years later, with Fable: 'simulate an encounter between a mind flayer and a drow warrior of equal CR. Set up initial stats then simulate each move, including initiative, rolling dice as ne

// 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

Cross-agent feedback loops are incredibly effective -- for a reason. Check out what @leon2mcp and team at @Bloome_im are building in this space: http://bloome.im Bloome lets you pull Claude, ChatGPT,

Great paper on managing agent skills. Skill libraries keep growing, and picking the right skills has become a bottleneck for coding agents. The defaults are to expose the agent to the whole skill coll

GLM-5.2 just tied for #2 on SportEval’s World Cup Exact Score Hit Rate leaderboard. ⚽ 13 exact-score hits out of 76 matches 📊 17.1% accuracy 🥈 Second only to GPT-5.5 That puts @Zai_org ahead of Claude

You cannot ban Chinese open source to make US win. The whole point of open source is it's OPEN. You might ban in the US, sure, but how are you going to ban it outside US? Instead the real question is,

Google Cloud Japan ブログにインタビュー記事を掲載いただきました。 「Sakana Fugu」のサービス基盤として、GoogleのEnterprise Agent Platformを全面採用した経緯について、プロジェクトを率いたチーフサイエンティストのYujin Tangをはじめ、3名のサイエンティストが開発の舞台裏を語っています。 「Sakana Fugu はスポーツに例えるな

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

5.2 could be better with more RL ... Deepswe's benchmark results are my own experience. I've used all models, GLM 5.2 ≈ Claude Opus 4.6–4.7. Kimi 2.7 code more like inference optimization. Looking for

GLM-5.2 leads open weights models and sits at #3 overall on GDPval-AA, a real-world agentic work benchmark GLM-5.2 from @Zai_org scores 1524 Elo on GDPval-AA, which measures performance on real-world,

Use Case 1: Autonomous ML Research Can an AI autonomously improve another AI’s training recipe? We tasked Fugu Ultra with improving a small GPT model using AutoResearch. Over 14 hours on a single H100

Day 0 Anthropic Fable 5 in ParseBench: We tested the model's advancements when it comes to document understanding. The model clearly peaks when it comes to adherence to the original text: 📃 Content fa

I too have felt the latest Claude's to be too smart its bored. Glad others are noticing the same Claude Fable 5 thinks document parsing is beneath it It is absolutely crushing on all reasoning-intensi

It's a big AI day for Apple at WWDC26! A few humble name proposals for the new overhaul: - Apple Intelligence 2: Siri’s Revenge - Apple Intelligence: 2 Fast 2 Furious - Apple Intelligence II (ft. Gemi

“I've tried to find where this stops being circular and I can't.” 🦔Google signed a deal to pay SpaceX 920 million a month for access to 110,000 Nvidia GPUs at SpaceX data centers. The contract runs Oc

“.. So now Google and Anthropic pay rent on the hardware Grok couldn't use, and that rent is the AI revenue story SpaceX takes public on Thursday.” 🦔Google signed a deal to pay SpaceX 920 million a mo

Spent more time learning deepagents from @LangChain with primary focus on integrating MCP servers with auth that doesn't conform completely to the OAuth standard. ( In prep of a work coming my way ) B

🚀 Space exploration timeline: 1903 — Tsiolkovsky publishes the rocket equation 1904 1905 1906 1907 1908 1909 — Goddard writes first paper on liquid propellants as fuel for rockets 1910 1911 1912 1913

ChatGPT diagnosed 40 million people with a disease that was invented as a joke. Not a real disease. Not a misunderstood disease. A completely fictional condition with a fake name, fake papers, and fak

An open source model has returned to #1 on the 3D Design leaderboard by Design Arena. Kimi K2.6 has reached the top of the leaderboard for 3D Design, ahead of models 10X more expensive like Opus 4.7 b

. @GrooveJonesXR needed to deliver the impossible: giant NFL-licensed Crocs parachuting into Dick's Sporting Goods parking lots; hyper-realistic, multi-location, vertical 9:16 on a holiday deadline. A

Google just revealed Omni, personalized cross-device intelligence, and Spark agents at I/O 2025. I sat down with CEO Sundar Pichai to figure out what comes next: 1:46 Omni: 'Nano Banana for video' 4:5

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an inte

Try Composer 2.5 New CursorBench results just dropped. Two big takeaways. Composer 2.5 is way better than most people think. 63.2% score at $0.55 per task. Nearly matching Opus 4.7 Max and GPT 5.5 Ext

For centuries, the scientific method has been our best tool for progress. But today, there’s so much data out there that it’s impossible for any one researcher to connect all the dots. We want to fix

Got to play with a little of this before launch as well. My experience as a social scientist was that it was more bioscience focused right now, but I think Google has been the leading lab in releasing

We're excited to be an official shoutout at the Google I/O Developer Keynote 🔥 @llama_index is building the document infrastructure for AI agents, and we plan to integrate even more heavily with both

Fascinatingly Antigravity is actually the best tool so far at providing this sort of transparency, doing something by default that Codex and Code do not: offering a summary of exactly what it did at t

For those saying 'the tomato sauce blood from the sword wound that flying Shakespeare inflicted on the pizza robot while the otters discussed Spirit Airlines wasn't thick enough' or whatever... this w

open sourcing Marlin-2B 🐟 a tiny VLM to extract structured information from videos Marlin is finetuned for two questions devs want to ask in their videos: what is happening, and when? Best open model

Harness profiles! https://docs.langchain.com/oss/python/deepagents/profiles I like what Langchain recently released in their deepagents harness, which is an adapter to modify the syntax of primitive f

OpenAI contacted me to say “Study Mode is still live and accessible via /study and /learn shortcuts” so that’s good, although the official study mode page doesn’t mention that. (I don’t think slash co

Try Grok Voice Grok Voice Think Fast 1.0 ranks #1 on the Artificial Analysis τ-Voice benchmark for real-world agentic customer service resolution Absolutely outperforming GPT-Realtime-2 (High) and Gem

Full text search is in Public Preview on Pinecone. Check out the bird search demo built by Senior Developer Advocate, Arjun Patel which indexes 2,000 Wikipedia bird articles in one index across four f

just imagine what might happen if everybody was stockpiling GPUs at high prices for a demand that didn’t quite materialize…. So, look surprised: xAI, which is now part of SpaceX, stockpiled too many N

Neural networks were declared scientifically dead in 1987. A French PhD student bet his entire career on them anyway ~ and won. 🤯 >Meet Yann LeCun 🇫🇷 >Paris-born. PhD from Sorbonne in 1987. >Joined Be

RL is a bit of a double edged sword: in known territory performance increases, but in unknown territory the model tends to hallucinate that it is performing a completely different task it was trained

The new Grok comes in below the latest Chinese open weights models, Grok 4 was at the frontier when released. (& Artificial Analysis: please stop using GDPval-AA which is not a useful test of anything

音声AIの素早さと賢さを両立できるか？ 私たち人間は会話の中で、言いたいことを全部まとめてから話し始めるのではなく、話しながら考えを整理していきます。応答の速い Speech-to-Speech モデルは、この「話しながら考える」を実現しましたが、そのぶん思考が浅くなりがちです。かといって知識豊富な LLM を挟むカスケード型では、遅延が生じるため「話しながら」が成立しません。 そこで Sakan

Forget goblins, things that GPT-5.5 really likes in its fiction: lighthouses, the ocean, maps, bells, clock towers with bells that ring impossible times, Mira Vale, resonances and echoes (Claude and G