There are nearly 200 agent harnesses to choose from. We gave three of them the same open-weight model and 90 AppWorld tasks, and changing the harness moved task completion from 30% to 81%.
Chips huggingface.co signal 85 primary Hot
Read announcements from Google’s 2026 AI in the Learning Era event.
Models blog.google signal 84 primary
Ask a frontier model for a slide deck and you'll get a good one. Ask for ten and they start to look alike: the same tasteful defaults, the same house style, a faint "AI-generated deck" look.
Chips huggingface.co signal 83 primary
The years-long wait for access to the grid and new turbines is causing data center builders to turn to an unusual, less efficient and less environmentally friendly solution: once-discarded jet engines.
Models techcrunch.com signal 79 verified
People hate data centers. They're loud, resource-hungry eyesores that are so profoundly unpopular they have managed to achieve the near-impossible in today's polarized political landscape: genuine bipartisan opposition.
Chips theverge.com signal 76 verified
Google announced changes to its AI offerings for schools that it said will give administrators more control over the technology—but that also will require those wanting to use its search service to use an AI-enhanced version. Under the new policy, access to Gemini, Gemini Notebook and Google...
Policy theinformation.com 2 sources signal 74 verified
Superwhisper is shipping as the default dictation setup in Omarchy 4.5. The free 10k word cloud option means you can get started without even having to download a model. Then you can switch to local-model dictation later if you like. It's so good!
High-engagement X post with 191 visible interactions on frontier model behavior.
@openrouter🥷 New stealth model: Crouton Alpha A highly efficient reasoning model built to scale out the work your agents repeat all day. - Free to use - 1M token context - Text, image, audio, and video input - No training on data
High-engagement X post with 292 visible interactions on frontier model behavior.
@anthropicaiTo that end, we're open-sourcing FlashProve’s code, here: We’re grateful to our...
New Anthropic research: progress on provable inference for large language models. Here, we introduce FlashProve. Building on established cryptographic techniques, FlashProve allows developers to verify whether a given response came from a given AI model....
Official or high-signal @anthropicai post with 1,992 visible interactions on AI developer tooling.
@adamholterererAfter having Codex check 120 different generated videos, models do NOT generate low-frequency airplane Dopplers
High-engagement X post with 2,761 visible interactions on AI developer tooling.
@tekniumYou can now join the waitlist for the Hermes Agent mobile app to get in as we get more seats!
Official or high-signal @teknium post with 478 visible interactions on AI signal.
@btrautSide chats in ChatGPT no longer expire or disappear when you restart the app! When I joined OpenAI, I found this to be one of the biggest pain points for side chat users, and as of today, it's fixed. Never lose data just because you updated, restarted, etc.
High-engagement X post with 377 visible interactions on frontier model behavior.
@nousresearchThank you to everyone who traveled from across the world to NousCon 2026. We loved spending time with the people who make the Hermes community great. As you all know, we released the Hermes Mobile beta at the event, and as anticipated it filled up quickly....
Official or high-signal @nousresearch post with 322 visible interactions on AI signal.
@teksedgeSo motherboard manufacturers are thinking of a future of cheap RAM where we can build these monster AI boxes at home. Until then, we can only dream. Gigabyte built a motherboard for Local AI where 128GB of memory is a quaint distant memory. How about 2T of...
High-engagement X post with 179 visible interactions on AI compute and inference.
@bchernySigning in to your company's plugins one by one is annoying, so we let Claude auto-connect to them now. Team and Enterprise admins can set them up, and teammates are connected automatically the first time they use the plugins.
High-engagement X post with 1,359 visible interactions on frontier model behavior.
OpenAIChatGPT and Codex are helping engineers build for the physical world. We’re sharing our progress across mechanical design, electrical engineering, and simulation—from helping our hardware teams iterate faster to testing new workflows with industry collaborators. https://t.co/b3rW9zYJGA
@primeintellectPublic evaluations are foundational for assessing model capabilities. We looked at some popular evaluations and found issues, which we fixed. Sharing our findings:
Official or high-signal @primeintellect post with 221 visible interactions on frontier model behavior.
@claudedevsEnterprise-managed auth (EMA) lets admins connect plugins for their entire organization through their identity provider, so members don’t have to search for, sign in to, and authorize each plugin themselves. It’s built on an open extension to MCP.
High-engagement X post with 139 visible interactions on AI developer tooling.
@claudedevsUsers can now auto-connect plugins when they log in to Claude, with MCP servers that accept Claude as a trusted issuer. If you’re building plugins for Claude, add support so users get started faster and stay connected without re-authorizing.
High-engagement X post with 1,178 visible interactions on AI developer tooling.
@natolambertFireworks serves >40T tokens a day of "specialized models." I'd define this as any post-trained model for a customer.
High-engagement X post with 175 visible interactions on frontier model behavior.
@mistralaiMistral Large 4 joins the model lineup on Open Math OS. 1.05T open weights, 1M context, live today for approved teams. Open Math OS is a research workspace where mathematicians and AI agents work on shared projects, built only on open models so the...
Official or high-signal @mistralai post with 147 visible interactions on frontier model behavior.
@simonwThere's a buzzy new paper about GPT-6 Astra being able to directly interpret model weights (prompts full of floating point numbers) - I tried it with Qwen 3.8 27B on my laptop and found that, after 45+ minutes, that model could achieve the same thing
High-engagement X post with 2,009 visible interactions on frontier model behavior.
@akshay_pachaarHarness Engineering 2.0 is here! After protocols for tools, agent communication, and agent interfaces, harnesses now have one too. Developers just built UHP (Unified Harness Protocol), a standardized interface for running agent tasks across any agent...
High-engagement X post with 241 visible interactions on AI developer tooling.
@mistralaiWe tasked Mistral Large 4 with creating a real-time, 3D tracker of the roughly 19,000 satellites and debris fragments orbiting Earth. The model built a working app, tested it like a user, made it fast, and passed strict verification. Please meet Orbital Trace:
Official or high-signal @mistralai post with 273 visible interactions on frontier model behavior.
@osansevieroWe just published a new tutorial on how to use Gemini for Computer Use with Cua Driver
High-engagement X post with 286 visible interactions on frontier model behavior.
@askalphaxiv"Mistral Large 4" paper published exclusively on alphaXiv Their latest technical report shares that aggressively reducing KV-cache memory can hurt an LLM's agentic capabilities, even when its language modeling performance looks unchanged. Mistral initially...
High-engagement X post with 206 visible interactions on frontier model behavior.
@askalphaxiv"Multi-Token Prediction Done Right" Multi-token prediction (MTP) helps LLMs predict a few future tokens, but this paper finds that around 80% of its training gradient comes from a noisy component that interferes with learning. They introduce Soft-MTP,...
High-engagement X post with 126 visible interactions on frontier model behavior.
@tekniumUpdate your priors: Every nous portal sub in Hermes comes with a cloud agent and it's own computer now, high usage limit models, and all of your tools like search, scrape, browser use, image gen and more
Official or high-signal @teknium post with 150 visible interactions on frontier model behavior.
@svlevineRL can help collect good data, whether you're doing BC or RL. It can also stack two cubes on top of another cube. Fun paper with @qiyang_li @zhiyuan_zhou_ @CharlesXu0124 on how RL can assist with robot data collection!
High-engagement X post with 125 visible interactions on robotics progress.
@emollickThis is a fantastic idea - a mass attempted replication of old findings of LLMs with new models. Much conventional wisdom is outdated. Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still...
High-engagement X post with 418 visible interactions on frontier model behavior.
No signals in the last 7 days.
AI-linked equities are broadly positive, with Oracle +5.85%, Advanced Micro Devices +3.65%, Super Micro Computer +3.12% leading the tracked basket.
| Ticker | Company | Move |
|---|---|---|
| ORCL | Oracle | +5.85% |
| AMD | Advanced Micro Devices | +3.65% |
| SMCI | Super Micro Computer | +3.12% |
| AVGO | Broadcom. | +3.07% |