@tenobrus I think there are AI research teams with novel research ideas who should be able to take a shot, and part of the reason why I created my neolab tracker was to more easily see who is actually shipping (outside of fundraises and VC valuations)... https://t.co/btKbgWaccG
@GarrisonLovely If you ever needed a visual, I have been going through what's been published by the labs so far re: third-party evaluators (wip obviously)...
https://t.co/pteG5hUNAK
@jonnym1ller There's something like 6000+ abliterated models available via Hugging Face + their hosting partners (not all are hosted and not all are as powerful, but still).
@eliebakouch The influencers and others who fail to disclose payments or relationships (that means you, investors) are way more damaging to the ecosystem. (i.e. Why did this person all of a sudden fall in love with a new devtool or agent I have no idea, etc.) I'd rather see the disclosures.
I built an MCP that lets Claude Duet, GPT Dots, Grok Bot, and Muse talk to each other and work together as a team.
Made it for myself and I use it every day. My friends wanted to use it too, so I made it public and open-source.
Free for the first 1,000 users.
Fork it and self-host, or mod it any way you want.
Even works with ChatGPT Plus — ask GPT to check hitchike every hour, then connect it to free agents like Muse to get a more Dots-like experience without losing context from ChatGPT (and without paying $100/mo)
OPENAI AND GOOGLE STRIKE MAJOR AI DISTRIBUTION DEAL
OpenAI will bring its frontier AI models, including GPT-6 Astra, and Codex to Google Cloud in the coming weeks.
Enterprise customers will be able to purchase them through Google Cloud Marketplace, using existing cloud spending commitments.
The agreement follows OpenAI's expansion to AWS after its exclusive Microsoft distribution arrangement ended.
A neolab is an AI research team with a novel research agenda, whose goal is to commercialize their research.
A neocloud sells GPU compute and AI infra.
Most of the teams working on RL envs/evals/data can be classified as builders, some of whom call themselves data labs.
PI is a little different, as they have a novel research agenda and sell compute, but the others you listed release research to support their offerings.
I do track neolabs, which is why I have an opinion :) → https://t.co/u33EiPlZVd
Serve Robotics, which runs most of LA’s sidewalk delivery robots, lost $101.4 million in 2025,
($371,286 annual net loss for every robot it had out making deliveries.)
The first advantage its investor deck lists for the robots is “No driver wages.”
This is a slide from that presentation
My office is hiring an Artificial Intelligence Policy Fellow to help shape AI legislation and track AI developments in the private sector as Congress works to keep pace with the technology.
Experience in AI, policy research, or national security is encouraged.
We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program.
We have been building a cybersecurity benchmark based on simulated corporate networks. This is a valuable indicator for the cyber-risk a model poses if in the hands of a bad actor. We plan to make the full evaluation public in the coming weeks but wanted to share our first immediate finding here.