Skip to content
Back

AI posts

AI posts

People and companies shaping AI. Filter your watched accounts, catch up on recent posts, or find the most liked and viewed.

Compiled by the Toolstash team · Updated

209 watched accounts · 2984 matching posts · Showing up to 3,000 recent stored posts. Counts are snapshots, not live X totals. This page refreshes every minute while visible. Last collected .

  1. Arena.ai

    @arena ·

    What if AI agents could safely improve their own harness at runtime? UC Berkeley PhD candidate and senior researcher @AlexKrentsel from @GoogleResearch recently presented his work on self-improving AI agents at Arena HQ.

    This architecture, Exo, borrows the exokernel idea from operating systems, separating an agent into a protected mechanism layer (state, history, secrets, and a "guardian") and a fully evolvable policy layer.

    Alex walks through the design, results on benchmarks and open-world RuneScape tasks, where Exo was 31-62% cheaper than Codex and Pi, and more.

    Big thanks to @AlexKrentsel for joining us to present his work with @istoica05, @martin_casado, @ankrgyl, Sylvia Ratnasamy, @BerkeleySky, and @netsyslab.

    https://www.youtube.com/watch?v=Jglgg0m8EIs

  2. merve

    @mervenoyann ·

    just in time for defenders 👨🏻‍🎤

  3. Mia

    @MiaAI_lab ·

    Another DGX Spark joins my lab!

    The RTX PRO 6000s and the additional DGX Spark will help me do things I couldn't do before, and accelerate everything I do - faster and better.

    I'd like to thank @NVIDIAAI for sending me these 💚

    No one can deny their local AI efforts. I've been saying this for a long time - they have the best GPUs and the best software stack out there.

    Expect new and interesting stuff from me once I settle everything the way I want. Time to push local AI to new limits!

    Media shared by Mia
  4. kache

    @yacineMTB ·

    Yep. Just like yoshua benigo

  5. Maziyar PANAHI

    @MaziyarPanahi ·

    Mistral just put Mistral Large 4 on Hugging Face, so i gave it my medical exam

    1,273 USMLE-style questions, 3 runs: 94.3% right, no answer cut off

    a model this good that a hospital can run inside its own walls is a big deal 🔥

    what should it read next?

    Media shared by Maziyar PANAHI
  6. Romain Huet

    @romainhuet ·

    Thanks for your feedback! Enjoy double the usage limits and happy building!

  7. Joshua Achiam

    @jachiam0 ·

    People who have been in research fields develop a felt sense that most "research" (in almost any field that isn't exactingly technical) is research theater, and many superstar careers enjoy star status not because of a unique connection to the truth but because of a unique connection with an audience. No one really has the time or stature to, from a position of objectivity and trust, bring popular researchers to heel. But it will get increasingly hard to hide from serious critique. The machines can do critique now, at the level of depth and rigor previously only possible for the academically deranged. We're going to have to revise a lot of what we "know" about the world that isn't really knowledge but is actually just received popular consensus - like floating voltages without grounding wires. The grounding wires are coming.

  10. Shubham Saboo

    @Saboo_Shubham_ ·

    Crisp, concise and clear writing is such a huge moat when AI does all the writing.

  11. Claude

    @claudeai ·

    A few upgrades to the Claude mobile apps this week:

    On iPad and iPhone Duo: what Claude makes now opens beside the chat instead of over it, and goes full screen when you want a closer look.

  13. Miles Brundage

    @Miles_Brundage ·

    We must bring back Emeril Live

    Media shared by Miles Brundage
  14. Daniel San

    @dani_avila7 ·

    If you have a Claude Max plan, you can claim $200 in API credits to power your apps and agents

    I already claimed mine, and the credits are in my account 🙌

    Claim yours here: https://clau.de/api-credits

    Media shared by Daniel San
  15. Andrew Curran

    @AndrewCurran_ ·

    OpenAI has updated their model spec around five core values:

    - faithfulness to human intent - harm minimization - honesty and transparency - human agency - care.

    'The Assistant,' which has been around for a very long time, has been replaced with 'GPT.'

    Media shared by Andrew Curran
  16. Derya Unutmaz, MD

    @DeryaTR_ ·

    Wow! Rejoice, all pro OpenAI accounts: usage limits are not halving, keep Codexing!

    Dots is also an incredible value that does not count toward your limits; it’s essentially a super-capable AI agent working for free! Mine is working non-stop☺️

  17. kache

    @yacineMTB ·

    I'm a good programmer. God damn

  18. Santiago

    @svpino ·

    There's no way call centers look the same after this.

    Listen to this conversation.

    You can't tell the person on the other end is an AI agent.

    This is now the #1 voice model on ServiceNow's EVA-Bench right now.

  19. dax

    @thdxr ·

    hex has transforms now so i can voice prompt and it'll go in all lowercase so you can't even tell that this was voice prompted it even lets you specify a custom function to implement whatever logic you want

  20. Mia

    @MiaAI_lab ·

    OpenAI is doubling ALL usage limits on $100 plans and above. This is a move in the right direction.

    STOP WITH THE RESETS!

  21. Matt Pocock

    @mattpocockuk ·

    Don't tell me /pr is in the model weights already

  22. Alex Imas

    @alexolegimas ·

    This is just epic. No words. The Power Broker is a singular book, there is just nothing quite like it. Reading it was an event in my life.

  24. Harrison Chase

    @hwchase17 ·

    a lot of the challenges in harness engineering are now best solved by closer integration of harness and infra

    async subagents and browser use are two examples of this

    now in MDA 0.10!

  25. Miles Brundage

    @Miles_Brundage ·

    You’re telling me this for the first time

  27. OpenRouter

    @OpenRouter ·

    We just shipped our biggest improvement to our provider router since the founding of OpenRouter.

    Before, we prioritized mostly based on price. The new router now rewards the best quality inference for the actual cost of a request, so requests stay cheap but are also fast and reliable.

    Providers that deliver more value automatically win more traffic, so every provider has a reason to keep improving. And every time a provider gets better, so does your inference on OpenRouter.

    Media shared by OpenRouter
  28. Josu San Martin

    @josusanmartin ·

    This is what the LZB train-safety system looks like, the system that failed ahead of the Spanish train crash.