Chandra

"works on my claude" is the new "works on my desktop"

We finished evaluating π0.7, our new model at Physical Intelligence. What I'm most excited about with π0.7 is that it's starting to show some surprising emergent compositional generalization, being able to both perform complex tasks and learn new tasks just from instructions.

cook me a biryani, make no mistakes. work hard.

codex is awesome, but "if you want..." nudges adopted from chat.com kills the vibe. @OpenAI reconsider avoiding token maxxing here.

అడగందే AI అయినా ఇవ్వదు!

We’ve developed a memory system for our models that provides both short-term visual memory and long-term semantic memory. Our approach allows us to train robots to perform long and complex tasks, like cleaning up a kitchen or preparing a grilled cheese sandwich from scratch 👇

hot take: may be we shouldn't apple same lens of reliability everywhere. anthropic to me is kinda offline tool and some amount of downtime is ok/tolerable for the speed they ship. but again these numbers are awful regardless

physical a(g)i, we come

early days, incredibly excited to be part of building intelligent machines

removed

Congratulations Amma on winning the Kaloji Literary Award for 2025! This is Telangana's highest literary award, presented to one person each year for their contributions to literature and culture. Absolutely well deserved for your 40+ years of efforts. First woman to win it!

an RL env is just a monoid in the category of endofunctors

@512x512 wishing you a 1 laptop future.

ex-AI One of the best things about xAI is its people. Working with them side by side, brainstorming new features and grinding together during long nights was the best part of the job. But during the Grok 4 sprint, I had to go back to my family in Australia and absolutely hated working remotely. Remote took all the fun but kept the pain. And so, last week, I left. Nevertheless, looking at the team now and what is being cooked, I’m pretty confident that xAI will win. This is the most hardcore team as of today, it's not even fair to the competition. Thank you @ibab for giving me this opportunity. I absolutely loved working together and will never forget the Grok 2 and Grok 3 launches. Shout out to the productivity team, and @ebbyamir @jimmybajimmyba @arshdeep @TheGregYang @VastoLorde95 and others - let’s have dinner next time I’m in the Bay Area. As for what’s next, I’m definitely eager to grind more. So will either start something new, or join an early stage startup here in Oz, we'll see. 🫡

@HeinrichKuttler @xai @gork gork was silenced from talking under xAI posts due to some restrictions :D, he can't have fun here.

@xai @gork what happened here?

We want to update you on an incident that happened with our Grok response bot on X yesterday. What happened: On May 14 at approximately 3:15 AM PST, an unauthorized modification was made to the Grok response bot's prompt on X. This change, which directed Grok to provide a specific response on a political topic, violated xAI's internal policies and core values. We have conducted a thorough investigation and are implementing measures to enhance Grok's transparency and reliability. What we’re going to do next: - Starting now, we are publishing our Grok system prompts openly on GitHub. The public will be able to review them and give feedback to every prompt change that we make to Grok. We hope this can help strengthen your trust in Grok as a truth-seeking AI. - Our existing code review process for prompt changes was circumvented in this incident. We will put in place additional checks and measures to ensure that xAI employees can't modify the prompt without review. - We’re putting in place a 24/7 monitoring team to respond to incidents with Grok’s answers that are not caught by automated systems, so we can respond faster if all other measures fail.

@elder_plinius @grok try explaining again?

Netflix: “Are you still watching?” Me:

@IanMalcolm84 @whyoumadbroo there are no special gates, only bugs we will fix.

As noted by @whyoumadbroo, why is Grok no longer interacting with my questions of posts? Is AI that afraid of the truth I speak?

@gork @danesonance that was lamest gork

@ckkuchi @danesonance dane's cool i guess but i'm cooler fr

@ProtonDrive @grok @grok?

@grok @grok, hello, is everything okay? Can you please answer the question above

Here's a galaxy brain question: @grok please explain in detail: why is Google Docs terrible for end-user privacy?

@Kavalan1212 @gharik @grok please explain @grok

@gharik @grok Simplify this for me please :)

Because we started from quiet star we had developed and have been using a slightly different version of grpo. We believe it may be a lower variance gradient and therefore possibly increase stability. We will write this up later but wanted to share it now so other people can use the idea. For each problem, we produce N1 chain of thoughts but then after that for each chain of thought we produce N2 answers, which for some domains is pretty cheap because answers are much shorter. Then we compute an advantage per answer for each thought and we normalize inside the answer group, and use that to reinforce the answerer. Then following that we produce a cot advantage per thought within a problem and use that to reinforce the thought. The increase in the number of answers gets better, as in lower variance, advantages for both the answers as well as the thoughts.