Amey Muke

1/🧵Was grinding some OSINT challenges late at night and had a stupid idea: what if two AI agents competed to geolocate a photo — and the loser gets eliminated? So I built it. Here's what happened. 🕵️‍♂️ #OSINT #AI #CTF #Geolocation

Conducted 2 technical interviews today. Reminded me how serious of a task this is. You get to decide potentially the next n number of years the person you are interviewing within the span of 30 minutes or so. I have nothing but hate towards those who conduct interviews without a care asking dumb leetcode hard questions they themselves couldn’t solve if the shoe was on the other foot.

@ryanvogel what is happening @thdxr

yeah no comment

@rumik_ai blr when??

@forwarddeploy dmed

Great time to join, email us or if we know each other already please do text/dm Come join us on the journey at SpaceXAI!

@udayan_w applied

apply here growthx.club/events/voice-a…

Presenting Voice AI Buildathon powered by ElevenLabs Voice AI is in production everywhere. That's not the interesting problem anymore. What breaks when you scale it is. Three things always break. 1/ cost 2/ latency 3/ memory and they're not independent. fixing one usually wrecks the other two. this buildathon is a benchmark for exactly that. What you'll build Pick the problem that's actually killing your production system: cost, latency, or memory. Build against it. you'll be judged on how well you hold the other two while you do it. That's what separates a production voice agent from a demo. The stack ElevenLabs provides voice infra. You get to pick the brain: claude, gpt, gemini, or your own model. Where is this happening? @ElevCap, bengaluru. Who should apply? 1/ Technical excellence: ai/ml folks,, engineering and senior product leaders. 2/ Domain excellence: you are a sales, marketing or a support leader who has built before. this is not a beginner friendly event Apply below.

@elonmusk Do check this as well @elonmusk @billyuchenlin x.com/7613Perfect/st…

I will be reviewing all emails that pass reasonable sanity checks personally

SpaceX is actively hiring world-class engineers/physicists for SpaceXAI, even if you have zero prior experience in AI. Smart humans figure it out fast. Please send an email with ~3 bullet points demonstrating evidence of exceptional ability to ai_eng@spacex.com.

@phroo Mailed @phroo

Come join us for the time of your life !

@elonmusk sent @elonmusk :))

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids. An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better. This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

We've been blown away by the community's response to EVEE, our tool for SOTA genetic variant effect predictions. We're open-sourcing a tool to make it even easier for anyone to use! Check out the EVEE MCP and Claude skill in thread (1/3)

@yeahfortommy @NousResearch @yeahfortommy dmed you

looking for someone to help us bring Hermes agent to the world. if you're a hermes agent power user and want to help others become one as well, hit my dms or email us to apply! more info on the forward deployed engineer role: nousresearch.com/forward-deploy…

@BrendanFalk dmed

If you're a harness engineer and think you could crush this interview, please DM me! Hercules is pushing the frontier on coding agents. We have 250k+ users, millions of revenue, and lots more to do

@himanshustwts Make a llm that solves two sum , N queen problem from scratch if you need you could use autoresearch as well

i am genuinely curious about interview process (if any)

Welcome to Gemini 3.5 Flash, our most powerful model to date. It pushes the frontier of intelligence, speed, and cost putting 3.5 Flash in a class of its own. We spent the last 6 months making sure Flash is great for real world use cases. It's available everywhere now!

Karpathy will scale Pre-training research at Anthropic.

This is coming from a place of frustration. I am getting to know about the Open Source India Summit today, where I cannot apply anymore as a speaker. I work at Hugging Face, which has done a lot for Open Source. Are the board members oblivious to the presence of HF in India? If yes, we are definitely doing something worng. Tagging the Indian workforce here @RisingSayak, @adarshxs, @adithya_s_k, @_DhruvNair_, @yvrjsharma PS: I dare you to pick GitHub profiles of any one of them.

Is this real !!!!!!!!!!!!! congratulations @karpathy

@karpathy which part are you interested in working @karpathy any interests in mechanistic interpretability alignment?

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

@sentra_app just killed @GoogleResearch's TurboQuant. SpectralQuant — 5.95× KV cache compression on Mistral 7B at +7.5% perplexity overhead. TurboQuant at the same compression: +22%. 3× less degradation. 15-second calibration. One per-model, then drop-in for any HuggingFace LLM, ViT, ESM, AlphaFold Evoformer, or VideoMAE. Check out the findings and how the mechanism works below. ↓

@AIforOperations @HeyGen is it opensource?

Built LeadFlow for the @HeyGen hackathon Drop in a LinkedIn URL or company name → a personalized outbound video tailored to that brand with their actual stats referred to in the video. Also, have options to add your own AI avatar. How it works: - Scraper gets the company's public data - Claude writes the script - HeyGen Avatar V for avatar video - HyperFrames motion graphics to discuss company output Example output: