⭐ How can we set up LLM pretraining to improve the model’s ability to learn new data upon further training? In this new preprint, we find that weight decay during pretraining helps!
Preprint: https://t.co/ObxgFCtOio
Thread below🧵⬇️
🚨 Just aired!
I had the opportunity to speak with #CBS News about the latest in AI text detection, a topic that's critical from education to online communication.
🎥 Watch the full segment here: https://t.co/XW2OMwSpbf
You can’t just do things. You need money to buy enough compute and a life free of visa-stress to build anything these days. Stop peddling this tech-bro bs. 😒
Google's AI just made math discoveries NO human has!
—Solved optimal packing of 11 and 12 hexagons in hexagons.
—Reduced 4x4 matrix multiplication from 49 operations to 48 (first advance in 56 years!)
and many more.
AlphaEvolve is the AlphaGo 'move 37' moment for math. Insane.
AlphaEvolve, our new Gemini-powered coding agent, can help engineers + researchers discover new algorithms and optimizations for open math + computer science problems.
We’ve used it to improve the efficiency of our data centers (recovering 0.7% of our fleet-wide compute resources on average). We’re also using it in chip design and to speed up Gemini’s training, the very models underpinning AlphaEvolve itself — an exciting flywheel of progress!
Every international student in the US is living in utter fear.
At Harvard and Northwestern last week, I heard
—We should Uber to avoid a speeding ticket
—I stopped partying to avoid noise complaints
—We took debates out of the conference to avoid conflict
It's really bad.
Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date.
For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation.
BREAKING DeepSeek has #1 best non-thinking LLM.
— Better (beats or ties GPT4.5, etc)
— Cheaper, by 100-200x ($0.27/1.10 vs $75/$150 for 1M input/output toks)
— Faster, by 5x (60 tok/s vs ~12 tok/s)
— Smaller (685B MoE vs 2T??)
— Free to distribute (MIT license)
— Open source
Wow. AI agents are here
I've been using Manus AI the last week and it actually is insane
While at dinner I prompted it to build me a full app. By the end of dinner the app was done
In this video I walk through Manus and show you how to build incredible apps (ya, bookmark this)
The Time Has Come for Robots.
I build AI Agents to replace office workers, but these demos convince me! All physical labor will be gone to robots, too. (even the world's oldest profession).
Just watch it if you disagree. The biggest robot thread ever (50 demos):
One of the features I've mosted wanted in AI Studio for a long time! Just paste a YouTube link into the command line and ask Gemini 2.0 questions about it - it's multimodal understanding is kind of mindblowing. Try it here: https://t.co/3ioM7tihCG
Manus, the new AI product that everyone's talking about, is worth the hype.
This is the AI agent we were promised.
Deep Research+Operator+Computer Use+Lovable+memory.
Asked it to "Do a professional analysis of Tesla stock " and it did ~2wks of professional-level work in ~1hr!
There's a dangerous app that helps you cheat on Leetcode style software engineering interviews without being detected.
Don't use it.
Please don't engage in cheating. It's morally reprehensible. Don't do it. I'll leave a link here just so you're aware what NOT to use:
Introducing our AI co-scientist, a multi-agent AI system built with Gemini 2.0.
We think of it as a virtual collaborator for scientists, using advanced reasoning to synthesize a huge amount of literature, generate novel hypotheses, and suggest detailed research plans. We’re seeing promising early results in important research areas like liver fibrosis treatments, antimicrobial resistance, and drug repurposing. As a next step, we’re opening up a trusted tester program for scientists around the world.
Excited to introduce the Perplexity Deep Research Agent: available for free to all users. Paid users only need to pay $20/mo to access an expert level researcher on any topic for 500 daily queries, and need to wait less than three minutes for getting a full research report.
today we launch deep research, our next agent.
this is like a superpower; experts on demand!
it can go use the internet, do complex research and reasoning, and give you back a report.
it is really good, and can do tasks that would take hours/days and cost hundreds of dollars.
Today we are launching our next agent capable of doing work for you independently—deep research.
Give it a prompt and ChatGPT will find, analyze & synthesize hundreds of online sources to create a comprehensive report in tens of minutes vs what would take a human many hours.
Powered by a version of OpenAI o3 optimized for web browsing and python analysis, deep research uses reasoning to intelligently and extensively browse text, images, and PDFs across the internet. https://t.co/AJHftUBs4m