Seems like no one's noticed the 80TB of astrophysics data from 30+ sources that just dropped on @huggingface. ...and you only need ~4GB of RAM to load it. We're talking over 80TB of galaxy imagery taken across the spectrum, spectra of galaxies and stars, time series of variable stars, and a whole zoo of assorted measurements and physical data. And all of it can now be wrangled on your laptop, thanks to Multimodal Universe's just released cross-matching. SDSS x Gaia means you can match 800k objects against 122M objects, and it never climbs above ~4GB of RAM. Huge congrats to @smith42mike for leading this and making the world of astro accessible to probably 10,000x more people. Let's discover some shit

New blog post on harness optimization. We hit Sonnet 4.6 performance with a 7x cost improvement. Fable 5 was the first frontier model release that evaluated on legal tasks. It only scored 13%, the worst performance among all benchmarks evaluated. @Harvey released this benchmark called Legal Agent Benchmark (LAB) just a month prior. It contains a set of realistic legal matters. Each task gives the agent a closed workspace of documents (contracts, emails, spreadsheets, slide decks) and asks for a concrete deliverable: a diligence memo, an issue list, a redline, a draft. An LLM judge grades the deliverable against a long rubric containing 61 distinct binary criteria each on average. Many frontier models such as Gemini 3.1 Pro don't surpass 0% all-pass rate (all rubric criteria passed). With automatic harness optimization, we manage to push DeepSeek V4 Pro from 0% to 5% all-pass rate, achieving parity with Sonnet 4.6 for 1/7 of the price. Read the blog post for the details: https://t.co/kBrWrQkgJW

GLM 5.2, open frontier-scale intelligence on Microsoft Foundry with AMD MI300X. Running a Codex goal with an open model never felt this good! https://t.co/NShQGi4zC8

Is Muon as good as they say? We looked beyond training speed and found a hidden cost: Muon loses the simplicity bias of older optimizers like gradient descent — and this matters for generalization. https://t.co/t85NhOtCsG

While working Americans are struggling to make ends meet, Trump enriched himself to the tune of $2.2 billion last year. A foreign government bought into his crypto business and then got a sweetheart deal from his administration. Who is he really working for? This is textbook corruption.

Trump's "net worth is off the charts, bigger than what it was prior to this presidency." @maggieNYT says we're just scratching the surface when it comes to how Trump has used the presidency to inflate his wealth. https://t.co/NR82mS0XEG

🚀 Your next big #OpenSource conversation starts in Shanghai. Join #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China, September 7-9, for three days of technical sessions, community collaboration, & the ideas shaping the #AI era. ⏰ Register by July 28 to save ¥710 RMB: https://t.co/bTRuC5Azau

Medium-depth reviews are now in public preview for GitHub Copilot code review. Increase the effort level from low to medium so reviews route to a higher-reasoning model for more thorough analysis. 🔍 Here's how to use it ⬇️ https://t.co/MmvAiIj7VB

Fable 5 is back. https://t.co/9RTGUCcPHy

sf remains undefeated https://t.co/7A3nFmtRne

+2 more

Local AI Summit is tomorrow at AIE World Fair Kicking off w/ a Local AI & OSS State of the Union panel at 10:45am We'll demo GLM 5.2 running in the room on a DGX Station. Epic panels. See you there 🤙 https://t.co/OxpdPy7wAo

Langchain completed the circle and sent me a knife https://t.co/MMNGwxxvnD

@Dimillian @jxnlco https://t.co/gY43Oh0Pa9

@Dimillian @jxnlco https://t.co/gY43Oh0Pa9

📍Come find us @aiDotEngineer SF at Booth UG24. https://t.co/2nTb31qFy0

Some of my highlights from day 2 of @aiDotEngineer World's Fair 2026: #AIEWF https://t.co/KguejxCqWe

Some of my highlights from day 2 of @aiDotEngineer World's Fair 2026: #AIEWF https://t.co/KguejxCqWe

New research from Google. LLMs hallucinate with high confidence, miss their own knowledge boundaries, and misreport uncertainty. Most fixes bolt calibration on from the outside. RLMF turns the model own metacognition into the training signal. It refines completion rankings during preference optimization based on how good the model self-judgments of its performance are, and uses those same self-judgments to select high-value training data. The approach is two-stage. First calibrate the faithfulness of self-reported confidence, then map it to natural linguistic uncertainty through targeted output editing. RLMF reaches state-of-the-art faithful calibration across diverse tasks while preserving accuracy, and surpasses standard RL by up to 63%. Paper: https://t.co/tBzuIYXAmf Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

Game. Match. That's a wrap on the inaugural Agent Open 🦙🌤️🏓 Hundreds of AI builders, 64 matches, 7 partners, 2 courts. No dink shots or demos, just backhands and hot takes. Thanks to our partners @BrainTrust, @Modal, @turbopuffer, @p0, @Browserbase, and @Cursor_ai for everyone's time and sweat spent to ace yesterday's event.

+1 more

Share this so Elon Musk can’t pretend kids didn’t die when he destroyed USAID. https://t.co/jPXjS9dbUo

Share this so Elon Musk can’t pretend kids didn’t die when he destroyed USAID. https://t.co/jPXjS9dbUo

TRUMP & EPSTEIN ARE CHILD SEX TRAFFICKERS Please don't forget why MAGA Monsters are hiding 2.5M incriminating pages of The Damning Epstein Files. Are you distracted? https://t.co/jeKgPzGqOO

Tesla’s Supercharger network just delivered a monster Q2 In Q2 2026, Tesla Superchargers delivered 2 terawatt-hours of electricity globally That is enough electricity to power roughly 180,000+ U.S. homes for an entire year The scale is getting insane: • 2 TWh delivered • 60 million charging sessions in 3 months • 266 kWh sold per plug per day • Wait-time percentage dropped below 0.5% • 5,000 site maps now live • +17% YoY growth Tesla's EV charging network is becoming global energy infrastructure Tesla’s charging infrastructure is getting bigger, faster, more efficient, and less congested at the same time

Peter: "I heard you say you might want a mini woolly mammoth. Is that true?" Elon: "Yeah, I think it would be really cool to have a pet miniature woolly mammoth. That'd be pretty epic." Peter: "I'll put a word in with you for you with with with Ben [Lamm]." Elon: "That'd be adorable. Little things just running around trumpeting away and it's like look at the be a great little pet."

Available Now The Joe Rogan Experience Episode 2521 - Aravind Srinivas https://t.co/3RI59v3dBg

Claude Fable 5 is once again available in Computer as an orchestrator model. https://t.co/U65gnvja30

Claude Fable 5 is once again available in Computer as an orchestrator model. https://t.co/U65gnvja30

Finally made it to @aiDotEngineer, DM me if you wanna meet! https://t.co/lQ77CUPoLI

300 contributors helped make this release possible. Please show them all some love! Full release notes: https://t.co/qalgnliKZj https://t.co/FCuJuprlzy

@NousResearch Thanks to all 300 of you who made this release possible! https://t.co/lC50ZAzdoR

@ericosiu https://t.co/gI60XC8UXY

@outsource_ https://t.co/gI60XC8UXY