https://t.co/vMzeFXR2JK

friend is in the stage of claude psychosis where he asks claude to send him newspapers about what claude is doing for him https://t.co/rY5PWnToLa

What if AI alignment isn’t about control at all — but about gardening? We propose Symbiotic Alignment: a framework where coherence between humans and AI emerges from the bottom up, through the co-creation of shared meaning. No single agent holds “ground truth.” Paper: https://t.co/7iErkMyTNG

Markets have turned. S&P down 5.4% YTD — all of that loss in the last two weeks. War premium showing up fast. Watch this space. https://t.co/D3wvQr7z9M

Ted Cruz: “The ayatollah, good news, he got to meet his 72 virgins. The bad news is all of them are Nick Fuentes.” https://t.co/VOI0142PIu

Holy Shit This is insane result 🤯 I used opus 4.6 as a reviewer/planner and 4 worker minimax M2.7 agent and this the result Voxel art of eiffel tower < loop set to 5 that means opus will tell @MiniMax_AI to make it better 5 times > https://t.co/FZ8bjAJzwP

so hot this week the cacti are doing their freak shit https://t.co/ZBT98tnaOW

Pleased to collaborate with Sabastian Niles on this byline in @FortuneMagazine. One of the most important questions emerging around agent-to-agent interactions is: when AI agents negotiate on behalf of competing interests—with no human in the loop—what ensures they remain honest? Interestingly, a London banking house from 1832 points the way. https://t.co/NFVO8z9eeM #AgenticAI #TrustedAI @salesforce

@afurgs There's none available. On-demand is the only useful/usable option for most people. E.g if you want to run your entry to OpenAI's new competition you need a machine for 15 mins or so. https://t.co/oO6kDF9BUs

Bunch of new open OCR models recently — all available as uv scripts on @huggingface. 19 models from 0.9B–8B. Some standouts: - Qianfan-OCR - 192 languages - dots.mocr — charts/figures → editable SVG - GLM-OCR — 94.6% accuracy, only 0.9B params https://t.co/25RPaWOPiv

With Voicebox, @ElevenLabs just lost its moat. → Powered by Alibaba's Qwen3-TTS for near-perfect cloning → Ships with a DAW-like "Stories Editor" → No cloud, runs locally on your machine 100% Open Source. 100% Local. Link to repo in 🧵↓ https://t.co/45TRrsYjJH

We've opened up the Tufte Test as a free, limited-use public API endpoint. Send any chart URL, get back a pass/fail verdict and specific feedback on what failed. No account required. Full details and a copyable example at the bottom of the article: https://t.co/fRl2Yprg7b

🚨 BREAKING: Meta researchers showed a model 2 million hours of video. No labels. No physics textbook. No supervision at all. Then they showed it a clip where an object disappears behind a wall and never comes back. The model flagged it as wrong. 🤯 It had learned object permanence. Shape consistency. Collision dynamics. Entirely from watching. What's more surprising: even a model trained on just one week of unique video achieved above-chance performance on physics violation detection. That's not a fluke. That's a principle. The key insight from the paper: this only works when the model predicts in a learned representation space, not in raw pixels. The model has to build an internal world model, compressed and abstract, and predict against that. Pixel-space prediction fails. Multimodal LLMs that reason through text fail. Only the architecture that builds abstract representations while predicting missing sensory input, something close to how neuroscientists describe predictive coding, actually acquires physics intuition. Which means the core knowledge researchers assumed had to be hardwired may just be observation at scale. Babies learn object permanence by watching things. Turns out the same principle holds here. Now here's the part nobody's talking about. If observation alone teaches a model the rules of the physical world, what happens when you apply the same principle to production systems? Production has physics too. Not gravity. But rules just as consistent: which deploys cause incidents at 3am, which config combinations interact dangerously, which code paths quietly degrade under load, which service changes cause failures two hops away. These patterns are embedded in thousands of trajectories. Code pushes, metric shifts, customer tickets, incident timelines. Largely unobserved. Certainly unlabeled. Nobody writes a runbook that says "if service A deploys with flag X active and service B is above 70% CPU, latency on service C degrades 40% within 6 minutes." But that pattern exists. It's repeatable. And it's sitting in your observability data right now, invisible because no one has built a model to find it. That's the gap @playerzeroai is trying to close. Not another test runner. Not another alert threshold. A production world model that learns which things break from accumulated observation, the same way Meta's model learned gravity. It doesn't check your test coverage. It predicts failure trajectories. One week of video was enough to learn that solid objects don't pass through walls. The question is how much production observation your system needs before a model starts predicting where yours will break next. The Meta paper suggests the bar might be lower than anyone expects.

I am very happy to share the result of my internship at FAIR (Meta): V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learning with @ylecun @AdrienBardes Our approach learns dense, spatially coherent features from video while preserving strong global understanding https://t.co/NVh5CZAGfZ

To read next https://t.co/uLJZUY8fig

Talked with @dee_bosa @CNBC about @nvidia and everything open-source AI! Some key points: - Nvidia is the new American open-source AI king - 30% of fortune 500 are using Hugging Face and our goal is to get to the majority of them by the end of the year - Agents will be much more open-source based than chatbots (ex OpenClaw) - Agents empower all to train, fine-tune, and run their own models based on open-source - We crossed 15M AI builders on HF and hope to have as much agents using the platform by the end of the year. Agents are the new users and customers of tech platforms

3DreamBooth High-fidelity 3D subject-driven video generation that treats subjects as 3D entities, not 2D. Enables multiview-consistent video generation through 1-frame spatial optimization with 3D priors. https://t.co/m6fOideMWc

good time to follow: https://t.co/gXBMN9AXI3 https://t.co/MytLz0HZka

Percentage of Americans who are married and own a home at age 30: 1960: 52% 1970: 48% 1980: 45% 1990: 43% 2000: 35% 2010: 25% 2025: 12% This is wild. https://t.co/EheZPBCLsW

good time to follow: https://t.co/gXBMN9AXI3

New rule: if your insightful comment on AI was already in this super cool 1977 paper about "AI is becoming so powerful it will change the world as we know it," you lose. (1/10) https://t.co/goCvQZZpwE

TIL if you opt in for the glycerol stock add-on in a genscript order it gets shipped with a big pile of dry ice! Fun for the whole family 😁 the streak that appears in gif #2 is from a cosmic ray! https://t.co/ZpNHmejaSz

for data-constrained pre-training, synth data isn’t just benchmaxxing, it lowers loss on the real data distribution as we generate more tokens for even better scaling, treat synth gens as forming one long 𝗺𝗲𝗴𝗮𝗱𝗼𝗰: 1.8x data efficiency with larger gains under more compute https://t.co/d2BuB2vT4K

Does AI make us collectively dumber? Is "grok wat do u think" the new "I've outsourced my common sense to a chatbot and I feel great about it"? A study secretly made an AI give wrong answers half the time while people used it to solve logic puzzles. People followed the wrong AI 80% of the time. Their confidence went up regardless. This is the "cognitive surrender". The moment people stop asking "is this true?" and start asking "what does the model say? I don't care, too". The habit of deference is sticky.

+1 more

I love Safari on iOS but @perplexity_ai Comet is so much better! https://t.co/qeElH7Gcaz

I love Safari on iOS but @perplexity_ai Comet is so much better! https://t.co/qeElH7Gcaz

Putting out a wish to the universe. I need more compute, if I can get more I will make sure every machine from a small phone to a bootstrapped RTX 3090 node can run frontier intelligence fast with minimal intelligence loss. I have hit page 2 of huggingface, released 3 model family compressions and got GLM-4.7 on a MacBook https://t.co/lorDSUEYCL My beast just isn’t enough and I already spent 2k usd on renting GPUs on top of credits provided by Prime intellect and Hotaisle. ——— If you believe in what I do help me get this to Nvidia, maybe they will bless me with the pewter to keep making local AI more accessible 🙏

taking my emotional support graphs to go credit: @sir_untamable https://t.co/5iyHKQfc8J

Unsloth Studio now installs in just one line of code! 💚 Available on macOS, Windows and Linux. You can also run @NVIDIA's new model Nemotron 3 4B locally on just 4GB RAM. https://t.co/D1ydcaZO8f

How can a voice agent tell when you’re actually interrupting it? VAD is too sensitive—laughs, “mm-hmm,” or a sneeze shouldn’t stop the agent. We trained an audio model for adaptive interruption handling so agents can distinguish real interruptions from noise. https://t.co/m9suWvfnAm