We compared AI observability pricing at production scale. The cheapest competitor costs 8x more than Logfire. The most expensive, 100x. https://t.co/MWQxxxXK57 If you're building production AI applications and paying more than $2 per million spans, run the numbers.

@petergyang 7. fashion, expression, sophistication 10/10 https://t.co/OkCKyHMsYT

+2 more

Code for our new world model planner is live! https://t.co/PnCk0OzqTl Includes our implementation on dino-wm, as well as implementations on jepa-wm and le-wm, and minimal pseudocode for anyone to re-implement themselves.

Code for our new world model planner is live! https://t.co/PnCk0OzqTl Includes our implementation on dino-wm, as well as implementations on jepa-wm and le-wm, and minimal pseudocode for anyone to re-implement themselves.

🚀 LeWorldModel datasets & checkpoints are now available on Hugging Face! https://t.co/aiBkDTsNyX You can plug them directly into stable-worldmodel (https://t.co/2eQB7Q0l9i), the engine behind LeWorldModel, to instantly load, run, and start building on top of our models.

🚀 LeWorldModel datasets & checkpoints are now available on Hugging Face! https://t.co/aiBkDTsNyX You can plug them directly into stable-worldmodel (https://t.co/2eQB7Q0l9i), the engine behind LeWorldModel, to instantly load, run, and start building on top of our models.

In what world is it okay for the Defense Secretary to make multi-million dollar investments in defense companies: a) Ever; b) Just before launching a war. FT has the scoop: https://t.co/ulRcXTH2Vp In a functional democracy, he would offer his resignation tonight. https://t.co/013BL2pZEB

As Yann LeCun and his team has brought a new wave of World Models research into 2026, we have compiled a reading list to recap some of the must-read papers to get you up-to-date about JEPA & World Modeling! https://t.co/ylgwqtRbaH

US removes 3 from Global Magnitsky and 2 from Russia sanctions lists. Absolute disgrace! https://t.co/70ZPEowrcX

US removes 3 from Global Magnitsky and 2 from Russia sanctions lists. Absolute disgrace! https://t.co/70ZPEowrcX

i didn't know codex had that dawg in him https://t.co/1LmVXlgfTo

Review issues Address feedback Commit changes Open pull requests @GitHub plugin in the Codex app. https://t.co/bNILtODGLF

Review issues Address feedback Commit changes Open pull requests @GitHub plugin in the Codex app. https://t.co/bNILtODGLF

DMCAs for Claude code source code are going out. https://t.co/rz8mirq08k

BREAKING MASSIVE NEWS President Trump just signed an Executive Order to provide for election integrity: 1) Election officials will be provided w/ federal data to ensure that ONLY American citizens are able to vote 2) Orders the Postmaster General & the United States Postal Service to verify that ballots are ONLY mailed to eligible voters & ballots are returned ONLY by eligible voters. This is a huge step in the right direction. Pass the SAVE America Act!

@9th @thecosmos I feel seen https://t.co/FuL8epDyyy

Odessa A’zion stars in new campaign for Cosmos. https://t.co/lBC82l3mkr

Paper review: LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels https://t.co/TpFFnwPWkc Nice clean github: https://t.co/HOuqEf0HaF This is the application of the LeJEPA results to world models, trained offline on experience from three different robotics style tests with one to two million steps in each dataset. Re-states the benefits of the SigReg loss relative to prior world model approaches. Uses ImageNet standard 224x224 RGB pixel input images with an unmodified ViT-Tiny vision transformer from HuggingFace to generate latents. One extra post-projection step is needed to give SigReg the necessary freedom to perturb the latents into independent gaussians, since ViT ends with a layernorm’d layer. Also tested with ResNet-18, which still performed well, but slightly worse. Uses a 192 dimensional latent. Performance slightly dropped when doubling the latent size to 384; it would be nice to know if it was stable there, or if it continued worsening with excessive latents. There is a relationship between batch size and SIGReg, the larger latent may have improved performance if the batch size was increased. The predictor is implemented as a ViT-S backbone – Why a vision transformer when the latent is flat? Uses a history of 3 sets of latents for two of the benchmarks and 1 for the other. Performance was markedly better with the “small” ViT model than the “tiny”, but the larger “base” model degraded notably, which is interesting. Dropout of 0.1 on the predictor significantly improved performance. 0.2 was still better than 0.0, but 0.5 was worse. Trained with a batch of 128 x 4 trajectories. I wish their training loss graphs were more zoomed in with grid lines. Performs planning at test time instead of building a policy by training in imagination like Dreamer / Diamond. Rolls out 300 initially random sets of actions up to a planning horizon H of 5 (at frame-skip 5). Iterates up to 30 times using the Cross Entropy Method (CEM). The main paper body mentions using Model Predictive Control (MPC) strategy, where only the first K planned actions are executed before replanning, but appendix D says they execute all 5 planned actions. After training, they probe the latent space to demonstrate that it does capture and represent physically meaningful quantities. They also implement a decoder from the latent space back to pixels – not used by the algorithms, but helpful to see what things the latent space is actually representing. They tested incorporating the reconstruction loss into training, but it hurt performance somewhat. They wound up with a 0.1 lambda for SigReg, as opposed to 0.05 in the LeJEPA paper. 1024 sigreg projections, but observe the number has negligible impact I like the JEPA framework, but so far my attempts to use it on Atari games with value functions have not matched my other efforts.

Apple's crackdown on vibe coding apps puts it on the wrong side of history vibe coding is the biggest democratization of software since the App Store itself. But Apple is blocking the tools that make it possible.... the same week it turns 50. The irony is that Apple should be leading this moment and instead its holding it back w/ @jaswu_

wow, blocking mobile phone internet for just 2 weeks: - reversed 10 years of brain cognitive aging - reduced depression MORE than antidepressants before taking any drug, go amish for a week https://t.co/GFfaTD4OYL

Holy smokes! Opus 4.6 set a new record on the Remote Labor Index! At 4.17%. Anyone who claims that we are close to AGI is either lying or lost. https://t.co/OAb6YsX4Qk

ICYMI - R3 Bio working to grow headless human bodies to harvest organs "for research," as Trump phases out federal animal experimentation: "If we can create a nonsentient, headless bodyoid for a human being, that will be a great source of organs," says investor Boyang Wang. https://t.co/VTpm84wyDd

📢Seen2Scene Real-world 3D is incomplete, typically requiring training on synthetic scene data. @QTDSMQ introduces visibility-guided flow matching, enabling training on real partial scans for scan completion & text-to-3D scene generation! Check it out: https://t.co/jqZ164QX0W https://t.co/GV5Q4ZvBPj

Oh shit, it seems like all the HF Research team pretraining data has been accidentally leaked to the public. The web, PDFs, and synthetic datasets are expode on hf FineData org... Apparently, an intern used CC to push the data with private=False. https://t.co/gUcaKQBzxi

Tesla robotaxis are sometimes driven remotely, by humans. They just admitted it. https://t.co/hLAaUXqBsQ

AMD Advancing AI 2026 event to take place on 22–23 July: here’s what to expect. https://t.co/u7goYTiL7r

AMD Advancing AI 2026 event to take place on 22–23 July: here’s what to expect. https://t.co/u7goYTiL7r

OpenClaw 2026.3.31 🦞 🇨🇳 Bundled QQ Bot — private, group, and guild chat + media 📹 LINE now sends images, video, and audio 🧵 Real background task flows: list, show, cancel 🇯🇵 Better CJK: context, memory, and TTS OpenClaw's next release has been leaked🦞https://t.co/EntA23WzSz

Psalm 34:1 - "I will bless the LORD at all times: His praise shall continually be in my mouth." We investigate whether injecting biblical Psalms into a large language model's system prompt produces measurable changes in performance on standardized ethical reasoning benchmarks. https://t.co/wryEK6WLA3

People don't understand how insane the explosion of Christianity was: - There were maybe a thousand Christians at Jesus's death. - In 100 AD there were still only about ten thousand Christians. - One hundred years later it was still less than one percent of the Roman empire. - And then a hundred years later there were over SIX MILLION Christians, more than 10 percent of the population. - Go forward another 150 years and there are over 30 million Christians and it is the official state religion of the Roman empire. It's one of the most stunning transformations in world history. And yet there is wide disagreement on why it occurred.

@PatrickMoorhead @wholemars I know. But it knows more about me than I know about me. Which is a freaky thing. :-) "Hey Grok simulate a conversation between Patrick Moorhead and Robert Scoble about the future of privacy and the social contract when it comes to AI-driven glasses." Here you go: https://t.co/j4PWh4KMeD Sounds pretty much like both of us, except I have @evenrealities and @getVITURE and @MentraGlass here. :-)

Apple vs vibe coding. The kicker: https://t.co/XocVdcBITc