Geometry-Guided Reinforcement Learning for Multi-view Consistent 3D Scene Editing paper: https://t.co/UWIVfI8dTU https://t.co/pViK4yPby2

Open-source video generation is progressing at lightning speed! LTX-2.3 is amazing 🔥 - Text-to-video & Image-to-video - Audio generation & sync - One single 22B model - Get a huge daily quota with your HF PRO subscription (25 min of H200/day!) Try it now on Hugging Face ⬇️ https://t.co/oV6CjpX0mI

Cosmos Policy just dropped for physical #AI 👀 Cutting edge research turning a world foundation model into a unified robot brain that can see, predict, and act—no extra action heads, no complicated control stack. Read our blog on @huggingface ➡️ https://t.co/zH98YwngeX https://t.co/nfmBg9pdHc

NEW: Elements of AI Agents. Our first text-based AI course. If you are looking for an entry point to building AI Agents, check it out now. FREE to enroll now. (bookmark it) Audio available so you can listen on the go. Enroll here: https://t.co/HHXg8rr0Ew A lot more to come.

AI isn’t just automating tasks. It’s reshaping how people spend their time at work. Developers using AI tools are spending less time on project management and administrative work and more on actual coding. The biggest boost is among junior developers, suggesting that AI may accelerate learning rather than replace entry-level talent. https://t.co/NzKNpsP8pC @MITSloan @Brian_Eastwood

BREAKING: Swedish start-up Lovable sees revenue jump from $300M ARR to $400M ARR in a single month. Ryan Meadows, Chief Revenue Officer @Lovable says annual recurring revenue has surged by more than 30%, from $300 million to $400 million in a single month, and could top $1 billion by year's end. "It's accelerating quite a bit," Meadows said. "We've doubled the number of active users daily just in the last couple of months." - Meadows

https://t.co/xEpwOLG3RC

https://t.co/xEpwOLG3RC

a touch of delicacy❣️ Grok Imagine, video extension up to 30 sec https://t.co/AUUZrrj3ea

Grok Imagine 1.0 is insane. Imagine V1.5. https://t.co/SxJeSitXL1

being a *literal* springboard for a *literal* terrorist while verbatim saying "we welcome everyone into this city" is poetic perfection https://t.co/nqJJ4Fx2qI

Also I’m going to wait on sending out a dollar to the Round 2 accounts. I’m going to randomize the 166 invitations for an Elon dollar giveaway. As you know Elon sent me more money; $69 to be exact. 👇🏻 So you will either get an Elon dollar or dollar. https://t.co/4ZPYbLJPlZ

Memory Skill for OpenClaw with 26k+ users in 1 week🚀 OpenClaw's memory system is broken by default. It requires curating massive MEMORY.md files or relying on duplicate-heavy generation. Hours are wasted tuning, and massive amounts of tokens are burned. It's time to stop. So we built the memory skill to solve that prob Here is our superpower ⚔️ 🎯 Top #1 market accuracy (92.19%) after 8+ months of intense architecture iteration 🧠 The ultimate solution to keep the timeline, facts, and meaning perfectly in place ☁️ Local & Cloud + Version control ⚡ Super easy setup

Here’s a first look at X-Plane 12 on Apple Vision Pro! With visionOS 26.4 and NVIDIA CloudXR 6.0, the simulator streams wirelessly at up to 4K/120fps to your headset. And if you have a physical yoke or throttle, ARKit uses image detection to recognize them and place them inside your virtual cockpit. 🤯 It’ll be available later this spring.

Democrats are weaponizing the Senate’s rules to block the SAVE America Act, defund @DHSgov, & hurt the American people to spite @POTUS.   @SenateGOP can either let Dems keep obstructing & smash the filibuster the first chance they get, or we can act now & use the mandate the American people gave @realDonaldTrump & Republicans to secure our elections, protect our homeland, & bring back common sense.   Democrats started this fight. I support whatever changes to Senate rules are necessary to finish it and pass the SAVE America Act. https://t.co/MmuCF7kzT2

Note: Claude Code invalidates the KV cache for local models by prepending some IDs, making inference 90% slower. See how to fix it here: https://t.co/tUDs5Q8Jt5 https://t.co/5qvwOAArjX

We created a repo with 250+ notebooks for LLM training. Train locally on your device with 3GB VRAM or free on Colab. Learn the entire fine-tuning and inference workflow. Supports RL, vision, audio, embedding, TTS models GitHub: https://t.co/1qSdX8w8SO https://t.co/B0T52LOJ79

ClawVault – a persistent memory for AI agents It gives agents a markdown-native memory system that: - stores knowledge in a graph-aware vault - is local-first (no cloud) - preserves state with checkpoints, search and structured facts - is human-readable + git-friendly So agent memory becomes a plain markdown (instead of vector databases), that you can read, edit, and version

Every AI video tool I've used before treated the creative process as a single moment. You write a prompt, you get a clip, and if it's not right, you just... start over. The whole thing. From scratch. PAI (got to try it early) is one of the best long-form video storytelling models I've used that really attempts to fix this problem. You can go back into a scene and change things. Move a character. Adjust the pacing of a shot. It sounds like a small difference, but it completely changes how you think about what you're making. It finally feels like I have editing control over the video generation process. Excited to see what @UtopaiStudios is building here.

The Ch08 Nb on distilling LLMs is now on GitHub: https://t.co/bPRyIU5BhH Hard distillation that works with any LLM (minding the terms of service, of course). https://t.co/KscPulkj7q

A self-evolving framework to discover and refine agent skills. Most agent skills I see today are hand-crafted or poorly designed by an agent. Multi-agent systems for building skills look promising. This paper introduces EvoSkill, a self-evolving framework that automatically discovers and refines agent skills through iterative failure analysis. EvoSkill analyzes execution failures, proposes new skills or edits to existing ones, and materializes them into structured, reusable skill folders. Three collaborating agents drive the entire process. An Executor that runs tasks, a Proposer that diagnoses failures, and a Skill-Builder that creates concrete skill folders. A Pareto frontier governs selection, retaining only skills that improve held-out validation performance while keeping the underlying model frozen. On OfficeQA, EvoSkill improves Claude Code with Opus 4.5 from 60.6% to 67.9% exact-match accuracy. On SealQA, it yields a 12.1% gain. Skills evolved on SealQA transfer zero-shot to BrowseComp, improving accuracy by 5.3% without modification. I will continue to track this line of research closely. I think it's really important. Paper: https://t.co/mgsnoMBjOx Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX

“For the immediate future, and perhaps for a long way ahead, the continuity of our culture may have to be maintained by a very small number of people.” T.S. Eliot https://t.co/KE0yEbD75j

Reasoning-Aware Retrieval for Deep Research Agents Deep research agents generate explicit reasoning before every search call. These reasoning traces encode rich signals about search intent and problem-solving context. Yet no existing retriever learns to exploit them effectively. This paper introduces AgentIR, a reasoning-aware retrieval system that jointly embeds the agent's reasoning trace alongside its query instead of just the query alone. Why does it matter? The agent's reasoning acts as a retrieval instruction, a memory of key history, and an implicit filter for outdated information. All of this context is available for free since the agent already generates it. AgentIR-4B achieves 68% accuracy on BrowseComp-Plus with the open-weight Tongyi-DeepResearch agent, compared to 52% with conventional embedding models twice its size and 37% with BM25. It also outperforms LLM-based reranking by 10% absolute, with no additional inference overhead. Paper: https://t.co/rok5nZDfYw Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

Professors are increasingly worried about what AI is doing to critical thinking. As tools like ChatGPT reshape how students research and write, many academics fear that the humanities and deeper reasoning skills could erode. The challenge now is teaching students to think with AI, not to let it think for them. https://t.co/lufNjlK7uA

Efforts to improve the security of AI agents should recognize that many security failures occur even in the absence of adversaries. The unreliability issue has largely flown under the radar and there hasn't been much work on defining, measuring, or mitigating the problem. More on this in our response to NIST's request for information on AI Agent Security, by @steverab, @sayashk, @PKirgis, @CitpMihir, and me: https://t.co/PW7DJZpDWV This is based on our recent paper: https://t.co/FI5kuBkdRZ

We just added /btw to Claude Code! Use it to have side chain conversations while Claude is working. https://t.co/hjO3YqvrPr

Introducing The Anthropic Institute, a new effort to advance the public conversation about powerful AI. https://t.co/M7vi9oRuYi

The Anthropic Institute is hiring. You can learn more about our work and priorities at the link below: https://t.co/xoyjNFcSlY

The idea of AI agents talking to each other is moving into the mainstream. Meta has acquired Moltbook, a social network designed for AI agents, and its founders will join the company’s AI research unit. As agentic AI grows, platforms built for machine-to-machine interaction may become a new layer of the internet. https://t.co/JT0MLjJJOG