Well, this is a first: a ChatGPT-generated story won a prestigious literary prize (The Commonwealth Prize). "Not X, not Y, but Z" sentences everywhere, the "hums" trope, and plenty of other obvious markers of AI writing. A major milestone for AI, at any rate... @GrantaMag https://t.co/U6jWejprFv

vLLM and PyTorch worked together to fix a long-standing aarch64 install headache — as of PyTorch 2.11.0, pip install torch on GB200 / GB300 / GH200 just works. 🎉 What changed: PyTorch 2.11.0 now publishes CUDA-enabled aarch64 wheels to the default PyPI index. No more custom --index-url flags. No more transitive dependencies silently swapping your GPU build for the CPU wheel. New users on Grace Hopper and Grace Blackwell systems can follow the standard install instructions and have vLLM work the first time. In our latest blog, @KaichaoYou (co-founder @inferact, Lead Maintainer @vllm_project) shares the full story: 🐛 A 2024 hackathon bug bringing up vLLM on GH200 🔧 vLLM's in-tree workarounds (use_existing_torch.py and [tool.uv] build-isolation passthrough) 🤝 From GitHub issue to PyTorch Foundation TAC discussion 🚀 The fix landing in PyTorch 2.11.0, driven by NVIDIA and PyTorch core. A great example of cross-project collaboration under the PyTorch Foundation umbrella — and a reminder that boring infrastructure wins compound. Read the full story: https://t.co/JGnJ1X7sxl ✍️ : Piotr Bialecki (@nvidia) — @ptrblck_de, Alban Desmaison (@Meta), Andrey Talman (@Meta), Nikita Shulga (@Meta)

NEW paper from Meta. (bookmark it) It's an agent system that autonomously discovers neural architectures that beat Llama 3.2 at 350M, 1B, and 3B scales, all under a 24-hour compute budget. They get this work by splitting the search into two agents: > AIRA-Compose searches the macro architecture. > AIRA-Design implements the low-level mechanisms. For devs: If one agent in your stack is doing both strategy and implementation, split it. Run a planner that picks the structure and an implementer that fills in the mechanisms. AIRA shows this beats a single end-to-end agent on a real, non-toy search problem. The same split is useful for pipeline assembly, query planning, prompt scaffolding, and tool-use programs. Paper: https://t.co/CYALI6CFjJ Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX

Reachy mini rap battle at @aiDotEngineer Singapore! https://t.co/vic68kneyL

Reachy mini rap battle at @aiDotEngineer Singapore! https://t.co/vic68kneyL

you can just do things at ai engineer singapore @aiDotEngineer https://t.co/zxr6LiDjJ5

you can just do things at ai engineer singapore @aiDotEngineer https://t.co/zxr6LiDjJ5

When a federal agency tasked with protecting all Americans decides to mandate a specific religious worldview, it ceases to protect democracy and begins to threaten it. https://t.co/cNxGJSIBPu

This is the single most important chart. If AI were driving prices, you'd see a cluster top-right. You don't. States with huge load growth (VA, TX, NV, ND, IA) sit at ~0c change in 5y. States with massive price hikes (CA, NY, MA, CT) have basically NO load growth. https://t.co/3s9DS0WdNA

"AI is hiking your energy bill" is the most popular political talking point of 2026. The data doesn't support it. A thread: https://t.co/jkmrdr5MEV

Verdict. https://t.co/3LVgAoYyMI

Verdict. https://t.co/3LVgAoYyMI

NEW paper worth reading. GPT-5.4 nano plus a critic-comparator orchestration loop hits 76.4% on SWE-bench Verified, matching standalone Gemini 3 Pro and Claude Opus 4.5 Thinking. The trick is to select from k=8 weak-model proposals using execution and proof signals. What does this mean? Many of the patches you'd expect from a frontier model are already inside a weak model's top-8 candidates. When you have 8 candidate patches from a weak model, don't ask the model which is best. Run them and verify them. That's enough to match a frontier model's accuracy. The takeaway for AI devs: a weak model's top-k often already contains the right answer. What limits you is the quality of your selector, not the capability of the model. Paper: https://t.co/Gx7j7EP9BM Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

The U.S. faces a chaotic #AI regulatory patchwork with over 1,200 state bills introduced in 2025 and no unified federal framework, researchers @JeffSonnenfeld and Stephen Henriques of @YaleSOM and @GaryMarcus of @CSMAP_NYU write in a recent issue of @FortuneMagazine. The current political and digital information environment and competing legal guidance exposes the dynamic, adaptive technologies to the internet of things – including #misinformation and #disinformation – and hampers efforts by companies and developers to innovate and stay ahead of international competitors and adversaries. The researchers provide a framework that first issues interpretive guidance, then suggests narrow legislation for election disinformation and 'sandbox' situations for high-risk challenges like national security. They detail a three-stage test for each piece of legislation that evaluates national security, harm reduction, product innovation and free competition – and instills confidence and continuity rather than chaos and uncertainty. 👉 https://t.co/6KqEAHVmBx

Start work on your computer, continue your local session anywhere. 📲 Remote control for GitHub Copilot CLI and @code sessions is now generally available. https://t.co/wwSEBd5lqL https://t.co/Yc5R6tBfBl

Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model. https://t.co/N87ojcXlOC

the aie friendship tour continues 🫶 https://t.co/UTODRi1Ogd

the aie friendship tour continues 🫶 https://t.co/UTODRi1Ogd

Since the entire X is here 🤪 https://t.co/Bh75qwlxuW

Since the entire X is here 🤪 https://t.co/Bh75qwlxuW

Anthropic is acquiring @stainlessapi, an SDK and MCP server platform that has powered every Anthropic SDK since the earliest days of our API. Read more: https://t.co/ZQbsZKnicv

Starting today, we're opening our Agentic Dialog Platform to every enterprise builder. Our dialog agents have resolved 1 billion+ customer conversations for clients like FedEx, Unicredit, PG&E, Marriott, Foot Locker, and many more. These aren't easy conversations. They solve problems like: > A patient booking medical transport who needs insurance verified on the spot. > A homeowner calling their utility company about a gas leak. > A cardholder figuring out why their must-have purchase was declined. Standard conversational AI was never built for this. It was designed for chat, adapted for voice later. It generates responses, but can't do what dialog requires: hold context under pressure, navigate ambiguity in real time, and actually resolve problems. So we built a better model. Our proprietary model Raven was built from the ground up specifically for dialog. Agent harness in the weights, not bolted on through prompts that drift under pressure. And in our platform, you can deploy Raven as your default or bring in GPT-5, Claude, Gemini, whatever model fits your use case or regulatory requirement. Now that the Agentic Dialog Platform is open, any team can create, test, and deploy dialog agents on the same model and infrastructure the world’s top brands trust on their hardest days. This opens up the pool of builders across your entire enterprise. The person who knows customers best, who runs operations, who owns the customer journey: they're all builders now. Two ways to build: > Poly Agent Builder: Describe your use case in natural language, and it configures your agent, knowledge base, and conversation flows automatically. Production-ready in ten minutes. > Agent Development Kit (ADK): Developers use this to build dialog agents the same way they build everything else. Use your own IDE, a coding assistant like Claude, version with Git, deploy from your terminal. Get started now: https://t.co/ifZOy1uEBz

i found this poem on @trq212's blog and i liked it very much. just what I needed after multiple weeks of studying and stressing and ruminating on my shortcomings. today I got some pho and ate it sitting on the ground in a park. it was a cool and lovely evening. https://t.co/6WNAgwi5GO

post @aiDotEngineer tour of Singapore tiny AI hub: @Lorong_AI Welcome @ZixuanLi_ and the @Zai_org team Hope you enjoy your trip to Singapore and the food https://t.co/o3rJ3A1wZR

Join us this Saturday at the Google I/O Hackathon! https://t.co/AMwHfqLQkS

Join us this Saturday at the Google I/O Hackathon! https://t.co/AMwHfqLQkS

Phew, Grok Build is really thorough, pretty incredible. Relay feature for PasteLocal is done, pushing to Github now. And if you want to know what this adds, here's a bit more about it, this is a feature I really wanted in there, and honestly didn't really expect to have done this week. • Per-peer E2E encryption: clipboard data is encrypted individually for each paired device using X25519 + HKDF + AES-GCM (no plaintext ever leaves the client) • Durable persistence: relay state is now stored in an atomic state.json with proper TTL handling and compaction. Pending clips survive relay-server restarts • Safe compaction: fixed unsafe map mutation during expiration/compaction that could cause data loss or undefined behavior • Full bidirectional CLI: new commands: pastelocal relay send, inbox, fetch, and status, plus pastelocal-remote --relay --peer --send • Auto-sync: when watch.enabled + https://t.co/lFc5oep12e_upload are on, meaningful clipboard changes are automatically pushed to paired devices • Improved DX: doctor checks are now only shown when relay is enabled, and the TUI shows basic relay status • Multiple review cycles: went through full implement → review → fix → re-review (effort 4), plus a final targeted regression fix round Try it out and please send any and all feedback!

This is crazy. My MacBook Pro is closed at home and shouldn’t be on, but I just connected to Codex and am able to code and access my laptop from my phone. Maybe it’s because I enabled computer use? Whatever it is thank you @sama I love it. https://t.co/5ERJ4WIrue

I've finished a video of just over 4 minutes, entirely created using Grok Imagine's Agent mode (see image). It's a much smoother way to make videos because you can rely on Grok's understanding. The Image reference is now available and it's very, very useful. All the dialogues were done with Grok, no need for sound dubbing and that's a huge time saver and adds credibility to the video. The realism is really well done. I post it in one hour.

Just started testing the @grok Build beta. First feel: UX is nice, still some rough edges, but model speed is genuinely cool. If task quality on hard stuff matches opus 4.7 (or even slightly below) at this speed, it's a game-changer. Good chance they steamroll the competition. https://t.co/jpKmvnfiv3

Elon Musk on “Suicidal Empathy”: Gad Saad has a good way to describe it “Suicidal empathy” - where you prey upon people’s empathy so much that it becomes suicidal to your own country and culture You feel sorry for the criminals, but show zero empathy for the victims This is why someone arrested 47 times for violent offenses gets released and then murders somebody We've become so blindly compassionate toward predators that "we're actually allowing our women to get r@ped and our children to get killed" We should have empathy. But that empathy must go to the victims - not the criminals This suicidal empathy has gone too far

How do you know your document parser is ready for production? 🤔 Existing benchmarks miss what AI agents actually need. That's the gap ParseBench, the first doc OCR benchmark for AI agents, fills. We'll unveil all the magic behind it in a live webinar👇 https://t.co/qPIslwCimz