Black Eye 2.0 is getting close. A huge leap forward for cameras in Unreal: • New Orbit / Freelook gameplay camera • New Black Eye Panel • Camera Manager, priorities + blend matrix • Trigger volumes, gameplay events + animation hooks • Through-the-lens blending • Save in Play Free upgrade for all current 1.x users. Price goes up at release.

Seedance 2.0 Advancing Video Generation for World Complexity paper: https://t.co/v0TZmavCUr https://t.co/pWuFZOiX7g

Had a great time at PyCon & PyData DE. Highly recommend it. Great open-source, community-focused conference with lots of builders in the Python AI, LLM and agent space. Taking a short family break, my first "vacation" in years (hopefully, I won't miss the DeepSeek V4 drop 🤞😆) https://t.co/kbKtRZK9aa

Cool to see that Meta conducted and published a pre-deployment investigation of Muse Spark behaviors like reward hacking, honesty, and evaluation awareness! https://t.co/i1Yy7HsEup

GM!! will be speaking at Stanford at 5pm today!! come say hi!! 🫡🥰❤️🦾 https://t.co/voBqShB1Pu

Ferrari for the mind https://t.co/RfTa1xsQXl

Current frontier models are increasingly saturating common AI benchmarks. Are they still useful? We think benchmarks remain important, but they can both over- and understate AI capabilities. To better survey this space, the field is turning to a new paradigm: open-world evals. https://t.co/LsDg0rWfvp

Come see my track at @AICouncilConf where the creator of this model will explain the history and science of quantization I'm telling you: The inference systems track at this years AI Council will be the highest density of this content ever assembled https://t.co/YzABskvJgi

Image generation is now live in Codex! You can now generate visuals, edit existing images, and create GIFs from a single image directly inside Codex. I spent a lot of time testing different use cases while working on this feature, and it was genuinely impressive to see how creative and useful the outputs could be. Hope you enjoy 🚀

Ternary Bonsai: state-of-the-art intelligence at 1.58 bits. The models are so small they can even run locally in your browser on WebGPU! ⚡️ Here's the 8B version (just ~2GB in size) running at 60 tokens per second on my M4 Max. Try the demo out yourself! 👇

🇺🇸 This absolute monster named Tesfaye Cooper, the same thug who tortured a disabled teenager on Facebook Live, is back on the streets. The twist?? He just got arrested again... okay, who am I kidding? Of course he was. He chased down a cyclist who simply waved at him, beat him, spat on him, screamed “Gangster Disciples,” and stole his bike. Released after only 7 years for the first savage attack. In a saner penal system, this predator would never see daylight again. Disgusting. Lock him up for good. Source: @CollinRugg

MERRIN A Benchmark for Multimodal Evidence Retrieval and Reasoning in Noisy Web Environments paper: https://t.co/UZpJdGxIxY https://t.co/ZmRa2TcuAu

🚨 I am truly honored to have the support of President Donald J. Trump! The stakes this midterm could not be higher. That's why when I started running, I promised I would drop out unless President Trump supported me so that Republicans could avoid wasting resources on a needless primary. Much like President Trump, I didn’t come from the world of politics. I spent my entire career in the business world. I had the honor of working with some incredible people and composing music for some of the greatest video games of all time. I'm not going to DC for fame or money. I'm going because I'm a grandfather who deeply cares about what type of country his grandchildren will grow up in. I appreciate the trust that the President has put in me and look forward to working with him in Congress to continue delivering for working families and keeping our nation safe. Now, it's time to ramp this thing into high gear and RETIRE Susie Lee in November.

81% of Grade 4 children in South Africa (that’s 914,000 kids out of 1.1 million) cannot read for MEANING in ANY of South Africa’s 11 languages! It got WORSE — up from 78% in 2016. South Africa ranked DEAD LAST out of 43 countries, as per PIRLS 2021 numbers. Biggest drop of ANY nation. Average score? Just 288… while the world average is 500. Poor kids in townships and rural areas — are the ones suffering the most. They can’t read, can’t learn, and their futures are being stolen. South Africa’s racist laws like BEE (Black Economic Empowerment) put skin colour first and skills last. Elon Musk is 100% right. These race-based rules decide who gets hired as teachers, principals, and education bosses. Competence doesn’t matter, race quotas do. Result? Broken schools. South Africa deserves better. The kids deserve better. Stop the racism.

@elonmusk South Africa has a lot of race laws to cut https://t.co/hvTyheI8c8

We’ve been thinking a lot about scaling laws, wondering if there is a more effective way to scale FLOPs without increasing parameters. Turns out the answer is YES – by looping blocks of layers during training. We find that predictable scaling laws exist for layer looping, allowing us to use looping to achieve the quality of a Transformer twice the size. Our scaling laws suggest that for a fixed parameter budget, data and looping should be increased in tandem! 🧵👇

read my vogue interview here. thank you john cale, jim jarmusch, stefano gallici, gabbriette, a. g. cook and george for contributing some words https://t.co/HAW3tEY90n https://t.co/HHsHcQ6P8e

We shipped a bunch of new features in Codex today, but I’m especially excited about plugins. So many new plugins just landed that make Codex even more powerful! https://t.co/I4kVFpm3G3

You can now visualize Pi traces that you upload on @huggingface! Let's make sharing agent traces 10x more common to make agent AI more open and collaborative! Also, because it's fun to analyze @badlogicgames's traces 😂😂😂 https://t.co/LLclFIZeWS

I will have the best MiniMax compressions on the market. REAP is out, it's a v0, I started repruning the models around feedback (; I am going to quantise this to 4bit & GGUF https://t.co/7BlITpUkTq https://t.co/vorWncLAAb

Quick career update: joined @HuggingFace as a Research Engineer to make RL go brrrrr 🚀 P.S. I’ve been a huge fan of hf ever since I started working on ML and contributing to open source. I can’t imagine what open-source AI would look like without HF. I’ve looked up to so many people here, and it’s truly a pleasure to now work with them.

Both Starship and the Super Heavy Booster have successfully completed the static fire tests and are ready to take to the skies Every test brings us one step closer to making humanity multi-planetary "Engineering is the closest thing to magic that exists in the real world" — Elon Musk

📢📢A double launch today! We’re releasing a paper analyzing the rapidly growing trend of “open-world evaluations” for measuring frontier AI capabilities. We’re also launching a new project, CRUX (Collaborative Research for Updating AI eXpectations), an effort to regularly conduct such evaluations ourselves. I think open-world evals are the most important development in AI evaluation over the past year. Our paper explains why we need them, what they can and can’t tell us, and how to do them well. In CRUX #1, we tasked an agent with building and publishing a simple iOS app to the Apple App store. The paper has many “lessons from the trenches” from running this experiment. We hope you find it interesting! CRUX #2 will be about AI R&D automation. The core team is @sayashk, @PKirgis, @steverab, Andrew Schwartz, and me. We’re delighted to have assembled an amazing group of collaborators, many of whom have conducted important open-world evaluations: @fly_upside_down, @RishiBommasani, @DubMagda, @ghadfield, @ahall_research, @sarahookr, @sethlazar, @snewmanpv, @DimitrisPapail, @shostekofsky, @hlntnr, and @CUdudec. Paper: https://t.co/M15jgh4PCP HTML version: https://t.co/iuVW7RAlr5 CRUX website: https://t.co/g937gpS65j

Benchmarks are saturated more quickly than ever. How should frontier AI evaluations evolve? In a new paper, we argue that the AI community is already converging on an answer: Open-world evaluations. They are long, messy, real-world tasks that would be impractical for benchmarks. https://t.co/CrvbEd9l7f

So excited to share that we're bringing Computer Use to Codex. Computer Use lets Codex see, click, and type into your Mac apps, with its own cursor. It's a magical feeling to have agents using your apps in the background, and still get to use your computer at the same time. https://t.co/wdgxiHAKyX

This tweet was sent by Codex via Computer Use https://t.co/UyxAw6MVj0

This tweet was sent by Codex via Computer Use https://t.co/UyxAw6MVj0

Biggest lesson from OpenClaw is that a good teammate doesn't start from scratch everytime you check in. They remember what was decided, what's still open, and proactively help you. Today we launched heartbeats in Codex: automations that maintain context inside a single thread over time. Instead of each run starting fresh, Codex wakes up in the same conversation, with the history and context it needs already in place. You can also have it schedule its own next steps – just ask Codex. Think about the overhead that quietly accumulates every morning: scanning Slack channels, catching up on email, piecing together what moved overnight. With a heartbeat, you offload that once, and wake up to a brief already waiting in a pinned thread. If you want to try turning Codex into a chief of staff: connect Slack, Gmail, and Notion, and paste the following prompt into codex: Please check @Slack @Gmail @Notion and write me a morning brief every weekday at 9am in this thread. I want you to collapse all the chaos at work into a single note every morning over some coffee ☕️

Shocking result on my pelican benchmark this morning, I got a better pelican from a 21GB local Qwen3.6-35B-A3B running on my laptop than I did from the new Opus 4.7! Qwen on the left, Opus on the right https://t.co/kDlbnJv6YI

Today we’re announcing Ternary Bonsai: Top intelligence at 1.58 bits Using ternary weights {-1, 0, +1}, we built a family of models that are 9x smaller than their 16-bit counterparts while outperforming most models in their respective parameter classes on standard benchmarks. We’re open-sourcing the models under the Apache 2.0 license in three sizes: 8B (1.75 GB), 4B (0.86 GB), and 1.7B (0.37 GB).

one small step towards something super one of the most exciting things I've seen is 1) very good skill triggering 2) computer use works in the background so I can multitask 3) better pdf handling and connectors too! https://t.co/ZeBf7a5kQQ

With computer use on macOS, Codex can now use any app by seeing, clicking, and typing with its own cursor. It runs in the background without taking over your computer, working on tasks like frontend iteration, app testing, or any workflow that doesn't expose an API. https://t.co/iO9iubLZX9