https://t.co/4aVgLYY50L

Trump administration officials tell WIRED that if Anthropic wants to rerelease Fable 5, it will need to ensure the model's guardrails can't be circumvented. Security experts say that can't be done. https://t.co/itfeqzRgSS

Two years ago I became obsessed with making AI videos. No filmmaking experience. Just a genuine love for the craft. My only real goal has been to make the best videos I possibly can. Last week, this spec ad I made for @Redfin was nominated for AI Commercial of the Year at the @generatedawards. The category included work from @Google, @kfc, and @runwayml (congrats @jforozcop for the win) Still not exactly sure what the Generated Awards are, but it feels pretty good. Really gratifying to follow your gut and actually see progress. Onto the next one.

@JayKurtz90 I have been sharing a few thoughts in my timeline about this: https://t.co/lBtzwkkeZb and in some recent articles about how I am handling these challenges. But lots to explore.

HyperFrames is now officially live on @Grok Grok knows everything. HyperFrames shows it Any answer becomes a polished video in minutes Proof: this video was created entirely with it https://t.co/6y6VY8lRqh

The @github Copilot app is now generally available. Hand off from agent powered flows in the app to full coding in @code without losing context or the conversation. 🔗 Learn more: https://t.co/wCY5fQJHt4 https://t.co/fPetm6W6xU

🚨 A Netflix engineer built an open-source proxy that cuts AI token usage by 60-95%. Zero code changes. Benchmarks show ±0.000 accuracy regression. ✨ 29.9k stars on GitHub. It sits between your app and the LLM, so every tool output, code block, and conversation history gets compressed in-flight. 🚫 No summarization, no loss. 😎 Just 60-95% fewer tokens with the same answers. Works with Claude Code, Cursor, Copilot, and any OpenAI-compatible client. One pip install, one env var, done. Netflix uses it internally. Apache 2.0. Built by Tejas Chopra. https://t.co/u1OIlMF5gm

Had my Hermes Agent evaluate headroom. Key conclusions: Most of what it would do for Hermes actually is net greater token cost 🧐🧐 There was one aspect it discovered would save ~60% of tokens used on the search file tool, which we will be integrating! Read the full report: https://t.co/8X7et0ctu2

I have a fun, oddly useful AI benchmark: "build me a procedurally generated 3D simulation showing the evolution of a harbor town from 3000 BC to 3000 AD, it should look beautiful & allow me to have some control over it" Play the gallery of 20 models: https://t.co/zN2uHY1gl8

Unreal Engine 5.8 ships today with experimental MCP server support: Your sources, your pipeline and your workflow—simply configure the MCP plugin and connect to any agent. Get familiar with the MCP server and the PCG Primitive Plugin today and see what teams can build together: https://t.co/cDITLWWv2F

We're hiring a DevX Engineer lead at DeepMind India🥳 This person will shape our regional strategy, work with an ecosystem of millions of developers, and work in the GTM for our models and tools. Please share! https://t.co/gzZcdxx5s3

GPT-5.4 helped drive a medicinal chemistry project from literature review to a validated experimental result. Paired with https://t.co/gcDaph8b2B’s Maria AI and specialized lab, the model proposed an unexpected way to improve a widely used reaction in drug discovery. https://t.co/KmyBlHLX8y

Maria tested the idea across 10,080 reactions, and human chemists later validated representative results by hand. Under the optimized conditions, yields improved for 88% of the boronic acids and 83% of the sulfonamides tested. Human chemists then repeated 14 representative reactions by hand: 11 showed higher yields, including 8 with a more than twofold improvement.

I didn't know what JEPA is, and at this point I am too afraid to ask ...😬 so I made a video tracing the last 30+ years of self-supervised learning, covering ideas from contrastive learning, distillation, masked modeling, JEPA, and world models. https://t.co/4HQp48gqYp https://t.co/wrLWqd6WcF

Nominations are officially open for the 2026 PyTorch Foundation Contributor Awards. These awards celebrate outstanding impact across PyTorch Foundation hosted projects, including PyTorch, vLLM, DeepSpeed, Ray, Helion, and Safetensors. Whether it be through code development, documentation improvements, mentorship or community leadership - we want to recognise the individuals who help move our ecosystem forward. Submit your nomination by July 17. Link in comments

@Ferbin08 @odysseyml You should watch this: https://t.co/FOREXksvXC

Email dashboards had a good run. Two decades. Billions of emails. I built two companies on them. But the dashboard was never the product, it was the bottleneck. We removed it. This is Nitrosend. → https://t.co/AwAv8mP5KM https://t.co/1VbBlt4cJd

You can now train on 350+ RL Environments from OpenReward with TRL with just a few lines of code https://t.co/E3Zy3VTi6x

You can now train on 350+ RL Environments from OpenReward with TRL with just a few lines of code https://t.co/E3Zy3VTi6x

BREAKING: Iraq has approved Starlink’s operating license, clearing the way for its satellite internet service to enter the country. Iraqi Prime Minister Ali al-Zaidi and U.S. Special Presidential Envoy for Iraq Tom Barrack welcomed the approval of Starlink’s operating license. https://t.co/Kuf79uZPta

AI is making marketers lazy. So we made the website do the work instead. Today, we're launching @ployai: the all-in-one marketing platform that turns your website into your hardest working employee. And we're coming out of stealth today with a $27M seed led by @ycombinator and @firstround. I spent 12 years at Webflow as the founding CTO where I built the product, but also started our marketing and sales teams that drove our fastest periods of growth. That experience made one thing obvious: the website is the center of your business. And it's only more important in the age of AI. Foundation model apps can generate assets. Point solutions can optimize pieces of the funnel. But nothing runs the whole growth system: your site, brand, CMS, CRM, campaigns, analytics, SEO, AEO, and customer data all working together. Until now. Teams at @hex_tech and @clay_run, and growth agencies like Tonik and TNT Growth, are already powering sites on Ploy. Hex is generating on-brand ABM pages at scale, Clay is using its data to power a programmatic SEO engine, and TNT Growth is spinning up a landing page for each of their clients’ ads. Wake up every morning with a report from Ploy - with what it did, and what it wants to do next. Approve it. Ship it. Or be lazy and just watch it cook.

Breaking: @YCombinator startup heyClicky website has been reported deceptive by @Apple 😱 https://t.co/rq47vO9dOr

+1 more

@odysseyml So happy to see you all got the gas to complete the dreams your founder laid out here: https://t.co/K4Xlu9ijOt

So blessed to have been the first to really understand @odysseyml's founder's dreams here in this video in a Silicon Valley park: https://t.co/K4Xlu9ijOt

We’re announcing 3 things as part of the Crosby Intelligence launch today: 1/ RedlineBench with @micro1_ai – publishing the first benchmark measuring how frontier models handle multiple steps of a complex, real world contract negotiation*, hosted on @huggingface 2/ The Crosby Intelligence Research Fellowship – funding two fellows pursuing frontier research with support from @OpenAI: $25K + $12.5K Codex credits each 3/ Hosting the most interesting conversations in applied AI at our Soho office, featuring @paraga, @rahulgs, @PeterHndrsn, @NeelGuha, and more *built externally with no client data Read more about all three at https://t.co/WuoSfT4g2E

Today, we're launching Crosby Intelligence to push the frontier of legal AI forward. https://t.co/ia3NGm3390

Today, we're launching Crosby Intelligence to push the frontier of legal AI forward. https://t.co/ia3NGm3390

Our CEO Jerry Liu is joining founders from LangChain, CrewAI, and others at @Databricks #DataAISummit today for a panel on the The Agentic Stack — what the stack looks like, where it's headed, and what happens when agents become the primary consumers of infrastructure, not humans. Panel begins at 11.30 am, see you there. 👋

I made Physical AutoResearch sound simple (conceptually), but it took a village to pull off and lots of design thinking into the robot /loopcraft. The hardest part is everything we need to setup *before* pressing Enter. Here's a behind-the-scene tour: 1. Safety harness Letting 8 robots run unattended overnight means safety has to be more than a hint in the system prompt. ENPIRE hardwires it in 2 layers: (1) hard kinematic limit that trips an immediate task failure and auto-resets as soon as a robot leaves its safety envelope, and (2) a torque-limited compliant gripper so a bad contact or misaligned insertion ends in a safe stall, instead of crushing the robot or the object at hand. We make safety more conservative than usual so humans can sleep tight. In reality, we still need a few human operators to watch over the "robots of loving grace". 2. Definition of /done An agent that can edit its own reward will game it for sure. ENPIRE fixes the goalposts before the fleet can move them. Here's the recipe: Collect a few minutes of success & failure demos -> Ask agent to write code using computer vision tools to classify success and measure against groundtruth -> Agent hill-climbs on classifier until reliably good -> This classifier becomes the real-time reward function that directly computes on sensor streams -> *Freeze* the reward function before AutoResearch. It's sacred, enshrined in a Gym env that no one can touch. 3. System telemetry design Robot-seconds is by far the scarcest resource, followed by GPU-seconds, and finally tokens. We instrument all three and surface them to ENPIRE for live resource awareness rather than letting it hill-climb in a vacuum. We define: - Mean Robot Utilization ("MRU"): the fraction of wall-clock time when the robot is actively executing an experiment. Otherwise the hardware is sitting idle and waiting for the next code commit. - Mean Token Utilization ("MTU"): tokens consumed per minute, our proxy for how hard the agent is actually thinking. A low MTU means the agent is stalled, waiting on a robot rollout to finish instead of doing research. - GPU utilization: fraction of wall-clock time when GPU is active. ... and evaluate on two budget-to-outcome metrics: 1. Tokens-to-Success: token budget the fleet burns to complete /goal. 2. Time-to-Success: wall-clock time to /goal

Introducing Epic’s version control system: Lore! Built from scratch and open-sourced, we made scalability and performance core tenets from day one. We’re releasing Lore as open source today.All source code, along with documentation of data formats and protocols are released under an MIT license and free to use. Help shape it into something even more powerful, together: https://t.co/zplVMjkRof

We're releasing MolmoMotion, a 3D motion forecasting model. Given one or a few video frames, 3D points on an object, & an instruction like "Put the white bowl on the table," MolmoMotion predicts where those points will go over the next few seconds in a shared 3D world frame. 🧵 https://t.co/F9wfwSQHRN