The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…

codex writes the most digusting code idk who's responsible for pre-training over there but you gotta flip the script

We heard you wanted to use Codex rate limit resets on your own time. Starting today, we’re rolling out the ability to save rate limit resets to use later. We’re starting Go, Plus, Pro, and Business users with one free reset:

Anthropic's Opus 4.7 and 4.8 models are experiencing degraded performance, which is causing a higher rate of failures for users selecting these models in Notion AI. To mitigate impact, all Anthropic models have been disabled in the model picker and requests have been rerouted to alternative providers. Most users should now be able to continue using Notion AI with minimal disruption, though Anthropic-specific features remain unavailable. Please refer to notion-status.com/ for the details.

You just sold your company for $ 100M. What’s the FIRST thing you buy?

Hot take: If AI tools had no token limits, Claude would win back a lot of developers overnight.

My personal research question for today: Should we optimize the model for a harness or should the harness be optimized for the model?

Ideogram released Ideogram 4.0, a new open-weight image model focused on high-end design, typography, layout control, and photorealism. The model is now live across Ideogram plans and the API, with weights available for download so users can fine-tune it on their own data and run it on their own hardware.

Un tipo de Reddit construyó accidentalmente el sitio web más útil de internet. Se llama FMHY (Free Media Heck Yeah). Google llegó a eliminarlo de los resultados de búsqueda tras recibir reclamaciones por derechos de autor. Aun así, sigue online, y se ha convertido en uno de los directorios más completos de recursos gratuitos de internet. Aquí te explico cómo funciona 👇

Today we’re releasing two new models for Japanese: LFM2.5-Audio-1.5B-JP (audio) and LFM2.5-1.2B-JP-202606 (text). 🧵

Today, we’re excited to introduce Miso One, the most emotive voice model in the world. Miso One is an 8-billion-parameter text-to-speech model for highly expressive speech generation. It emotes like a human and responds faster than a human, with just 110 milliseconds of latency. We’ve open-sourced the model weights, with API access coming soon. Hear how Miso One sounds in the thread below.

Introducing Monako Glass 👓 The world's first wearable Linux computer in glasses form. Run Claude Code, Codex, and any coding agent — anywhere.

I built a Notion Worker Custom Agent that takes whatever we post on @NotionHQ, localizes it, and schedules it to 3 regional X accounts (coming soon). One English post → ES, DE, PT versions. All drafted, reviewed, and scheduled in Sprout for the next morning. It’s a 5-step loop: 1. Router agent creates language pages the moment an HQ post goes live. 2. Translator agent drafts each version using translation glossary. 3. Human translator rewrites for tone cultural nuance. 4. Worker schedules everything in Sprout. 5. QA agent compares Original → Draft → Final and improves translation glossary. The system gets smarter with each run. Now, one local translator per market is enough to run product updates and educational content to 3 regions at ~ the same quality as HQ. Next: hiring SMM contractors in each region so the content feels local... same product heartbeat, but with local events, culture, and community.

which is the most powerful model for coding right now?

Thousands of families are already using Ollie to manage their households. We’re backed by Khosla Ventures and AI2 Incubator.

Ollie’s most powerful skill: memory. Unlike your husband who might forget your daughter’s birthday, Ollie won’t. It: > makes sure you never miss an appointment > triages important updates across your email and WhatsApp > plans meals and groceries based on your family’s preferences This kills the mental load of parenting.