> Seeds won't require a stargateI really wonder how far DSec + DSpark + everything else have come in the last 10 weeksAlexander Doria: Virtualized environments…
vague sense of gnostic dread and an urge to get blackout drunk after reading locallama
> ByteDance: installingcacklingbut yeah! We're still living in the Hopper era. + TPUs/Trainium. "Blackwell class models" like Mythos or GPT-5.x have virtually nothing to do with…
> serve at DeepSeek-API-killing prices. That loop is why xAI runs Grok on SGLang and third parties> beat DeepSeek's own API by 5x on cost.what is…
Join us for a fireside chat on where AI research and infrastructure are headed, led by @tri_dao. Hosted by Together AI, @nvidia and Lyra Labs at…
RT サメQCUhttps://x.com/voooooogel/status/2074217167600734647"unfortunately it's still not perfect, the main limitation imo being the reliance on the vocabulary. "personally im kind of tired of the categorical over tokens…
RT Carter Rabasahttp://x.com/i/article/2074146239483195392
Very important whether these assumptions hold.will depue: A Stargate for DataLabs are on a trajectory towards >$100B/year of data spend by 2030. As we begin the…
Fascinating toolon-the-fly model surgeryThat's the kind of tool that I imagine underpins Anthropic's science of pretraining which elevates them from the industry-standard alchemy to serious physics…
Finally a new bench where GLM-5.2 crashes to mid-tier modelsArtificial Analysis: Announcing AutomationBench-AA, our independent leaderboard for Zapier’s AutomationBench, testing whether AI agents can automate real…
Interesting stuff. And the visualization at the end is worth trying: https://www.neuronpedia.org/qwen3.6-27b/jlensAnthropic: New Anthropic research: A global workspace in language models.Of everything happening in your brain…
RT Gaurav PatilRe @huggingface has one of the craziest upload and download speed I have ever experienced on free tier, and the amount of storage they…
Convincing.Anthropic: For most things, Claude actually doesn’t need its J-space. If we delete the J-space, Claude still speaks fluently, recalls facts, and classifies text—but becomes bad…
A traffic light on a quiet corner, keeping time for a neighborhood that never stops moving. Through every season, it stays. TOMO, by Maximilian Kempe. Made…
Fascinating specimen that spells out what is often quietly assumedDr. Bunsen Honeydew: @Bluebearmonkey In wartime, yes. In peacetime, the opposite is better because you can always…
RT Philip TurnerThe profit motive that won the last venture capital round is falling apart.Since this is taking longer than expected, I'm returning to STM development,…
> resumed after 318h58m0s idle (provider cache expired) — elided 1 stale tool results to cheapen the cold restartI like this new feature of Reasonixcache-centric agents
RT Hannes RudolphOpenClaw FREE MODEL alert!!Hy3 just released from @TencentHunyuan and you can use it in @openclaw now FREE through @OpenRouter for 2 weeks!Run this to…
Really looking forward to asking @drfeifei some questions in person. Only a couple spots left!Michael Grinich: Few people have shaped the trajectory of AI more than…
Taiwan is Japan-maxxing…Nikkei Asia: Why are wealthy Taiwanese pouring money into Japanese real estate just as mainland Chinese buyers are pulling back?https://s.nikkei.com/3SL46Af
RT OpenClaw🦞OpenClaw landed on @huggingface local apps 🦞🤝🤗1. Pick any GGUF/MLX model on hf2. Copy the openclaw onboard setup3. Volla you've got a tool-calling agent running…
Re We also partnered with Neuronpedia to create an interactive demo of our methods on open-weights models.Try it here: http://neuronpedia.org/jlens
RT Paul GrahamIf you want to start a startup, don't learn "entrepreneurship." Learn how to build things. The hard part of startups is not "entrepreneurship" but…
great companyWill 🏴: Can Mullvad get any better?