for those keeping track at home it was 34 days between signing this deal and launching Mythos-class model GA to the world. https://
x.com/leerob/status/
2052059466821198061?s=20
… building on @nvidia stack means you can just do things™.
TECHNOLOGY
-
34 days from signing deal to Mythos-class model GA launch
By
–
-
Anthropic’s new Fable 5 safeguards quietly limit effectiveness
By
–
Anthropic’s new Fable 5 safeguards are fascinating. When the model is used for frontier LLM development, it apparently does not simply refuse or warn the user. Instead, it quietly limits its own effectiveness through techniques like prompt modification, steering vectors, and
-
Apple’s Core AI runs models entirely on-device
By
–
Apple finally did it. Its new framework, Core AI, runs models entirely on Apple silicon, so inference happens on the user's device with zero server calls and zero token bills. That means Qwen, Mistral, and SAM3 running natively across iPhone, iPad, Mac, and Vision Pro. It's a
-
SpaceX launches AI1: AI computing and solar energy in space
By
–
SpaceX has just launched AI1, a satellite designed to relocate AI computing to space and capture energy where it is nearly infinite: the Sun. We are witnessing, before our eyes, the transition from a Type 0 civilization to a Type 1 civilization on the
-
OSCAR: 2-bit KV cache for LLMs without accuracy loss
By
–
Can LLMs run on ultra-low-bit memory without tanking accuracy? Researchers from Together AI, University of Sydney, and UIUC present OSCAR — a method that uses offline, attention-aware covariance analysis to design fixed rotations and clipping thresholds for 2-bit KV cache
-
Google Unveils Gemini 3.5 Live Translate Model
By
–
Google has released a new Gemini 3.5 Live Translate model that supports low-latency translation across 70+ languages. The model is now available in Preview on AI Studio and APIs, and Google Meet will soon integrate this model for live translation as well.
-
Cost frontiers: more expensive and superior model, Fable surpasses Opus 4.8
By
–
The cost frontiers, as expected, show an upward and rightward shift, meaning we have a more expensive and notably superior model (the low of Fable remains well above the x-high of Opus 4.8)
-
SambaNova shows disaggregated inference with up to 2x speed at Computex
By
–
Same prompt. Same model. Two stacks.
— SambaNova (@SambaNovaAI) 9 juin 2026
At #Computex, we demonstrated disaggregated inference live: GPUs handling prefill, SambaNova RDUs handling decode, and CPUs orchestrating agent execution.
The result? Up to 2X the speed of B200-only configurations 🦾 pic.twitter.com/YYP8o6WYrKSame prompt. Same model. Two stacks. At #Computex, we demonstrated disaggregated inference live: GPUs handling prefill, SambaNova RDUs handling decode, and CPUs orchestrating agent execution. The result? Up to 2X the speed of B200-only configurations
-
Creating mobile app, promo video and pitch deck with Replit
By
–
I built a mobile app, promo video, and pitch deck for my travel app at the same time using Replit's parallel agents 👇 pic.twitter.com/alTA1mPULy
— Replit ⠕ (@Replit) 9 juin 2026I created a mobile app, a promotional video and a pitch deck for my travel app at the same time using Replit's parallel agents
-
Skeptic wrong about AI data centers in space due to heat
By
–
Remember that guy who was on here with a video that AI data centers in space would never in a million years happen because of the heat? It's incredible how certain people who are flat out wrong can be just because it's trendy now to be against any sort of technological progress