Your curated collection of saved posts and media
EfficientRollout A system-aware self-speculative decoding framework for RL rollouts from FuriosaAI & UC Berkeley that induces a quantized self-drafter to cut rollout latency by up to 19.6% and end-to-end training time by 12.7% without sacrificing model quality. https://t.co/GasdO9Jfz5
@SomeUKTeslaGuy @TopherScottJ @comma_ai And I have a lowered car with a race suspension. Looking at the scenery and talking with a friend is far more fun. https://t.co/USfB3SAQb9
Congrats to Ferrari for winning Formula 1 today. It says it will always force you to manually drive. Which on a curvy road with no traffic maybe that is your preference. This road above Silicon Valley. Where having the robot drive is sometimes the better choice. Went and to
We are really coming together here... Using @wandb I've achieved a much higher Training accuracy and less loss and lower latency this run. Geeked. @solana ai models and datasets/training/fine tuning have arrived. Imagine if @metaplex would just answer their dms? https://t.co/lz1FiF2YpH
@thehill https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@bioshok3 https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@Coinvo https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@HarryRajTyagi https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@Hamzaonchain https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@VoidToVector https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@DragnixMod https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@Mr_Salio https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@heyshrutimishra https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@cnet_japan https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@thenextweb https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@askalphaxiv https://t.co/wVuMh1DuCv
π¨ Anthropic Mythos National Security Crisis: Advanced AI is like a powerful industrial drill. The problem is how easily bad actors can get access unnoticed. A hacker used Claude for months to siphon 150 GB of private data from multiple Mexican government agencies. Exposed: ht
@giffboake @DavidSHolz Well you can always get an MRI scan: https://t.co/tYIpqtAvW0
π’ We're partnering withΒ @modal to offer a new development and exhibition opportunity for artists with sustained engagements in artificial intelligence and the arts. This global open call seeks proposals for creative projects that demonstrate the intentional use of AI to further artistic expression. Generative media, enabled by models trained at network-scale and proliferated by agentic systems, has opened an expansive toolkit for artists. As aesthetic practices evolve, how can artists help us interpret the new realities unlocked through scaling computation? Against a prevailing narrative where inference equates to automation, homogenization, and slop, we think thereβs another story. One where artists harness the inherent creativity of inference to expand perspectives, create novel forms, and suggest alternative ways of sensing the world. Selected artists will receive: π $2,000 honorarium π Up to $5,000 in materials and production stipends π Up to $20,000 value in Modal credits to use for custom model training and cloud GPU compute. πππππππ‘π: πππΉπ ππ± Full details at https://t.co/xlHP1JXu1S
Quick UX tip: Crossing out completed todo items makes them harder to read Checkmarks + dimming are usually enough https://t.co/x8ngxZFgDn
Codex continues to blow my mind everyday! Had to make an ad hoc Codex for OSS grant today, and all I did was send codex an app shot and "can you take care of this?" Codex went and looked through my slack DMs/ chats and chronicle logs to find an API @jxnlco's codex made couple days back It then proceeded to verify the details and make sure our trust and safety filters were happy with it and then waited for me to send the final confirmation and send it! App shots is all you need! LFG!!
Alleged 'ringleader' behind drone attack plot for White House UFC event ID'd as illegal immigrant https://t.co/WFBqkI2Gzy https://t.co/W2Mw0awN1W
"If civilization is not strong, if civilization doesn't grow, then nothing else matters. Profits don't matter if civilization collapses. We cannot exist absent civilization. You have to be pro civilization, because you cannot exist without civilization." β Elon Musk https://t.co/eYMM2qPqtd
NEWS: Flow Cayman partners with Starlink to keep networks online during hurricanes. βIf a hurricane cuts off one of our towers from the core fibre-optic network, then the tower will switch to the Starlink backup and our users will still be connected via satelliteβ β’ Starlink will provide satellite backup for Flowβs mobile network β’ Designed to keep customers connected during hurricanes and major outages β’ Network can switch to Starlink if traditional infrastructure is disrupted β’ Rollout expected before the peak of hurricane season Another telecom operator is integrating Starlink into critical communications infrastructure.
Introducing Airwallex AgentOS. A toolkit for AI agents to work with Airwallex directly from @claudeai, @cursor_ai, or other terminal and chat-based agent environments. https://t.co/Qx9RkyjQd0
Codex and Claude just became a social hangout. Messages, see which friend is online, chat rooms, leaderboards good oldβ ICQ / IRC chats vibes , inside Codex and Claude join -> shellbook . co https://t.co/QZVt8iGzzC
@CaptainHaHaa Yeah! And you know I'll post the scans on X. :-) My MRIs are already up: https://t.co/tYIpqtAvW0
Grok TTS is already sounding insanely human In Vapiβs blind voting Humaneness Index, Grok TTS ranked as the top AI voice model in the chart with a humaneness score of 96.....just 4 points below the real human benchmark β’ Top AI voice model shown β’ 96/100 humaneness score β’ Only 4 points behind the human benchmark What makes this even more impressive is that Grok TTS is combining natural-sounding speech with low latency and aggressive pricing The gap between AI-generated speech and real human voices is disappearing faster than most people realize Grok is starting to speak like a real person
anyone giving me a hard time here might rewind eg to last fall and this kind of thing: https://t.co/ZfyqnGncI9
Accenture ($ACN) last fall: Look at our massive $3 billion bet on AI which will transform our business! Accenture today: quarterly revenue is down slightly Reality beats magical thinking, yet again. https://t.co/ZfyqnGncI9
Tell me you work for an AI lab without telling me: https://t.co/cmvee1GV5j
@Polymarket @bert_designs https://t.co/d9ArOL64ZU
Tell me you work for an AI lab without telling me: https://t.co/cmvee1GV5j
Just added this blog as a project page to https://t.co/6yQIMR6Ltn! Hope that more people can learn about state-of-the-art full-duplex voice models this way :) https://t.co/IcZwVCjIp2
Moshi is one of the best open source full-duplex voice models out there. The architecture is dense, so we spent a few days studying it and wrote up what we learned, with diagrams to make it click faster. Let us know if it was helpful π€ https://t.co/ollrKCnj7f
Just added this blog as a project page to https://t.co/6yQIMR6Ltn! Hope that more people can learn about state-of-the-art full-duplex voice models this way :) https://t.co/IcZwVCjIp2
