Qwen 3.5: small models with impressive performancetwitter.com6 points·by moondistance·قبل 4 أشهر·0 comments
Language Models Are Injective and Hence Invertiblearxiv.org1 points·by moondistance·قبل 9 أشهر·2 comments
DeepSeek-R1 at 3,872 tokens / second on a single Nvidia HGX H200blogs.nvidia.com13 points·by moondistance·السنة الماضية·1 comments
ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheapertwitter.com4 points·by moondistance·السنة الماضية·0 comments
US-China Commission top recommendation: Manhattan project for race to AGI [pdf]uscc.gov4 points·by moondistance·قبل سنتين·0 comments
AI scans RNA 'dark matter' and uncovers 70k new virusesnature.com1 points·by moondistance·قبل سنتين·0 comments
moondistance·قبل 11 شهرًا·discussHow does this compare with Exowatt? Assume fresnel lenses are more efficient.
moondistance·السنة الماضية·discussChris Crawford is also famous for the Dragon Speech :) https://youtu.be/CBrj4S24074?si=Ph12RpW8BKsh8-qS
moondistance·السنة الماضية·discussYes, big time, but there continues to be lots of progress.Most importantly, models are maturing, and this means less custom optimization is required.
moondistance·السنة الماضية·discuss(this is also why AMD popped 10% at open yesterday - this is a new development and talks from their 2025 "Advancing AI" event were published late last week + over the weekend)
moondistance·السنة الماضية·discussYes, for many applications.Meta, OpenAI, Crusoe, and xAI recently announced large purchases of MI300 chips for inference.MI400, which will be available next year, also looks to be at least on par with Nvidia's roadmap.
moondistance·السنة الماضية·discussWhere do you find/buy land? How do you vet purchases? Can you point to a few websites, etc.? Thanks!
moondistance·قبل سنتين·discussHappy New Year!I know so few of you but feel like you’re my best friends. Love to all.
moondistance·قبل سنتين·discussMicrosoft is negotiating equity in OpenAI as part of the switch to a for-profit. Non-zero chance this is a negotiation flex.