moondistance
Re:Good people refuse to do bad things
on Monday September 21, 2026 @11:58AM (#49788956)
The root issue is a lack of honour in the world today.
Re:VibeThinker-3B achieves 80.2 on LCBv6
on Tuesday June 16, 2026 @05:09PM (#48562121)
Qwen 3.5: small models with impressive performance
Language Models Are Injective and Hence Invertible
Re:Standard Thermal: Energy Storage 500x Cheaper Than Batteries
on Monday August 25, 2025 @09:55AM (#45013891)
How does this compare with Exowatt? Assume fresnel lenses are more efficient.
Autonomous Trash Can
Grok 4 Benchmarks
Re:End of an Era
on Monday June 30, 2025 @06:07PM (#44428430)
Chris Crawford is also famous for the Dragon Speech :) https://youtu.be/CBrj4S24074?si=Ph12RpW8BKsh8-qS
Re:AMD's CDNA 4 Architecture Announcement
on Tuesday June 17, 2025 @03:48PM (#44303148)
Yes, big time, but there continues to be lots of progress.
Most importantly, models are maturing, and this means less custom optimization is required.
Re:AMD's CDNA 4 Architecture Announcement
on Tuesday June 17, 2025 @03:38PM (#44303016)
(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)
Re:AMD's CDNA 4 Architecture Announcement
on Tuesday June 17, 2025 @03:29PM (#44302933)
Yes, 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.
Re:Graphcore unveils GC200 and M2000 IPU Machine–1 petaFLOP "pizza box...
on Friday May 09, 2025 @05:19PM (#43940973)
This is from 2020.
Re:The Gang Has a Mid-Life Crisis
on Thursday May 01, 2025 @04:12PM (#43862818)
Thanks :)
Re:The Gang Has a Mid-Life Crisis
on Thursday May 01, 2025 @03:45PM (#43862473)
Where do you find/buy land? How do you vet purchases? Can you point to a few websites, etc.? Thanks!
Jensen Huang – Nvidia GTC 2025 Keynote
Exaone Deep 32B – beats DeepSeek 671B
DeepSeek-R1 at 3,872 tokens / second on a single Nvidia HGX H200
ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheaper
Re:Happy New Year 2025 HN
on Tuesday December 31, 2024 @10:14PM (#42563674)
Happy New Year!
I know so few of you but feel like you’re my best friends. Love to all.
US-China Commission top recommendation: Manhattan project for race to AGI [pdf]
Re:GitHub cuts AI deals with Google, Anthropic
on Tuesday October 29, 2024 @04:59PM (#41989347)
Microsoft is negotiating equity in OpenAI as part of the switch to a for-profit. Non-zero chance this is a negotiation flex.
AI scans RNA 'dark matter' and uncovers 70k new viruses
Llama 405B 506 tokens/second on an H200
Re:Llama 405B 506 tokens/second on an H200
on Sunday October 13, 2024 @09:21PM (#41833288)
Significant further optimizations. FP8!
Re:Mira Exits OpenAI
on Wednesday September 25, 2024 @10:06PM (#41653808)
VP Research Barret Zoph and Chief Research Officer Bob McGrew also announced their departures this evening.