HackerTrans
TopNewTrendsCommentsPastAskShowJobs

moondistance

no profile record

Submissions

VibeThinker-3B achieves 80.2 on LCBv6

twitter.com
8 points·by moondistance·28 days ago·3 comments

Qwen 3.5: small models with impressive performance

twitter.com
6 points·by moondistance·4 months ago·0 comments

Language Models Are Injective and Hence Invertible

arxiv.org
1 points·by moondistance·9 months ago·2 comments

Autonomous Trash Can

twitter.com
20 points·by moondistance·12 months ago·1 comments

Grok 4 Benchmarks

twitter.com
3 points·by moondistance·last year·0 comments

Jensen Huang – Nvidia GTC 2025 Keynote

nvidia.com
75 points·by moondistance·last year·74 comments

Exaone Deep 32B – beats DeepSeek 671B

lgresearch.ai
3 points·by moondistance·last year·1 comments

DeepSeek-R1 at 3,872 tokens / second on a single Nvidia HGX H200

blogs.nvidia.com
13 points·by moondistance·last year·1 comments

ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheaper

twitter.com
4 points·by moondistance·last year·0 comments

US-China Commission top recommendation: Manhattan project for race to AGI [pdf]

uscc.gov
4 points·by moondistance·2 years ago·0 comments

AI scans RNA 'dark matter' and uncovers 70k new viruses

nature.com
1 points·by moondistance·2 years ago·0 comments

Llama 405B 506 tokens/second on an H200

developer.nvidia.com
21 points·by moondistance·2 years ago·5 comments

comments

moondistance
·28 days ago·discuss
Paper: https://huggingface.co/papers/2606.16140
moondistance
·11 months ago·discuss
How does this compare with Exowatt? Assume fresnel lenses are more efficient.
moondistance
·last year·discuss
Chris Crawford is also famous for the Dragon Speech :) https://youtu.be/CBrj4S24074?si=Ph12RpW8BKsh8-qS
moondistance
·last year·discuss
Yes, big time, but there continues to be lots of progress.

Most importantly, models are maturing, and this means less custom optimization is required.
moondistance
·last year·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
·last year·discuss
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.
moondistance
·last year·discuss
This is from 2020.
moondistance
·last year·discuss
Thanks :)
moondistance
·last year·discuss
Where do you find/buy land? How do you vet purchases? Can you point to a few websites, etc.? Thanks!
moondistance
·2 years ago·discuss
Happy New Year!

I know so few of you but feel like you’re my best friends. Love to all.
moondistance
·2 years ago·discuss
Microsoft is negotiating equity in OpenAI as part of the switch to a for-profit. Non-zero chance this is a negotiation flex.
moondistance
·2 years ago·discuss
Significant further optimizations. FP8!
moondistance
·2 years ago·discuss
VP Research Barret Zoph and Chief Research Officer Bob McGrew also announced their departures this evening.