HackerTrans
トップ新着トレンドコメント過去質問紹介求人

moondistance

no profile record

投稿

VibeThinker-3B achieves 80.2 on LCBv6

twitter.com
8 ポイント·投稿者 moondistance·28 日前·3 コメント

Qwen 3.5: small models with impressive performance

twitter.com
6 ポイント·投稿者 moondistance·4 か月前·0 コメント

Language Models Are Injective and Hence Invertible

arxiv.org
1 ポイント·投稿者 moondistance·9 か月前·2 コメント

Autonomous Trash Can

twitter.com
20 ポイント·投稿者 moondistance·12 か月前·1 コメント

Grok 4 Benchmarks

twitter.com
3 ポイント·投稿者 moondistance·昨年·0 コメント

Jensen Huang – Nvidia GTC 2025 Keynote

nvidia.com
75 ポイント·投稿者 moondistance·昨年·74 コメント

Exaone Deep 32B – beats DeepSeek 671B

lgresearch.ai
3 ポイント·投稿者 moondistance·昨年·1 コメント

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

blogs.nvidia.com
13 ポイント·投稿者 moondistance·昨年·1 コメント

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

twitter.com
4 ポイント·投稿者 moondistance·昨年·0 コメント

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

uscc.gov
4 ポイント·投稿者 moondistance·2 年前·0 コメント

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

nature.com
1 ポイント·投稿者 moondistance·2 年前·0 コメント

Llama 405B 506 tokens/second on an H200

developer.nvidia.com
21 ポイント·投稿者 moondistance·2 年前·5 コメント

SenseNova 5.5 claims SOTA LLM benchmark results

twitter.com
2 ポイント·投稿者 moondistance·2 年前·0 コメント

New AI Training Technique Is Drastically Faster, Says Google

decrypt.co
84 ポイント·投稿者 moondistance·2 年前·38 コメント

Nvidia open source LLM Nemotron 4 340B at top of the charts [pdf]

d1qx31qr3h6wln.cloudfront.net
17 ポイント·投稿者 moondistance·2 年前·1 コメント

Uncensor Any LLM with Abliteration

huggingface.co
4 ポイント·投稿者 moondistance·2 年前·0 コメント

コメント

moondistance
·28 日前·議論
Paper: https://huggingface.co/papers/2606.16140
moondistance
·11 か月前·議論
How does this compare with Exowatt? Assume fresnel lenses are more efficient.
moondistance
·昨年·議論
Chris Crawford is also famous for the Dragon Speech :) https://youtu.be/CBrj4S24074?si=Ph12RpW8BKsh8-qS
moondistance
·昨年·議論
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
·昨年·議論
(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
·昨年·議論
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
·昨年·議論
This is from 2020.
moondistance
·昨年·議論
Thanks :)
moondistance
·昨年·議論
Where do you find/buy land? How do you vet purchases? Can you point to a few websites, etc.? Thanks!
moondistance
·2 年前·議論
Happy New Year!

I know so few of you but feel like you’re my best friends. Love to all.
moondistance
·2 年前·議論
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 年前·議論
Significant further optimizations. FP8!
moondistance
·2 年前·議論
VP Research Barret Zoph and Chief Research Officer Bob McGrew also announced their departures this evening.
moondistance
·2 年前·議論
I'd like to build a chatbot arena with pro-humanity questions.
moondistance
·2 年前·議論
This is the letter: https://righttowarn.ai/