Hacker News: bigdict

New comment by bigdict in "Cannabis usage in older adults linked to larger brain, better cognitive function"

bigdict — Wed, 04 Feb 2026 14:51:10 +0000

Was it a controlled study or just correlation?

New comment by bigdict in "Five Practical Lessons for Serving Models with Triton Inference Server"

bigdict — Sun, 18 Jan 2026 13:47:54 +0000

Huh? Triton inference server and Triton the kernel language are two distinct, very different things… Is this AI-generated?

New comment by bigdict in "LMArena is a cancer on AI"

bigdict — Thu, 08 Jan 2026 01:05:51 +0000

> What actually happens: random Internet users spend two seconds skimming, then click their favorite.

> They're not reading carefully. They're not fact-checking, or even trying.

Uhhh, how was that established?

Artificial General Cleverness

bigdict — Tue, 16 Dec 2025 14:31:48 +0000

Article URL: https://mathstodon.xyz/@tao/115722360006034040

Comments URL: https://news.ycombinator.com/item?id=46288950

Points: 4

# Comments: 0

New comment by bigdict in "Matrix Core Programming on AMD CDNA Architecture"

bigdict — Sun, 30 Nov 2025 05:46:54 +0000

cuz area and power

New comment by bigdict in "What is a manifold?"

bigdict — Tue, 04 Nov 2025 14:48:42 +0000

There's antimony, arsenic, aluminum, selenium…

New comment by bigdict in "Language models pack billions of concepts into 12k dimensions"

bigdict — Mon, 15 Sep 2025 05:21:55 +0000

What's the point of the relu in the loss function? Its inputs are nonnegative anyway.

New comment by bigdict in "Scarcity, Inventory, and Inequity: A Deep Dive into Airline Fare Buckets"

bigdict — Tue, 22 Jul 2025 01:12:27 +0000

Exactly, it's a natural way to write a college essay. I've never not cringed reading an article/blog post that is structured that way, it comes across very contrived. I've also noticed that LLMs tend to prefer it, and humans tend to avoid it in general.

New comment by bigdict in "Scarcity, Inventory, and Inequity: A Deep Dive into Airline Fare Buckets"

bigdict — Mon, 21 Jul 2025 23:35:33 +0000

"delves"

dashes

an explicit "conclusion" section at the end

New comment by bigdict in "Notes on Graham's ANSI Common Lisp (2024)"

bigdict — Sun, 13 Jul 2025 14:54:09 +0000

Lisp In Small Pieces

New comment by bigdict in "Dinesh’s Mid-Summer Death Valley Walk (1998)"

bigdict — Wed, 18 Jun 2025 01:24:39 +0000

Pairs great with the tale of the Death Valley Germans: https://www.otherhand.org/home-page/search-and-rescue/the-hu....

New comment by bigdict in "Emacs dired-mode as a file manager"

bigdict — Fri, 23 May 2025 20:18:02 +0000

Once you realize that elisp is a better shell programming language than bash, it ceases to appear insane.

New comment by bigdict in "Vision Transformers Need Registers"

bigdict — Mon, 28 Apr 2025 17:33:47 +0000

Has this been used widely since?

New comment by bigdict in "Gemma 3 QAT (Quantized Aware Training) 3x less memory"

bigdict — Thu, 03 Apr 2025 17:29:33 +0000

Thank you so much for continuing to support Gemma 3 with these updates.

New comment by bigdict in "Gemma 3 QAT (Quantized Aware Training) 3x less memory"

bigdict — Thu, 03 Apr 2025 17:24:33 +0000

Amazing, I've been wishing for this! Do you have any estimates on how much accuracy is first lost then recovered compared to the original bf16 and the naively quantized models?

New comment by bigdict in "Multi-Token Attention"

bigdict — Wed, 02 Apr 2025 22:56:18 +0000

Sure, you can get better model performance by throwing more compute at the problem in different places. Does is it improve perf on an isoflop basis?

New comment by bigdict in "OpenAI releasing new open model in coming months, seeks community feedback"

bigdict — Mon, 31 Mar 2025 20:36:30 +0000

Gemma 3 is.

New comment by bigdict in "Gemma 3 Technical Report [pdf]"

bigdict — Thu, 13 Mar 2025 02:38:21 +0000

* the code is at https://github.com/google-deepmind/gemma

* you download the weights at https://www.kaggle.com/models/google/gemma-3/

New comment by bigdict in "Open-Sourcing R1 1776"

bigdict — Tue, 18 Feb 2025 21:50:56 +0000

"undo their disgusting propaganda, apply our beautiful correct opinions"

This is so cringe.

New comment by bigdict in "Resist Authoritarianism by Refusing to Obey in Advance (2017)"

bigdict — Wed, 12 Feb 2025 03:18:04 +0000

Why is it timely?