Hacker News: slickytail

New comment by slickytail in "Digital Red Queen: Adversarial Program Evolution in Core War with LLMs"

slickytail — Fri, 09 Jan 2026 00:19:46 +0000

I think it would, for all practical purposes, be impossible to determine an optimal warrior, even at very small core sizes. Not only is the search space huge but the evaluation function can take unbounded time to resolve. We should consider the halting problem embedded inside the optimization target as a clue to the problem's difficulty.

New comment by slickytail in "The terminal of the future"

slickytail — Wed, 12 Nov 2025 07:06:52 +0000

As a non Emacs user, I would be really interested in a full writeup of how this works.

New comment by slickytail in "Do not accept terms and conditions"

slickytail — Tue, 21 Oct 2025 20:17:10 +0000

The site has a little game where you click to dismiss as many cookie-consent popups as possible, in 30 seconds. I suppose if you can't see the cookie consent popups, then at the end of the timer you just have zero points.

New comment by slickytail in "The Unknotting Number Is Not Additive"

slickytail — Thu, 09 Oct 2025 07:53:30 +0000

In the words of Kronecker: "God created the integers, all else is the work of man."

New comment by slickytail in "Is a movie prop the ultimate laptop bag?"

slickytail — Mon, 22 Sep 2025 22:30:08 +0000

Two different groups of people commenting different things.

New comment by slickytail in "Interview with Japanese Demoscener 0b5vr"

slickytail — Fri, 05 Sep 2025 18:16:06 +0000

In what sense is this 64KB? Clearly there's more than 64KB of code in the repo. And since it's typescript it's not like there's a binary that could be 64KB.

New comment by slickytail in "What Is the Fourier Transform?"

slickytail — Fri, 05 Sep 2025 16:37:50 +0000

In a discrete Fourier transform, the number of frequencies you get out is the number of datapoints you have as input. This is because any frequencies higher than that are impossible to know (ie, they are above the sampling frequency).

But in the continuous Fourier transform, the output you get is a continuous function that's defined on the entire real line.

New comment by slickytail in "New research reveals longevity gains slowing, life expectancy of 100 unlikely"

slickytail — Sun, 31 Aug 2025 09:36:22 +0000

Traditional suicide is incredibly stigmatized; ending one's life manually is a huge trauma to place on loved ones. The benefit of MAID is that it's dignified, and won't leave families searching for answers after a death.

New comment by slickytail in "From multi-head to latent attention: The evolution of attention mechanisms"

slickytail — Sat, 30 Aug 2025 10:21:49 +0000

The transformer was a major breakthrough in NLP, and it was clear at the time of publishing that it would have a major impact. But I will add that it is common in the Deep Learning field to give papers catchy titles (see, off the top of my head: all the YOLO papers, ViT, DiT, textual inversion). The transformer paper is one in a long line of seminal papers with funny names.

New comment by slickytail in "The warning signs the AI bubble is about to burst"

slickytail — Fri, 22 Aug 2025 16:52:18 +0000

I think the biggest sign of a potential crash is that Meta has frozen AI hiring. Such a quick reversal from just a few months ago.

New comment by slickytail in "The surprise deprecation of GPT-4o for ChatGPT consumers"

slickytail — Sat, 09 Aug 2025 07:39:39 +0000

I stand corrected! Thanks.

New comment by slickytail in "The surprise deprecation of GPT-4o for ChatGPT consumers"

slickytail — Fri, 08 Aug 2025 23:04:53 +0000

Source for this? My understanding was that this was true for dalle3, but that the autoregressive image generation just takes in the entire chat context — no hidden prompt.

New comment by slickytail in "Ask HN: How can ChatGPT serve 700M users when I can't run one GPT-4 locally?"

slickytail — Fri, 08 Aug 2025 22:43:22 +0000

The memory bandwidth on an H100 is 3TB/s, for reference. This number is the limiting factor in the size of modern LLMs. 100GB/s isn't even in the realm of viability.

New comment by slickytail in "Qwen-Image: Crafting with native text rendering"

slickytail — Mon, 04 Aug 2025 23:12:31 +0000

All of this only really applies to LLMs though. LLMs are memory bound (due to higher param counts, KV caching, and causal attention) whereas diffusion models are compute bound (because of full self attention that can't be cached). So even if the memory bandwidth of an M3 ultra is close to an Nvidia card, the generation will be much faster on a dedicated GPU.

New comment by slickytail in "Gemini Diffusion"

slickytail — Thu, 22 May 2025 09:16:33 +0000

The relative unimportance of the exact SDPA attention in use in modern transformers is already known: https://arxiv.org/abs/2111.11418

The FFN, normalization, and residual connections are absolutely irreplaceable -- but attention can be replaced with almost any other layer that shares information between tokens, such as pooling, convolution, random mixing, etc.

New comment by slickytail in "Everyone knows your location: tracking myself down through in-app ads"

slickytail — Tue, 04 Feb 2025 12:54:30 +0000

The linked wikipedia article below says that they asked you for your email password specifically -- is there any evidence that they would try to use your linkedin password itself?

New comment by slickytail in "Apple's AirPods Pro hearing health features"

slickytail — Tue, 22 Oct 2024 08:44:38 +0000

Generally the problem with this type of argument is that the two sources are not volume-matched. Try out an ABX test here, of lossless vs various lossy codecs: https://abx.digitalfeed.net/