Hacker News: spidersouris

New comment by spidersouris in "Speculative Speculative Decoding (SSD)"

spidersouris — Wed, 04 Mar 2026 10:54:05 +0000

Note that a similar idea had already been suggested by Shen et al. (2025) in Speculative Decoding via Hybrid Drafting and Rollback-Aware Branch Parallelism (https://arxiv.org/abs/2506.01979), but with lower performance.

New comment by spidersouris in "This time is different"

spidersouris — Thu, 26 Feb 2026 22:08:41 +0000

> Gen AI reached 39% adoption in two years

Source?

New comment by spidersouris in "TimeCapsuleLLM: LLM trained only on data from 1800-1875"

spidersouris — Mon, 12 Jan 2026 20:55:03 +0000

I'm not sure what you'd call a "pioneering scientific advancement", but there is an increasing amount of examples showing that LLMs can be used for research (with agents, particularly). A survey about this was published a few months ago: https://aclanthology.org/2025.emnlp-main.895.pdf

New comment by spidersouris in "Microsoft please get your tab to autocomplete shit together"

spidersouris — Thu, 25 Dec 2025 10:26:25 +0000

If you want to bind Tab to Accept suggestions:

Set-PSReadLineKeyHandler -Chord "Tab" -Function AcceptSuggestion

Understanding Conversational AI: Philosophy, Ethics, and Social Impact of LLMs

spidersouris — Sat, 20 Dec 2025 07:54:40 +0000

Article URL: https://www.ubiquitypress.com/books/m/10.5334/bde

Comments URL: https://news.ycombinator.com/item?id=46334365

Points: 2

# Comments: 0

New comment by spidersouris in "Show HN: Pulse 2.0 – Live co-listening rooms where anyone can be a DJ"

spidersouris — Fri, 28 Nov 2025 22:27:54 +0000

And plug.dj!

New comment by spidersouris in "Why do AI models use so many em-dashes?"

spidersouris — Sun, 02 Nov 2025 09:11:41 +0000

What we also learned after GPT-3.5 is that, to circumvent the need for new training data, we could simply resort to existing LLMs to generate new, synthetic data. I would not be surprised if the em dash is the product of synthetically generated data (perhaps forced to be present in this data) used for the training of newer models.

New comment by spidersouris in "John Carmack on mutable variables"

spidersouris — Sat, 01 Nov 2025 00:32:48 +0000

For type only though.

New comment by spidersouris in "Typst 0.14"

spidersouris — Sat, 25 Oct 2025 09:24:07 +0000

FWIW, there is an open-source collaborative editor for Typst that was posted a couple of weeks ago on HN: https://news.ycombinator.com/item?id=45481791

New comment by spidersouris in "Typst 0.14"

spidersouris — Fri, 24 Oct 2025 18:48:05 +0000

Maybe you can use Touying Exporter: https://github.com/touying-typ/touying-exporter

New comment by spidersouris in "Typst 0.14"

spidersouris — Fri, 24 Oct 2025 18:47:00 +0000

To add on what's been said already on slide decks, another great slide creation package in Typst is touying[1]. I've used it to create my own academic theme[2] for courses or conference presentations.

[1] https://touying-typ.github.io/ [2] https://typst.app/universe/package/touying-unistra-pristine/

New comment by spidersouris in "Electricity can heal wounds three times as fast (2023)"

spidersouris — Thu, 16 Oct 2025 16:06:42 +0000

> The project was recently granted new funding so the research can get to market and benefit patients.

How is it now? Has this been extended to real use outside of research?

New comment by spidersouris in "SynthID – A tool to watermark and identify content generated through AI"

spidersouris — Sat, 30 Aug 2025 14:04:08 +0000

There is a repo: https://github.com/google-deepmind/synthid-text

New comment by spidersouris in "GPT-5: "How many times does the letter b appear in blueberry?""

spidersouris — Sun, 10 Aug 2025 09:03:47 +0000

There is no inherent need for humans to be "trained". Children can solve problems on their own given a comprehensible context (e.g., puzzles). Knowledge does not necessarily come from direct training by other humans, but can also be obtained through contextual cues and general world knowledge.

New comment by spidersouris in "GPT-5: "How many times does the letter b appear in blueberry?""

spidersouris — Sun, 10 Aug 2025 06:49:22 +0000

For the AI to say this or to produce the correct answer would be easily achievable with post-training. That's what was done for the strawberry problem. But it's just telling the model what to reply/what tools to use in that exact situation. There's nothing about "self-awareness".

New comment by spidersouris in "I tried to replace myself with ChatGPT in my English class"

spidersouris — Tue, 05 Aug 2025 09:08:42 +0000

What's the issue with "… disproportionately affects …"? It seems to be a correct English construction (even though the frequency in COCA is relatively low; 72).

Some examples:

> Indeed, the recent cases of hyperinflation in Brazil, Argentina, and Poland illustrate that although hyperinflation is harmful to savers and disproportionately affects the poor (The Independent Review)

> A hearing is set Thursday on the new version of a legislative bill to eliminate scheduled pay increases for state employees that nixes a section that disproportionately affects rural legislative information offices (USA Today)

> Suicide is a key mental health issue which disproportionately affects men. (london.gov.uk)

New comment by spidersouris in "'Gwada negative': French scientists find new blood type in woman"

spidersouris — Sat, 21 Jun 2025 16:19:45 +0000

FYI, the only English article at the time of posting was Entrevue's, which is why it was initially chosen. But indeed, Le Monde's article is much better.

'Gwada negative': French scientists find new blood type in woman

spidersouris — Sat, 21 Jun 2025 07:38:42 +0000

Article URL: https://www.lemonde.fr/en/science/article/2025/06/21/gwada-negative-french-scientists-find-new-blood-type-in-woman_6742577_10.html

Comments URL: https://news.ycombinator.com/item?id=44335517

Points: 184

# Comments: 68

New comment by spidersouris in "Over 16B records leaked in "unimaginable" major data breach"

spidersouris — Thu, 19 Jun 2025 21:32:07 +0000

I feel like this is clickbait. There is no way to check for now.

edit: I mean, there's always the good old https://haveibeenpwned.com/, but there is no guarantee the leaked data is already in there.

New comment by spidersouris in "As a developer, my most important tools are a pen and a notebook"

spidersouris — Wed, 28 May 2025 20:27:26 +0000

> me getting challenged helps as well (I ask for it)

Can you elaborate? Do you have any examples of such interactions with LLMs?