Hacker News: bonplan23

New comment by bonplan23 in "sshole"

bonplan23 — Thu, 05 Nov 2026 15:00:00 +0000

I do actually. My guess is that this was not just one-prompted, but someone took the time to refine it based on their tastes and expectations until it matched something they would actually like.

However, if someone already developed a strong AI-slop-tingle it will still be triggered by some telltale signs and let their subconsciousness know that they can't possible enjoy this.

New comment by bonplan23 in "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"

bonplan23 — Mon, 26 Oct 2026 10:38:08 +0000

I don't want to crash the party, but there is also a non AI-bubble-is-about-to-burst way to read this: If producing something becomes cheaper (e.g. more efficient inference techniques, cheaper compute etc.) and there is some competition, then typically prices go down as well.

New comment by bonplan23 in "Using thousands of AI agents to uncover new leads in biological data"

bonplan23 — Wed, 21 Oct 2026 20:46:55 +0000

After AI more or less “solved” large parts of mathematics overnight and literally broke the internet by uncovering decades-old PHP vulnerabilities, you still think this is a hype that’s cooling off?

To me, this mostly shows how much our baseline for what counts as “noteworthy” has shifted since ChatGPT came out. I’ve seen more disruptive events in the last couple of years than in the 30 before that.

What would it take to get you excited again? AI discovering that aliens exist?

New comment by bonplan23 in "DeepSeek just open-sourced its inference engine"

bonplan23 — Tue, 20 Oct 2026 10:45:30 +0000

Also the concern that open weights models could contain dormant adversarial behavior (i.e. they are sleeper agents) is more popular amongst decision makers than one would think.

In that regard it was a clever marketing move by Anthropic to release their paper "Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training" ....

New comment by bonplan23 in "DeepInfer: High-performance inference framework for DeepSeek models"

bonplan23 — Tue, 20 Oct 2026 10:34:17 +0000

I saw the same kind of comments when the first Deepseek model was initially released. However the relevance that investors assign to something like this does not seem to match the excitement expressed here.

I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.

New comment by bonplan23 in "There's never been a better time to go into math (by: 3blue1brown) [video]"

bonplan23 — Mon, 19 Oct 2026 09:14:26 +0000

>just don't want to do that....

Nobody ever said that life is fair. Whether you still persist and find happiness typically depends on your willingness to adapt and evolve.

New comment by bonplan23 in "A Princeton PhD Applied to 2,300 Jobs. He Still Can't Get One"

bonplan23 — Sun, 18 Oct 2026 09:44:36 +0000

This kind of job is one in a million. In all other jobs you are expected to work in a team and if the interviewer thinks that you're lacking the basic interface for that - i.e. being able to articulate your thoughts in a concise way and not being annoying to work with - they just cannot hire you.

New comment by bonplan23 in "A Princeton PhD applied to 2,300 jobs and still can't get one"

bonplan23 — Sun, 18 Oct 2026 09:28:12 +0000

There are a lot of PhDs in my company and basically none of them was hired because of the specific topic* of research they were pursuing during their thesis. Yet having the PhD was a huge plus during hiring and in most cases even an (informal) requirement.

I think it happens quite rarely that a company is looking for a PhD based on the specifc topic of their thesis. If that ever happens, you can consider yourself lucky though, because they will probably pay you a fortnue if they really think that they couldn't put any other PhDs on that topic.

*that is not to say field of research like "Machine Learning", but rather something like "loss functions for graph neural networks".

New comment by bonplan23 in "Redesigning source control at Anthropic"

bonplan23 — Wed, 14 Oct 2026 12:21:56 +0000

Many of the comments here seem to stem from the implicit assumption that before claude, software products were thoughtfully designed, secure, reliable, and beautifully engineered works of art.

New comment by bonplan23 in "QGIS bans AI assisted contributions"

bonplan23 — Mon, 12 Oct 2026 14:00:22 +0000

Typically something is "objectively" the case if there is indisputable evidence for it (of which you provided none).

New comment by bonplan23 in "LeCun has "zero concerns" about AI wiping out humanity, recent "rogue" incidents"

bonplan23 — Sun, 04 Oct 2026 09:06:32 +0000

... zero concerns about LLMs wiping out humanity. But world models based on JEPA totally will! Please come invest in my company ... :)

New comment by bonplan23 in "Several vulnerabilities have been discovered in the Linux kernel"

bonplan23 — Fri, 02 Oct 2026 12:48:11 +0000

I think we will soon reach a point where code won't be considered secure unless AI has written or at least verified it. Company policies may then even consider that bad practice and/or forbid it.

New comment by bonplan23 in "The Navier–Stokes Millennium Prize Problem"

bonplan23 — Wed, 09 Sep 2026 08:27:25 +0000

If an entirely new problem, within a new field of mathematics were to appear tomorrow, I highly doubt a human mathematician would be useful at all on their own.

New comment by bonplan23 in "I trained a small transformer in 1.5hrs and it beats many LLMs"

bonplan23 — Tue, 01 Sep 2026 15:11:08 +0000

Just to be clear: It was well known that you can reach such scores with small models and without an LLM if you train on the task. The author highlights those models himself - e.g. HRM/TRM.

The novelty is more that it works with such a plain transformer and low compute price.

Zeiss, the company behind ASML optics, is also doing wildlife monitoring with AI [video]

bonplan23 — Tue, 20 Jan 2026 08:21:24 +0000

Article URL: https://www.youtube.com/watch?v=7kKJOphMxUw

Comments URL: https://news.ycombinator.com/item?id=46689233

Points: 3

# Comments: 0

New comment by bonplan23 in "Show HN: Use Claude Code to Query 600 GB Indexes over Hacker News, ArXiv, etc."

bonplan23 — Wed, 31 Dec 2025 12:45:06 +0000

"Someone" literally did that (+/- 2 years): https://link.springer.com/book/10.1007/978-3-540-68677-4

I think it was supposed to be a more useful term than the earlier and more common "Strong AI". With regards to strong AI, there was a widely accepted definition - i.e. passing the Turing Test - and we are way past that point already: ( see https://arxiv.org/pdf/2503.23674 )

New comment by bonplan23 in "Do the thinking models think?"

bonplan23 — Mon, 01 Dec 2025 08:46:05 +0000

Nice that LLMs can now argue to defend themselves. This is what gemini 3 thinks about its mistake and I find it perfectly valid:

"If you tell a human a joke they think they know, they often stop listening to the setup and jump to the punchline."

And when we say "stop listing" we don't actually mean that they shut their ears, but that they activate a once established neural shortcut - just as the LLM did.

New comment by bonplan23 in "THW: Germany′s army of volunteers for disaster relief"

bonplan23 — Wed, 21 Jul 2021 10:45:17 +0000

Interestingly the THW also operates the "virtual operations support team" (VOST). This group of volunteers is monitoring social media and detecting fake news with specialized software in order to aid crisis relief efforts. https://www.thw.de/SharedDocs/Meldungen/DE/Einsaetze/nationa...