Hacker News Reader

uh_uhon Boeing has started working on a 737 MAX replacement

> This is eerily similar to the comment on the thread about LLMs that says software teams have to maintain a "theory" or model about how the software works and when they cant, they can no longer function beyond limited forms of maintenance.

Link: https://news.ycombinator.com/item?id=45424366

uh_uhon AWS CEO says using AI to replace junior staff is 'Dumbest thing I've ever heard'

Which model?

uh_uhon Man develops rare condition after ChatGPT query over stopping eating salt

Do you think Hinton and Ilya haven’t heard these arguments?

uh_uhon Man develops rare condition after ChatGPT query over stopping eating salt

> LLMs don't think. At all.

How can you so confidently proclaim that? Hinton and Ilya Sutskever certainly seem to think that LLMs do think. I'm not saying that you should accept what they say blindly due to their authority in the field, but their opinions should give your confidence some pause at least.

uh_uhon GTP Blind Voting: GPT-5 vs. 4o

Same result.

uh_uhon Persona vectors: Monitoring and controlling character traits in language models

An LLM is an autocorrect in as much as humans are replicators. Something seriously gets lost in this "explanation".

uh_uhon Why you should choose HTMX for your next web-based side project (2024)

It cannot scale because it doesn't have a solution for reusable components. That's why I have abandoned it. Frameworks like React solve this in a much saner way.

uh_uhon OpenAI claims gold-medal performance at IMO 2025

But don't you think this might be a case where there is both self-congragulation and actual progress?

uh_uhon Reflections on OpenAI

What agentic tools have you tried?

uh_uhon Reflections on OpenAI

> I don't actually believe it's true, I think it's pure hype and LLMs won't even approximate AGI.

Not sure how you can say this so confidently. Many would argue they're already pretty close, at least on a short time horizon.

uh_uhon The Rise of Whatever

What other sources of information will people turn to? Kids are growing up asking ChatGPT in school. I just can't see a mass exodus happening.

uh_uhon Better Auth, by a self-taught Ethiopian dev, raises $5M from Peak XV, YC

Does it also embed two-factor authentication, confirmation/reset emails for me? Those are the reasons one might want to go with Firebase.

uh_uhon Claude 4

Pointing out that OP is using the wrong level of abstraction to explain a new phenomenon is not only useful but one of the ways in which science progresses.

uh_uhon Claude 4

I don't have to have a better explanation to smell the hubris in OP's. I claim ignorance while OP speaks with confidence of an invention that took the world by surprise 3 years ago. Do you see the problem in this and the possibility that you might be wrong?

uh_uhon Claude 4

Your explanation is as useful as describing the behaviour of an algorithm by describing what the individual electrons are doing. While technically correct, doesn't provide much insight or predictive power on what will happen.

Just because you can give a reductionist explanation to a phenomenon, it doesn't mean that it's the best explanation.

uh_uhon Ask HN: How do you promote your personal project in limited bugget?

Why did you get banned?

uh_uhon Ask HN: How do you promote your personal project in limited bugget?

Time and time again I'm surprised how many different types of cat toys are being advertised on Instagram/Twitter by cat video meme pages. Must be a pretty big market.

uh_uhon Ask HN: How do you promote your personal project in limited bugget?

How big is your tiktok following?

uh_uhon Gemini 2.5 Pro Preview

Noticed this too. There's something funny about billion dollar models being handicapped by stuck buttons.

uh_uhon Why is AI so popular when nobody wants it?

I'm not claiming that OpenAI will be the winner of the AI race, but _somebody_ will win big time.

Regarding people losing interest: are you willing to bet with me that there will be fewer than, let's say 500 million active users of LLMs in 5 years?

uh_uhon Why is AI so popular when nobody wants it?

Google has built a multibillion dollar business on top of "free" users. ChatGPT has more than 400 million weekly active users and this is obviously going to grow. You are overlooking how easily that "free" demand will be monetized as soon as they slap ads on the interface.

uh_uhon Why is AI so popular when nobody wants it?

I disagree with the characterisation of AI as "another blockchain: a solution in search of a problem". The two industries have opposite problems: crypto people are struggling to create demand, AI people are struggling to keep up with demand.

uh_uhon Why is AI so popular when nobody wants it?

I read more sceptical takes about AI on Hacker News than anywhere else (since I stopped following Gary Marcus, at least). My hunch is that some people here might feel professionally threatened about it so they want to diminish it. This is less of an issue with some of the 'normies' that I know. For them AI is not professionally threatening but use it to translate stuff, ideate about cupcake recipes, use it as a psychologist (please don't shoot the messenger) or help them lesson plan to teach kids.

uh_uhon CT scans could cause 5% of cancers, study finds; experts note uncertainty

The LLM had the argument that OP's comparison was off. Which you would see if you didn't have a stick up your arse about AI.

uh_uhon CT scans could cause 5% of cancers, study finds; experts note uncertainty

Why is an LLM more prone to hallucination than AlphaGo?

> It isn't my purpose try to convince you that your apparent presumption that the output of a human and a machine are somehow equivalent and should be treated equally is wrong.

You should judge arguments by their merits, not by who is saying them.

uh_uhon CT scans could cause 5% of cancers, study finds; experts note uncertainty

It was NOT hallucinating this time. OP was.

What do you mean all LLM output is hallucination? Would you say the same about AlphaGo? That system was also trained to predict human data initially yet it's competent to the point of beating most humans in Go.

uh_uhon CT scans could cause 5% of cancers, study finds; experts note uncertainty

I don't understand how was the machine hallucinating?

uh_uhon AGI Is Still 30 Years Away – Ege Erdil and Tamay Besiroglu

Not only that but LLMs unsurprisingly make similar distributional mistakes as humans do when asked to generate random numbers.

uh_uhon CT scans could cause 5% of cancers, study finds; experts note uncertainty

That would amount to 10 mrem of radiation per year. I don't believe this is a realistic estimate for a CT scan though. From epa.gov [1]:

- Head CT: 2.0 mSv (200 mrem)

- Chest CT: 8.0 mSv (800 mrem)

- Abdomen CT: 10 mSv (1,000 mrem)

- Pelvis CT: 10 mSv (1,000 mrem)

So for a head CT, one would need to spend more than 13 hours per workday in the station. OP was off at least an order of magnitude.

https://www.epa.gov/radiation/frequent-questions-radiation-m...