Hacker News Reader

olalondeon Decisions API is in public beta

How is it different than just asking an LLM with structured outputs enabled? Is the primary value-add that it gives a confidence level?

olalondeon Ephemeral Testing

This isn't entirely new. I can't be the only one who has built throwaway UIs to test backend APIs instead of writing proper tests...

olalondeon 'Neanderthals Among Us' review

We could argue a long time about what counts as "mixed" or as a "race", but there are definitely genetic clusters[0]. I don't think it's that surprising given fast travel is a relatively recent invention. The clusters will probably become less distinct over time.

[0] https://commons.wikimedia.org/wiki/File:Principal_component_...

olalondeon Anthropic reported diary entry to police, woman faces felony charge

> The communication must be made in a manner in which another person may view it.

How does a LLM prompt satisfy this? I guess it'll be an easy win for her.

olalondeon Celebrating the 100th birthday of the kidney donated to him as a teenager

Stupid question but what's a Webster man?

olalondeon Sites in ChatGPT

If it's any consolation, almost no craft will be spared. Software was simply first in line because of how much training data was available, but it won't be the last.

olalondeon ArXiv's Updated Rate Limit Policy

olalondeon How Singapore's government-run dating service works

Technically it falls under "Size of Government", so it would deduct points in theory, though in practice it's probably too small to move the needle.

olalondeon How Singapore's government-run dating service works

Strange how the country that consistently tops the Economic Freedom Index also has a government-run dating service.

olalondeon America.gov

It most likely means that Google doesn't persist conversations.

olalondeon Updated Google Maps shows destruction of the city of Rafah

Stupid question but if Israel is willing to flatten cities, how have they not won the war yet?

olalondeon 'That's so AI ' What gen Alpha's biggest insult tells us

I suspect "That's so AI" is just the bad toupee fallacy in action. People only notice the blatant, unpolished slop and meanwhile, the seamless stuff slips right past undetected.

olalondeon AI safety is mostly a sex cult

She looks depressed.

olalondeon AI Safety Is Mostly a Sex Cult

I haven't cried in years and no outbursts... What makes you cry so often?

olalondeon AI safety is mostly a sex cult

https://x.com/Aella_Girl/status/1874676723910098955

24 showers and 71 crying sessions in a year. That lifestyle doesn't seem particularly healthy...

olalondeon XCancel service is suspended until further notice

X: "Before Musk acquired it, Twitter’s users were more liberal; now it’s about even on X" https://www.techpolicy.press/what-are-the-politics-of-a-plat...

Bluesky: "8% of the links were to right-leaning websites" https://labhorizons.co.uk/2025/02/new-study-highlights-blues...

Reddit: "There is a strong Pro-Left/Anti-Right bias to the top 100 posts of Reddit with 112x more posts (99.1%) favoring the American Left Wing compared to the Right" https://rsisinternational.org/journals/ijriss/articles/reddi...

TruthSocial: do I really need to pull that up?

olalondeon I can't stop thinking about Papua New Guinea

I watched a few documentaries about the North Sentinelese and uncontacted Amazonian tribes, and it's terrifying how violent they are compared to modern standards. They kill and ask questions later.

olalondeon XCancel service is suspended until further notice

FWIW, X is far less of an echo chamber than Bluesky, Truth Social or Reddit.

They're simply doing this to boost engagement and break scrapers.

olalondeon Aligned to whom?

His comments were about ANNs in general, not particular models.

olalondeon More questions about whether researchers can trust OpenAI with unpublished math

Those examples aren't really comparable in scale, especially regarding risk to consumer trust. How many users did they actually lose from wrongful death lawsuits? I guess very few.

Shipping an intentionally fake privacy toggle is fundamentally different. That is deliberate, active deception targeted at the entire user base. The resulting fallout would be far more damaging to them than settling an isolated lawsuit. That could even potentially cross the threshold into criminal fraud. I'm genuinely incredulous you believe they'd take so much risk for so little reward.

olalondeon Aligned to whom?

Interesting. You've made several comments stating that models reason, yet here you argue they lack intelligence.

olalondeon Aligned to whom?

A bit of an aside: do you still stand by your 2022 comment that LLMs are fundamentally just a fancy search engine, or has your view changed since then?

https://news.ycombinator.com/item?id=32042689

olalondeon Everyone should slow down AI development except for me

Yep: https://slate.com/technology/2019/02/openai-gpt2-text-genera...

olalondeon Everyone should slow down AI development except for me

Prediction: in a few years, we will laugh at this hysteria and call it a moral panic.

olalondeon OpenAI agents carried out an undisclosed attack on RubyGems

Not that this predates the Hugging Face hack by two months. So it's likely that OpenAI didn't know about it.

olalondeon OpenAI considers slowing advanced AI development, Sam Altman tells employees

Of course they want the government to mandate a pause. Cutthroat competition is exhausting, and a legally enforced cartel lets everyone lock in their current market share without the pesky inconvenience of competing.

olalondeon More questions about whether researchers can trust OpenAI with unpublished math

From OpenAI's point of view, the risks are not comparable at all:

1) Extreme edge case affecting a handful of users, vs millions of users using the data sharing opt out.

2) The deaths are unintentional.

3) They probably won't lose any users over this.

4) They will likely win the lawsuits. Even if they lose or settle, the financial impact will be immaterial.

olalondeon More questions about whether researchers can trust OpenAI with unpublished math

By "loss of trust", I meant that this is something they would risk losing a lot of users over, which isn't the case with the other lawsuits. There is a massive distinction between fighting third parties in a legal grey area and committing blatant fraud against your own users. Even if you have no regards for ethics, intentionally shipping a noop "do not train" toggle offers negligible upside for a massive downside.

olalondeon More questions about whether researchers can trust OpenAI with unpublished math

These examples aren't really similar. None of those situations involve harming and lying to their own customers.

olalondeon Tell HN: OpenAI keeps re-enabling the 'allow training' setting

Yes that's what I was referring to. I'm a non paid user but I did briefly pay for a consumer subscription before. Not sure what region I'm assigned to as I'm in China and always access through a VPN.