HN user

puttycat

2,321 karma
Posts152
Comments288
View on HN
twitter.com 28d ago

France national supercomputer shuts down GPU/CPU clusters due to heatwave

puttycat
13pts1
grumpy.website 1mo ago

Hiding the sidebar toggle under your logo is NOT a good idea

puttycat
2pts0
www.iranintl.com 1mo ago

Iran appears set to restore internet access after 3-months blackout

puttycat
3pts0
www.haaretz.com 2mo ago

Starlink users can be located and sometimes identified through terminal mapping

puttycat
4pts3
www.nytimes.com 2mo ago

Upcoming El Niño: The World Is About to Get a Preview of Life in 2035

puttycat
10pts0
www.youtube.com 3mo ago

Defrag ASMR

puttycat
2pts0
medium.com 4mo ago

The First Pancake Principle. How to Get Started

puttycat
2pts0
old.reddit.com 5mo ago

Altman on AI energy: it also takes 20 years of eating food to train a human

puttycat
55pts118
www.theguardian.com 5mo ago

Spain approves decree to regularise half a million undocumented migrants

puttycat
3pts0
news.sky.com 6mo ago

Spain's fourth train in five days

puttycat
1pts0
www.theguardian.com 6mo ago

Matthew McConaughey trademarks catchphrase in bid to beat AI fakes

puttycat
8pts1
enigma.inc 6mo ago

Multiverse: The First AI Multiplayer World Model

puttycat
2pts0
www.cnn.com 7mo ago

FBI Foils New Year's Eve Attacks in Southern California

puttycat
12pts4
gptzero.me 7mo ago

Over fifty new hallucinations in ICLR 2026 submissions

puttycat
507pts422
old.reddit.com 8mo ago

WhatsApp starts hiding full URLs in messages

puttycat
2pts0
people.idsia.ch 8mo ago

Who Invented Transformer Neural Networks?

puttycat
2pts0
www.reuters.com 9mo ago

Iran reports bird flu outbreak in poultry

puttycat
4pts0
realitycheckk.com 10mo ago

Reality Check – AI/real photo quiz

puttycat
3pts0
www.axios.com 10mo ago

Anthropic to pay $3k per book in AI copyright settlement

puttycat
18pts0
en.wikipedia.org 10mo ago

Fractal Compression

puttycat
2pts0
www.nytimes.com 11mo ago

The Changing Politics of Masks

puttycat
4pts0
en.wikipedia.org 1y ago

Iron Beam

puttycat
3pts0
news.ycombinator.com 1y ago

Ask HN: Who else is not using AI for coding?

puttycat
13pts7
news.ycombinator.com 1y ago

Ask HN: Does Google train on Google Docs?

puttycat
9pts6
thenewstack.io 1y ago

Python Pandas Ditches NumPy for Speedier PyArrow

puttycat
17pts5
www.thediff.co 1y ago

The Factorio Mindset (2022)

puttycat
1pts0
engineering.berkeley.edu 1y ago

Brain-to-voice neuroprosthesis restores naturalistic speech

puttycat
1pts0
www.theguardian.com 1y ago

First case of bird flu in sheep detected in England

puttycat
4pts0
www.nature.com 1y ago

Causal chambers as a real-world physical testbed for AI methodology

puttycat
1pts0
www.instagram.com 1y ago

Official Sora sample videos on Instagram

puttycat
1pts0

France is indeed ridiculously bad at ventilation (not to mention air conditioning). Restaurants, offices, even gyms - most have bad to non-existing ventilation. Coming from the States this is just insanity.

A periodic reminder that companies are paperclip optimizers that will stop at nothing to protect their profits and existence.

If you are developing anything in AI or related domains that is of immediate value and/or in competition with Anthropic (and the like), DO NOT use a CLI programming agent. Preferrably obfuscate your code and gut it of sensitive IP before showing it to agents. Do not trust the dont-train toggle.

+1 for office work as disease source. Plus taking public transport (mostly underground trains). Today I avoid the subway almost completely, taking my bike almost everywhere, to avoid getting sick during winter. (No kids though)

I have two languages defined in iOS (English and French) and the native swift keyboard recognizes which language I'm swifting in.

Maybe this is supported only for specific language combinations.

My guess is that an AI layer already present _is_ the reason you're seeing these unhinged suggestions. Earlier generation iOS keyboards were much more accurate and reasonable.

Sakana Fugu 1 month ago

Can someone explain this in layman terms? I don't understand any of it

The final weights in themselves don't tell you anything about what went into the training process, e.g censorship.

100%. Companies are paperclip optimizers, with money as the objective. For example, Uber used ride data to circumvent investigations by regulators. There is absolutely no reason to assume that AI companies would not use their data in any way possible to reach their objectives.

ChatGPT is the only bot that reliably cites sources (through Web search mode).

The other bots either make up links or simply don't provide any information that is distinguishable from the LLM predictive output.

Ironically Gemini is also very bad at this, while it should have been the best at Web search.

Gemini also does something very patchy, which is to provide "links" which are in fact GET queries into classic Google search. I'm guessing they did it this way because the links generated/hallucinated by the LLM were too unreliable.

The goal of automation is to automate consistently perfect competence, not human failures.

You wouldn't use a calculator that is as good as a human and makes mistakes as often.