https://github.com/RsyncProject/rsync/issues/929#issuecommen... has a partial list

That is a list of problems, not causes, which is what was being asked for.

It's possible it's some LLM randomness that caused bugs. That would suggest that some AI hygiene is in order.

If it is because of behaviour changes necessary to fix security issues, then the regressions might be from things that relied on unsafe features.

Do we know of actual specific causes yet?

Yesterday's comment listed suspected commits alongside the issues: https://qht.co/item?id=48334270

Imagine if someone tried making a sub from that, and then it delaminated lol

Carbon fiber

It's gross to scroll on desktop as well.

It's gross to scroll on tablets as well.

Is there any source with just the plain text? The css styling is headache inducing and reader mode doesn't work or has been defeated.

Firefox has a handy "Reader view" (Opt + CMD + R on Mac) that you can activate to get a stripped down view of just the text on the page. Unfortunately, it also removes the images which contain some of the sources they use.

More like repeating their firmly entrenched preconceptions. Their claims may (or may not) be right, but there's very little if any new evidence being provided by either camp.

The real uncomfortable thing is that because we cannot confidently know, the moral defacto position is to treat them like they are.

I know you "feel" pain because I can poke you and observe the result. If I do it enough there is permanent damage.

Show this same phenomenon exists in LLMs.

We did exactly that and published it last month right here: https://www.anthropic.com/research/emotion-concepts-function

They are confidently hallucinating a factual statement. Which is funny when claiming that confident hallucinations are the proof of LLMs' lack of intelligence.

One side is making a positive assertion. The other is making a negative assertion. One side can prove it's right. The other, logically, cannot.

One camp has to offer it's proof. If it has none then that _in and of itself_ is highly suggestive.

People have fully turned their minds off on this subject. It's disgusting.

> Or, as it turns out, from within. (Looking at you, Danielle Smith).

Pretty sure that's a prong of the southern strategy.

Relax, it's niche internet points, you'll be fine lol

> it could also be fraught if Stripe was in the business of blocking customers from their entire network based on one vendor's complaint

“You probably don’t want a system where one annoyed merchant can get someone blocked across the whole Stripe payment system. But there’s a pretty big gap between “automatically block this person everywhere” and “thanks for the screenshots, please consider Radar”, and this is where it gets frustrating.”

There is no gap between those things. Any fraud signal at all either causes people to get blocked, then it's a cross-merchant block, or it doesn't cause people to get blocked, then it's useless.

The gap between “block on first signal” and “ignore the signal entirely” exists, and it is not small.

The gap between "block transaction" and "don't block transaction" is nonexistent and that's the only thing that actually matters. What else are you, as a payment processor, gonna do besides block a transaction? Allow the transaction but send the user a sternly worded email?

False dichotomy is false; good day.

“Please don't post comments saying that HN is turning into Reddit. It's a semi-noob illusion, as old as the hills.” --hn guidelines (there are links to examples in the original)

They failed far more than 12 landings before they started reliably landing boosters, and people made the same “maybe this is a dumb idea” comments back then too.