bornfreddy

Nice writeup. But IIUC you can change it so that it automatically updates the time every 6 months?

Oh, the problem is not with agents that escaped, it is with the people who prepared the terrain so that they broke into 3rd party property.

That's awful for the misclassified customers who will not get the attention they need. Not to mention that people talk and if you treat some customers badly, the others (the ones from the top of your pile) might reconsider working with you. That doesn't automatically make this a bad idea, but I would tread very carefully.

I mean there is an abundance of shovels - but everyone wants to dig with excavators, and there's a shortage of those.

It's all downhill for OpenAI and Anthropic investors, because they will never recoup what was invested. NVDA is safe(r) - Huawei first needs to catch it, and as AMD learned, it's a fast moving target.

This. A competent admin and a clueless user is a perfect match for Linux. Either clueless admin or "competent" user however... not so much.

Genuinely curious: why do you use them then? It's not like there is no competition in this area.

If you want to be in the business of checking other people's PRs, sure. I would (personally) hate that, and I suspect the maintainers too. Money is not a solution to everything. Also, I don't see people standing in line to pay to get their PR reviewed.

It's as if building something was not a bottleneck for startups. Go figure.

If you can't tell that the contributor used AI, the PR is good - how would they know to forbid it? This rule hopefully stems the tide of low level fly-by contributions. As they said, it's also not set in stone, there will be enough time to change it later if needed.

It also creates wrong expectations. If I pay you (even if it is $1) then I expect you to take a look at my slop and argue with me how to make it better. Just closing it will prove that you're a scammer (and not even a successful one, because you'll get at most $5 altogether).

None that I can rely on being available wherever I see a terminal. It's bash or sh as far as I'm concerned.

On the other hand, this description of situation does make one less concerned about the whole super intelligence / extinction thing, so there's that.

Actually... Server can detect if you are piping or not and serve a modified version for inspection.

But really, there is no reason not to use prebuilt packages for distribution. Curlpiping needs to die.

Why don't you? No time like now.

> Nowadays I would attribute their response to chatbots considering how elaborated and friendly, but yet totally useless their messages were, but this was before LLMs

Outsourcing was a predecessor of LLMs.

Also, how does this number (53%) compare to pre-LLM numbers?

That is a good point, but the problem with ML/AI is that anyone can do it. It will be difficult to stem this tide.

Which doesn't mean they didn't escape.

Agree! My only concern is - is the judicial system fast enough, and resilient enough? Or will these creators get "off the hook" by using their agents to find loopholes, sway public opinion or even convince Trump to grant them immunity?

Still, I have no idea why OpenAI & co. are not being sued for these hacks.

"Move fast and break things" was supposed to break your own things, not other people's.

Wow, that's genius. Basically, user gets a VPN but so does provider, in a sense.

The agents also overload you with walls of text, always including lots of less relevant information. It's like they want you to give up and let them drive.

I already did, before AI:

  import antigravity

Well, it's not actually a trap, since it's not the management who pays the price - it is investors, employees and the society at large that do, while the MBAs have already moved on to their next victim.

Problem with incentives, I would say.

I think the taste of their beer would have been much different from what we drink today?

Were the two cars priced the same? Is public image of both brands the same? I'm not sure this was a good A/B test.

At its core it's a question of control. If I use this library or app, will I be safe? Will I be able to repair it, use it in the future, pay someone to fix it? Will it still be developed in a few years?

FOSS is a great answer to the first few questions, but it fails at the last one. As a user, yes, I would prefer a cloud-protection license, if it means that software will still be developed in a few years time. It's not like I want to use this software to compete with the author, I'm happy not to be allowed to do that.

They probably broke something on purpose so that they are not left out.

Often for trivial things that LLM is proud that it has noticed but bear no load whatsoever.