Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Debian votes to allow "responsible use of generative AI"

lwn.net

Post reply on HN
Print view View original

441–450 of 451 posts

Re: Debian votes to allow "responsible use of generative AI"

#441

Post by aleph_minus_one »

sebmellen wrote: post #308
a1o wrote: post #284

Earlier quoted context omitted.

Nope, most of the PRs are authored entirely by agents with people instructing them to “pr famous projects in my name to increase my GitHub profile value or “hire ability”. The original authors have no idea what their agents are writing, these are mostly badly quality models (people doing this are not only cutting corners, but also using the less amount of money/tokens to do so). If the agent creates 200, 400 or whate…

Just had a thought, what if you made new contributors write a human-authored essay on why they’re submitting a patch, and then check it against Pangram? Pangram is really accurate from what I’ve found…

> Just had a thought, what if you made new contributors write a human-authored essay on why they’re submitting a patch, and then check it against Pangram?

The problem is writing such an essay takes magnitudes more of an effort for people who are not native English (or whatever the language is in which discussions are done about the project) speakers.

Also, relatedly, essays written by non-native English speakers often get considered to be AI-written by common AI detection systems, even though no AI was involved when writing them; see for example

> I'm Kenyan. I Don't Write Like ChatGPT. ChatGPT Writes Like Me.

> https://marcusolang.substack.com/p/im-kenyan-i-dont-write-li...

Re: Debian votes to allow "responsible use of generative AI"

#442

Post by aleph_minus_one »

lokar wrote: post #440
thayne wrote: post #345

Earlier quoted context omitted.

That doesn't really solve the problem. I've gotten PRs that do do an adequate job of explaining why why something should be implemented. Then I look at the code and it doesn't fully solve the problem, or it does so in a way that will obviously break something else.

A colleague once setup a PR description template that asked for short sections like: - Why is this change needed (big picture) - Why are you doing it this way (small picture) - What are the trade-offs you made for this approach - How do you know this is correct, that it works the way you think it does? - What other changes or approaches does committing to this way block in the future? etc. If you had actually thought…

> If you had actually thought through what you were doing, it was a few min to fill it all out.

... for people who are native speakers of English.

Re: Debian votes to allow "responsible use of generative AI"

#443

Post by aleph_minus_one »

lokar wrote: post #229
hypfer wrote: post #212

Earlier quoted context omitted.

Nah, that was already blown up way before LLMs existed. Foss wasn't exactly not miserable before 2022. The dysfunction just got amplified so that we cannot fix it anymore by just throwing more human resources at the problem.

I tried to nod a bit to the inconsistency. Some projects clearly lost patience with imperfect contributions long ago. But many were still pretty supportive. This dynamic also plays out in many companies. I see a lot of posts from mid to sr engineers frustrated with plummeting code quality and unsure how to manage the reviews. They are increasingly pulling away from the idea that they should mentor people to do better…

> They are increasingly pulling away from the idea that they should mentor people to do better.

Two considerations on this:

1. In the USA, it is common to switch jobs often (in opposite to, for example, Germany). So, in the US-American system it is not often worth the effort to invest a lot of mentoring ressources into some employee if he will leave soon anyway.

2. I personally observe that I often do try to give less experienced programmers good hints on how to improve, but my kind of thinking about programming is often so different that my advice is often of no use for them. Or I often get the arrogant reaction by less experienced programmers that I care far too much about obscure details. I thus seriously do believe that the only way to help junior programmers to become better is to completely crush their ego and arrogance (just like a doctoral study programme in, say, mathematics does to the doctoral students), which is of course not allowed at work. :-(

Re: Debian votes to allow "responsible use of generative AI"

#444

Post by lokar »

aleph_minus_one wrote: post #442
lokar wrote: post #440

Earlier quoted context omitted.

A colleague once setup a PR description template that asked for short sections like: - Why is this change needed (big picture) - Why are you doing it this way (small picture) - What are the trade-offs you made for this approach - How do you know this is correct, that it works the way you think it does? - What other changes or approaches does committing to this way block in the future? etc. If you had actually thought…

> If you had actually thought through what you were doing, it was a few min to fill it all out. ... for people who are native speakers of English.

Software engineering is a highly collaborative process. To be successful you need to be pretty good at reading and writing whatever language the team is using. I've worked with plenty of non-native speakers who were well above the average I see for natives. In the last 10 years I found the level of just basic writing for new Jr engineers has really fallen.

If you can't write, you can't think

Re: Debian votes to allow "responsible use of generative AI"

#445

Post by Barrin92 »

newswasboring wrote: post #143
zaneyard wrote: post #85

Earlier quoted context omitted.

With AI people aren't reading, writing, or thinking about the code. I would not say something that helps me remember a function name is comparible.

That's because you are used to it. I am old enough to have been there during the rise of things like intellisense and language servers, people were making the exact same arguments about them. That not knowing your API to the core makes you a bad programmer. I'm sure for IDEs themselves there were similar arguments and it goes all the way back to Socrates' warning against writing. These arguments always end with "but…

>people were making the exact same arguments about them. That not knowing your API to the core makes you a bad programmer.

And they were correct. You're entirely right that this time it isn't different, but in the sense that the quality of both developers and code has been going downhill for decades. The only thing that has kept this industry alive is that hardware got more performant faster than developers became more inept.

The non-strawman version of Socrates warning, that writing-based cultures would lose memory, take recorded fact at face value and mistake knowledge for wisdom was true as well. Your post is an example of this, you read something about Socrates, you didn't investigate and were fine with the appearance of knowledge, and because books can't defend themselves you mistook that for understanding.

That's what he was concerned about, hordes of people with a vague assemblage of facts but no genuine capacity to critically think.

Re: Debian votes to allow "responsible use of generative AI"

#446

Post by eterm »

dooglius wrote: post #425
eterm wrote: post #400

Earlier quoted context omitted.

That's an interesting opinion, and I certainly won't downvote you for sharing it, although it's not a position I would ever support. You are right though, that it would strengthen the voted-on position if there were more options either side of the position. I have Things To Say about LLM reviews. They are an endless grind. My biggest source of burnout right now is never being able to finish anything, because the LLM…

I'm not advocating for this, but it's a example of a position reasonable enough I'd expect to see some level of support for. I'd view it as less extreme, for example, than requiring all new security-relevant code be in Rust, which is certainly something I've seen argued for in various contexts.

While not "only Rust", that development should be done in memory safe languages is the position of the US Government:

Held both weakly by CISA:

https://www.cisa.gov/resources-tools/resources/memory-safe-l...

And more strongly by the NSA:

> NSA advises organizations to consider making a strategic shift from programming languages that provide little or no inherent memory protection C/C++ and assembly, to a memory safe language when possible. Some examples of memory safe languages are Python, Java, C#, Go, Delphi/Object Pascal, Swift, Ruby, Rust, and Ada.

From https://media.defense.gov/2022/Nov/10/2003112742/-1/-1/0/CSI...

Re: Debian votes to allow "responsible use of generative AI"

#447

Post by Tade0 »

bigstrat2003 wrote: post #60
Tade0 wrote: post #22

Earlier quoted context omitted.

A year ago I would agree. But now with sufficient context length the models actually do what they're told.

Not really, no. Nothing has changed versus a year ago, you still have to babysit them to make sure they aren't doing something insane.

Hugely depends on the size of the task. I've found that telling them to ask questions removes a lot of uncertainty.

Re: Debian votes to allow "responsible use of generative AI"

#448

Post by tredre3 »

doublerabbit wrote: post #388
phyzix5761 wrote: post #174

Earlier quoted context omitted.

The problem is you can't prove if AI was used or not so Debian's response is actually the most logical one. If anyone uses AI they're still responsible for what they submit.

What happens when the AI submitted code has been submitted by AI, Who's then responsible?

Whoever controls the AI, as per the article we're discussing.

Debian isn't a free-for-all, they don't accept fly-by-night contributions, there needs to be a real and serious person behind every contribution.

Re: Debian votes to allow "responsible use of generative AI"

#449

Post by sebmellen »

aleph_minus_one wrote: post #441
sebmellen wrote: post #308

Earlier quoted context omitted.

Just had a thought, what if you made new contributors write a human-authored essay on why they’re submitting a patch, and then check it against Pangram? Pangram is really accurate from what I’ve found…

> Just had a thought, what if you made new contributors write a human-authored essay on why they’re submitting a patch, and then check it against Pangram? The problem is writing such an essay takes magnitudes more of an effort for people who are not native English (or whatever the language is in which discussions are done about the project) speakers. Also, relatedly, essays written by non-native English speakers ofte…

I’m as skeptical of the AI writing detection as anyone else, but I’ve been trying to beat Pangram v4 for the last week and it’s essentially impossible.

Re: Debian votes to allow "responsible use of generative AI"

#450

Post by newswasboring »

Barrin92 wrote: post #445
newswasboring wrote: post #143

Earlier quoted context omitted.

That's because you are used to it. I am old enough to have been there during the rise of things like intellisense and language servers, people were making the exact same arguments about them. That not knowing your API to the core makes you a bad programmer. I'm sure for IDEs themselves there were similar arguments and it goes all the way back to Socrates' warning against writing. These arguments always end with "but…

>people were making the exact same arguments about them. That not knowing your API to the core makes you a bad programmer. And they were correct. You're entirely right that this time it isn't different, but in the sense that the quality of both developers and code has been going downhill for decades. The only thing that has kept this industry alive is that hardware got more performant faster than developers became mo…

I am not arguing that things don't change and certain skills dont atrophy. I am arguing that the atrophied skills are compensated by new skills we develop. But your personal attack makes me feel like your purpose is to win, not to be right.
Post reply on HN