> What if code generation is copy-pasting GPL-licensed code in to your proprietary codebase?

This is obviously a big, unanswered, issue. It's pretty clear to me that we are collectively incentivised to pollute the well, and that it happens for long-enough for everything to become "compromised". That's essentially abandoning opensource and IP licensing at large, taking us to an unchartered era where intellectual works become the protected property of nobody.

I see chatbots having less an impact on our societies than the above, and interestingly it has little to do with technology.

> we are collectively incentivised to pollute the well

Honestly, there are two diametrically opposed incentives occurring right now. The one you describe may not even be paramount -- how hard is it to prove infringement, shepherd a case through court, and win a token amount. Is it worthwhile just to enrich a few lawyers, and get more AI-regurgitated slop to open up?

The second incentive is to not publish source code that might be vacuumed up by a completely amoral automaton. We may be seeing the second golden age of proprietary software.

Waiting for the LLM evangelists to tell us that their box of weights of choice did that on purpose to create engagement as a sentient entity understanding the nature of tech marketing, or that OP should try again with quatuor 4.9-extended (that really ships AGI with the $5k monthly subscription addon) because it refactored their pet project last week into a compilable state, after only boiling 3 oceans.

Glorp 5.3 Fast Thinking actually steals this diagram correctly for me locally so I think everyone here is wrong

I may have a new favorite HN comment.

https://news.ycombinator.com/favorites?id=Balinares says you don't

Using an LLM to generate an image of a diagram is not a good idea, but you can get really good results if you ask it to generate a diagram.io SVG (or a Miro diagram through their MCP).

I sometimes ask Claude to read some code and generate a process diagram of it, and it works surprisingly well!

It's microsoft's AI though, not even the totally crazed evangelists like that one.

You're holding the LLM wrong.

It's about as easy to hold as an old foam mattress

There's a room index at: https://search.jabber.network/

For me too, ejabberd is the admin-friendlier/lower-effort one. Being more "monolithic", your calls will work straight out of the box because it ships a turn server properly configured out of the box, manage certificates over ACME for you, etc. Prosody isn't bad but has a reputation of needing attention to be paid to which incompatible modules not to enable together and overall more protocol knowledge. Both will run on a first gen RPi effortlessly.

Re: Signal, it's even worse: they are openly opposed to federation and to letting alternative clients use their server. They demand control and obedience, which has always been suspicious-enough to defeat any goodwill effort on their side. Why would I have/want to trust them when XMPP is a viable federated alternative?

Signal focuses on security and privacy above all else, which they don't think a federated model can do well. Case in point, XMPP in practice is less secure than Signal but has the advantages you mentioned.

The other common anti-federation argument is spam/reputation, which is basically the reason email is becoming more centralized unfortunately, though it still survives.

Matrix has gotten to a complexity threshold that makes it near-impossible to have independent client/server implementations. Element is terrible, and many contenders are better in a way or another, but all lack some essential feature to turn them into practical alternatives.

There are several independent clients and servers. None of them support all features, not even synapse/element. But most keep improving steadily.

Same, years ago, I set up a Matrix server, because it was advertised as the-new-XMPP (and I had done XMPP as a user, a long time before, and thought it had been quirky enough to warrant a successor protocol).

What I found with Matrix was the same terrible experience you describe, so I gave old XMPP a new look, and it's been great and continuously improving since. I sleep much better at night having my whole family using XMPP over a self-hosted ejabberd than I can using Matrix to talk with them (and synapse... Forget using synapse federated).

So, the solution to too much AI is... Even more AI! You sound like you would fit just right at a LLM-shop marketing department.

I was like "oh common, that can't be a real comment, it's obvious to everyone how unstable this still is", then I saw that the comment was from Arathorn.

You know, for half of the time you spend commenting over here to save face (or something), you could work with your users and see their firsthand experience for yourself.

this is me working with my users and trying to understand their firsthand experience for myself :)

Literally no better audience to do this work on than places like here.

Or maybe the mindless rush to host it in azure?

Or both!