Notices by Stefano Zacchiroli (zacchiro@mastodon.xyz)

I said what I said about Mastodon moving too slowly to matter for the US midterms, so BlueSky being our best bet.

I said what I said about me being wrong about thinking that Mastodon would get it together by now.

I said what I said about BlueSky making it okay that Mastodon moves slower, and it not necessarily being over in the long run. We're watching people leave Twitter at a rate of 1 million a day. If BlueSky does go bad, and a better option is available, I anticipate a similar exit.

2/N

Honestly baffled

#fediverse #twitter #bluesky

@ttyS1 @lunar yeah... :-(

« #ReproducibleBuilds mourns the passing of @lunar » https://reproducible-builds.org/news/2024/11/14/reproducible-builds-mourns-the-passing-of-lunar/

@jzb @lunar thank you for your message. He truly did and will be missed.

In memory of @lunar https://upsilon.cc/~zack/blog/posts/2024/11/In_memory_of_Lunar/

#blogpost #Lunar #RIP

@BrKloeckner @3j0hn @rfc1149 my take is that the scientific editor didn't see that note (yes, if true this means they have been sloppy). Because if they did, they would have either backed the authors, relieving them of the "duty" of adding the citations; or backed the reviewers, asking the authors to put the references where they belong and discuss them properly.

#peerreview is doing great: « As strongly requested by the reviewers, here we cite some references [[35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47]] although they are completely irrelevant to the present work. » from the last paragraph in the introduction of https://www.sciencedirect.com/science/article/pii/S0360319924043957

h/t @rfc1149

@Andres4NY the ALT description on this picture is a piece of art. Kudos.

I'm looking for guest lecturers for a university-level #Rust class in #Paris. Favorite topics would be security-related, and/or Rust in the real world (a return of experience kind of thing), but #RustForLinux would be awesome too.

cc: @rfc1149

@sylvestre @fabrice @rfc1149 et tu n'as que du bien à dire du cours, n'est-ce pas? :-)

@lxo @chaz AFAICT even OSI recognizes this. Their main arguments against mandating training data in OSAID are of a pragmatic nature, related to (1) the current state of the industry, (2) the legal regimes that apply to data, which are different from those of (free) software.

I disagree with the decision taken based on those arguments, but I understand them. Either way, they don't call into question the fact that training data is *better* than no training data to exercise user freedoms.

@lxo @chaz It's a good approach, but I don't think it's needed.

If we start from first principles, there's no doubt that to fully exercise freedoms of study and modify, you need the training data. (You can exercise *some* of those freedoms even without training data, but in a suboptimal way. I can give precise examples if you're curious.)

The "data is too big" problem is IMO a distraction. There are relevant ML systems that are small enough to make retraining them from scratch viable.

Does anyone know if the training dataset for the ML weights shipped with #gnubg at https://git.savannah.gnu.org/cgit/gnubg.git/log/gnubg.weights is available from anywhere?

And, if not, is the #FSF going to exclude gnubg from #GNU? (It would be the right thing to do, short of obtaining and shipping the dataset as gnubg's preferred form of modification, under a #freesoftware license.)

@zimoun I'd take "normal" November rain over floods...

@pchestek it might have been the famous drop (of alcohol) that made the (cocktail) glass overflow, yeah

@jzb @slothrop you folks across the pond are definitely excused!

let's do #drynovember

@khinsen @civodul also, while "statistically plausible code" is an amazing negative term for what code LLMs do (chapeau to whoever coined it 😉), it hides the fact that code bases are full of *recurrent patterns*. Some of them should be refactored away (but aren't!, for whatever reason). Others are there to stay and part of the (broader) coding style of a project. One thing code LLMs are really good at is spotting, replicating, and instantiating those patterns (for better or worse).

Very happy to have contributed to this: "SFC Announces Aspirational Statement on #LLM -backed generative #AI for Programming" https://sfconservancy.org/news/2024/oct/25/aspirational-on-llm-generative-ai-programming/ #freesoftware #opensource