<unnamed>

Yes?
There is a link between caring about privacy and being anti-AI.
There is a link between caring about the environment and being anti-AI too.
There is a lot of these links. That’s why it’s biased…

2026-08-27 13:47:49 CEST

- reply

Your last sentence is weird?
I mean, are you acknowledging that there is a link between caring privacy and being anti-AI?

2026-08-27 13:35:24 CEST

- reply

Maybe that bias isn’t important, but you have to prove it, while you’re assuming it don’t have any importance until proven false, which is a really bad way to do statistics…

2026-08-27 13:33:02 CEST

- reply

If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence, not a proof that X is representative of the population either.
You are saying that people can be representative on certain subjects and not others, which don’t make any sense.
Either is representative on all subjects, ore none.
If X have more people accepting facsism than the Fedivers, then none of them are representating on any other subject, there’s already a bias.

2026-08-27 13:26:00 CEST

- reply

And the Fediverse is a privacy-friendly bar.
If you aknowledge that X and the Fediverse have different people that have differents opinions, then you’re already destroyed your own point.

2026-08-27 13:18:17 CEST

- reply

Being online does not mean accessing to everything everywhere on the net.
> "The Fediverse isn’t a random sample of the world population" - the statistics tell you otherwise
Does it? Do you have any study or numbers to prove it?
The fediverse have way much french people in proportion than the world have, that’s already a big bias…
And there is more LGBT people in proportion than other social medias too, that’s another bias.
There is more people that care about privacy than in the general population, too.
I mean, the fact that there is a lot more of people on BlueSky than on the Fediverse is already a proof that the Fediverse isn’t a random sample (BlueSky isn’t either, for what it’s worth, but they are populated by differents persons that will have different opinions on some subjects).
Like, if what you say is right, then we couldn’t find a subject where BlueSky, X and the Fediverse have different results.
But there is some…

2026-08-27 13:06:45 CEST

- reply

"In my local reality" is not a random sample, nor are online circles.
The Fediverse isn’t a random sample of the world population.
DDG users aren’t a random sample of the world population.
Yeah, the link was shared in a lot of places, but these places were not random. When you share a link, you’re doing it in your own "local reality".

2026-08-27 11:44:45 CEST

- reply

But people using DDG or being on the Fediverse is already a huge bias…

2026-08-12 13:11:45 CEST

- reply

The insurance part should be managed by the state directly, the potential catastrophe do no more harm than solar when you compare it at equal production (nuclear is for electricity safety the same as planes are to transport safety: it’s impressive when there’s a crash, but they are way fewer crashes than anything else), about water level we can always use oceans for countries that have an access to it, and the Palo Verde nuclear plant is in a desert, using waste water: if we have enough water for people, then we have enough for nuclear (if we don’t, then nuclear will be the last of our problems). Also, solar don’t like heat, and the best solution we have to cool it is water too, so…
The nuclear waste is less of a problem than the atmospheric waste of fossils (it’s easy to collect and store securely), and even renewables create more environmental pollution because of their construction (more metal and other material extractions for the same energy output, and it’s even worse with batteries, did I already say that nuclear produce **a lot** of energy?).
Decommissioning is not a problem, we already deconstructed some plants completely? It just take a long time, but since the lifetime of nuclear plant is really long, a long decomissioning doesn’t look like a big deal.
https://pleroma.breizh.pm/media/67/db/fc/67dbfc319ccabc77ad5e9bda7d070fc9dcdba953856e28c093af649710cff109.webp

2026-08-12 12:58:04 CEST

- reply

Yup, for the second part, that’s good advantages, but that doesn’t make solar magically the best way. It still use way more metal (and with batteries that’s even worse), we could make shade with wood and keep metal for better use for example.
I’m not sure about the price too, but there is a lot of factors to take into account (nuclear is expensive because of the high cost of private investment, if it was public it would be way cheaper).
Anyway, my point was that good arguments for solar aren’t necessary good arguments against nuclear. And in the short term, we need both anyway (we can’t make enough batteries in the time we have and we should use them in priority elsewhere, like transports for example, and we can’t build nuclear fast enough to count only on it).

2026-08-08 01:17:18 CEST

- reply

Well, good luck producing electricity at night with a solar panel?
I mean, agrivoltaic do looks like it can have great uses, but using it as an argument against nuclear is a bit… weird?
Especially since it’s not solar, but shade that was beneficial. The fact that we can use this to make electricity as a bonus is great, but that’s it.

2026-07-31 18:58:31 CEST

- reply

Toi aussi t’apprécies le porno groove ? :yay:

2026-07-14 16:15:55 CEST

- reply

But it’s not simple either? You lost me here.
(maybe that’s because I’m not fluent in english)

2026-07-14 16:14:00 CEST

- reply

But that was my point. When I was saying it was complex, I meant not easy…

2026-07-14 16:07:15 CEST

- reply

When you play with expansions, and start doing combinations, it can get very hard to understand pretty fast.
Even the `^string1^string2^` is a shortcut for `!!:s^string1^string2^`.
And you can have things like `!-3:4*:gs/old/new/`.
Personally I didn’t play that much with history, but I did play with paramater expansion a lot, and this is in one of my scripts at work: `${dispo[${keys[$choice]}]##*:}`.
In personal scripts, I have this kind of things:
```bash
title="${infos[1]:-${infos[3]}}"
[[ -n "${infos[4]}" ]] && length="$(( 10#${infos[4]%%:*} * 60 + 10#${infos[4]##*:} ))"
```
It’s not that hard to understand, but I won’t call it easy either.
And it’s probably because of the way I choose to write it than because of Bash syntax, sure, maybe that’s because I’m the grey beard here, I guess. But I have the honesty of saying that it’s not easy at least ahah.

2026-07-14 15:40:33 CEST

- reply

You mean @npub1hpw…pnfc ?

2026-07-14 15:37:14 CEST

- reply

Bash history expansion always looked very powerful, but also very complex at the same time, so I didn’t bother learning it.
This particular shortcut looks like it can be useful, that being said.

2026-06-04 12:30:44 CEST

- reply

https://pleroma.breizh.pm/media/f0/1b/86/f01b86c144a25a335d4f5a03957a064c728ddb361f16d93523164a145157ec1f.webp

2026-05-23 19:11:40 CEST

- reply

Je serais curieux de voir le résultat cela dit. Le compteur geiger serait la preuve que ce béton est effectivement inoffensif sur le plan radioactif…
(et voir cette personne expliquer le fait qu'une belle maison en granit est pire que ce béton :D)

2026-04-09 14:38:44 CEST

- reply

Règles du jeu :
- Se saisir du livre le plus proche
- L’ouvrir à la page 42
- Trouver la 2ᵉ phrase
- Poster ladite phrase dans un message avec le # et écrire les règles en réponse
Ne pas chercher son livre favori, celui qu’on trouve le plus cool ou le plus spirituel. Vraiment : prendre le plus proche.
(question : on peut prendre un magazine aussi ? Là j’ai pris le roman le plus proche, mais j’ai pas mal de magazines qui sont plus proches en général)

2026-04-09 14:36:57 CEST

> À part l’étrange courbure de son visage et ses cheveux épars, il avait des yeux d’un bleu très clair, presque blanc, la peau translucide et le teint terreux.
#page42

2026-01-29 23:27:12 CET

- reply

Well, some times ago, open-source TTS was pretty lacking, but now Kaldi / Sherpa is pretty good, did you check it? If yes, what was the problem with it?

2025-12-11 11:23:42 CET

- reply

D’un autre côté c’est le type d’erreur qui me surprendrais pas de la part d’un LLM, après tout ça sait pas compter ces trucs ahah.

2025-11-24 15:53:28 CET

- reply

https://www.youtube.com/watch?v=JvUMV1N7eGM

2025-11-19 18:30:15 CET

- reply

Actually, it is, yeah. Well, in some case it could be true (like when they replace outdated stock ROMs), but’s generally, other ROMs are worse than the stock one in terms of security (about privacy it’s a harder to say, since they fix some stock problems, but the security breaches can create more privacy problems than it’s solving).

2025-11-19 18:26:27 CET

- reply

You’re talking about the part he said "You may be talking about GrapheneOS scams ?" I guess?
The question was here. Okay, after it’s an affirmation but it’s more like a formulation error.
And the response is, no, Graphene didn’t talk about the scams, but they were really talking about Lineage and their forks.

2025-11-19 18:16:51 CET

- reply

Like he said, I know him. I think he didn’t realise it was a false claim… it was more probably a genuine mistake.

2025-11-19 18:12:03 CET

- reply

Well, it’s not a troll, only someone that don’t know what is this thread about, and try to understand. But only receive blunt response, and now even rude ones. That’s not helping…

2025-11-05 17:56:07 CET

- reply

What, why. I mean, it’s already another account so people should see if it’s a normal post or a PeerTube post 🤔.

2025-08-21 22:21:17 CEST

- reply

Si seulement. Quand t’as une bulle qui éclate et qu’une crise économique arrive, c’est toujours les même qui trinquent. Certes, les plus riches perdent parfois quelques plumes, mais ils sont rarement à plaindre.

2025-07-27 17:20:55 CEST

- reply

J’ai pas d’idée, mais certaines de ces propositions m’ont fait penser à cette chanson…
https://fichiers.breizh.pm/Musique.opus/Tri_Yann/2007.Abysses/07.Petite_Sir%C3%A8ne.opus

2025-06-26 00:41:02 CEST

- reply

I guess NC and ND already forbid most of the AI use. They just don’t care about it…

2025-06-17 13:29:08 CEST

- reply

Yeah, I already know that Graphene does a better work. That’s not the problem here, like I said.

2025-06-17 10:34:59 CEST

- reply

Even if you’re right, your tone is incredibly rude. I don’t think that’s the best way to convince people ^^’

2025-06-09 14:35:13 CEST

- reply

Yes, I know, and environment recognition can be done in a more... flexible way, but we need to look at what we call AI.
For example, automatic parking on cars inevitably has an adaptive component, but we can be sure that it won't do anything stupid either (and for example, Tesla chose to go with full machine learning on it, and they gives much worse results than the competition).
Agriculture was already using semi-automated vehicles, notably via very precise GNSS positioning, for example (and besides, I'm not convinced by a tractor without a cockpit, having a supervisor seems to me to be preferable no matter what level of automation you reach, except in some special cases (subways in particular, since they have a very controlled environment, but then you don’t need any kind of AI)).
In short, AI as it is generally referred to today is not relevant here, in my opinion. Then again, maybe it's a completely different type of AI that's relevant in this case, but that's not really the point (and it's been around a lot longer, and most people don’t complain about it).
That said, I interpreted your message as “these new AIs can have a relevant use like here", but perhaps it meant “there are other types of AIs that shouldn't be mixed up with the ones they're trying to sell us everywhere, and which aren't to be thrown away”, in which case we agree and I'm sorry for the off-topic.

2025-06-09 10:02:37 CEST

- reply

For this kind of thing I think a deterministic algorithm is probably more relevant.
(and that goes for all autonomous vehicles, whether subways, trains, cars or tractors...)

2025-04-05 23:10:50 CEST

- reply

The disadvantage is that it will delay the development of a fast neutron reactor, which could allow depleted uranium stocks to be used up (and maximise the use of the raw material, which even with a more efficient extraction technique, seems preferable to me). Even if we are talking in millennia...
It would have been nice if this techno had arrived just after we developed (at an industrial scale) the FBR technology, so we could have had the optimal cycle.

2025-03-21 22:30:13 CET

- reply

Magnifique, par contre ses yeux me perturbent sur le gros plan ^^’

2025-01-06 21:57:47 CET

- reply

The best option is to just remove it 🤷. Or making it a optional text field, but… storing this kind of information isn’t great if you don’t really need it, privacy-wise.