TIL, I thought it was more
keepthepace
Faire des actions pour peser dans le calcul de rentabilité de certains projets, ça oui, c'est efficace. Un truc comme Sea Shepherds je comprends.
Faire des actions médiatiques pour "se faire entendre", je comprends un peu moins: toute la population aujourd'hui a entendu parler du réchauffement climatique. Ceux qui ont décidé de ne pas y croire ne vont pas être convaincus par des jets de peinture sur des oeuvres d'art.
Par exemple les activistes ont réussi à convaincre les politiques, que le nucléaire, c'était devenu trop couteux politiquement et économiquement parce qu'il y aura 10 ans de guerrilla. C'est le plus gros "succès" activiste à ce jour (et un but contre le camp de la cause climatique AMHA). Faire reculer l'état sur NDDL. Rendre l'état parano sur l'établissement de ZAD qui peuvent lui tordre le bras. Ce genre de choses, ça a un impact.
Mais en effet, les trucs médiatiques au sein d'une foule qu'on va juste rendre hostile, je ne saisis pas.
Ben perso je paierai pas pour la musiquée générée par IA. C'est possible de la générer localement en temps réel sur un GPU modeste. Je doute que les gens se laissent avoir pendant très longtemps. Quant à la place des artistes humains, elle va probablement radicalement changer, y a des modèles à réinventer, mais les majors ont déjà pas réussi à s'adapter aux enregistreurs à cassette, sont en train de digérer l'arrivée du net, je sais pas s'il faut les attendre pour l'IA.
Strong disagree. There is a lot we can do all about it. Starting with voting and promoting politicians who want to do something about it (yes, voting matters). Then, starting a discussion about the social use of work and labor. If these entry-level jobs can be automated, what quality were they exactly testing new hires for? Let's make it an interviewing step instead.
Maybe make the entry level job one hierarchical layer higher. Instead of data entry, manage a fleet of data-entry AI agents.
And, also, wonder about the place of labor in our society. If not directly talking about basic income, talk about funding for more education, more training, on of the various modalities that exist: private or public funding, private or public college, in-company training, training through non-profits, etc.
The AI wave comes with huge social implication, and it is collectively that we will decide whether it makes individual situations worse or better.
This is only true for basic pre-training of the base model. The later stage fine tuning (I used to call it RLHF but now I think many different techniques exist) is to make the model understand the basic level of expectation. Despite having 4chan in their training set, you will never see modern LLMs spontaneously generate edgy racist shit, not because it can't but because it learnt that this is not the output expected.
Similarly with code, base models would produce by default average code, but fine tuning makes it understand that only the highest standard has to be generated. I can guarantee you that the code LLMs produce is much higher quality (on the superficial level) than the average code on github: documentation on all functions, error code and exceptions managed correctly, special cases handled whenever they are identified...
That lazy line again:
Nuclear power plants n France have scheduled maintenance in summer (because despite what the article suggests, this is a low in power consumption)
Capacity reduction happen in that context, with extremely stringent regulations in some rivers that prevent any sort of warming of the water above some temperatures.
It does not impact the power plants that rely on water evaporation which, let's remind it, release colder water than they take in (but in a lower volume)