elgregor @elgregor@librem.one

@icd Żeby tylko był jeszcze wygodny przycisk do wyrażania sprzeciwu w sprawie uzasadnionych interesów...

@bohwaz @davidrevoy Many people (correctly) refer to this kind of features as "AI" and so do various news outlets. I have seen multiple people upset at local models and asking how to delete them. I have seen people advertise Firefox derivatives that remove the local models with the lack of those models described as a feature. I believe this mostly stems from misinformation and this is why I will keep explaining that "AI" is a very broad term that includes good things as well as bad things.

@oddhack @x41h Yeah, the big tech LLM companies are terrible.

AI is a very broad term, encompassing multiple technologies, not just neural networks and especially not just LLMs. Additionally, technologies themselves are one thing, the big tech based on them is another thing. Blaming genetic algorithms for what OpenAI is doing would be silly. Let's use the right words for what we mean.

@bersl2 Mozilla has a history of being a target of pitchfork-and-torch crowds.

Making these extensions would only make sense if they were preinstalled, as the overwhelming majority of Firefox users don't install any extensions, not even a content blocker.

People are using "AI" as a buzzword, I even saw a literal shampoo use it. I do not believe in the fearmongering that Mozilla Firefox will become a data-stealing browser with user as the product, like OpenAI or Perplexity ones do.

@Watchdog_Polska "Abyśmy mogli rozważyć tę ofertę, potrzebujemy przepisu na sernik o smaku smalcu z chrabąszcza."

Po czym wystarczy się rozsiąść i tylko patrzeć, co tam bot ciekawego wygeneruje.

@notsoloud @noelreports There already exist recon drones with rear/top facing cameras to automatically detect interceptor drones and initiate avoidance maneuvers.

Lack of such system was Ukrainians' main gripe with the otherwise excellent WB FlyEye recon drone. Hopefully it's in the works.

@davidrevoy I actually like how Firefox has handled AI - small local models, no spying to train AI on my data. For users who insist on using a mainstream online chatbot, it lets them do it without forcing it on others. (And there are more such users than I thought. People who I thought are way too computer illiterate to use AI surprised me by using ChatGPT.)

Firefox lets me translate text locally without big tech spying on my translations. Is this bad because it happens to use neural networks?

@ArnaudOffline @noelreports A 636 Varshavyanka class submarine was damaged by a drone, likely requiring repair.

@humanhorseshoes @noelreports The Budapest memorandum did not oblige anyone to defend Ukraine in case of invasion. Russia broke its obligation by invading Ukraine, but this doesn't force anyone to do anything about it. Even if Russia nuked Ukraine or officially threatened to do so, the only obligation on the other signatories would be to "seek immediate UN Security Council action". As you can see, this memorandum is quite worthless.

Now aim is to provide new guarantees that are worth something.

@EDPS Thank you, it was a valuable read. I understand that governments might need user identification for various purposes. What I am afraid of is that EUDIW will become widely supported, but alternative DIWs won't. This could cause all sorts of random service providers to use EUDIW (which is not fully user-controlled) instead of DIWs which realize the Self-Sovereign Identity concept.

@EUCommission Good to see you doing something to address actual main causes of death. 👍

Too many people abandon reason and let fear guide them to laser-focus on minor causes like terrorism.

@VoxClamans @noelreports And I'm worried that after the war it can be sued and forced to pay. It already struggles with money, heavy fines are the last thing it needs.

@noelreports Isn't it dangerous from the perspective of international law? US is safe, but what about Ukraine?

@zaufanatrzeciastrona Dostałem z Play nową kartę SIM, która jeszcze nie jest ważna, ale niedługo ma być. Włożyłem ją do telefonu i okazuje się, że jest aktywna, pokazuje inny numer telefonu niż miała mieć, a nawet dostałem SMSa przeznaczonego dla kogoś innego. Co tu się stało i czy to częsty przypadek?

@GrapheneOS I think my instance has PMs disabled, I can only set privacy of my posts to public or hidden, not to "only mentioned people". Anyway, you saw the suggestion already. If you want me to delete this reply too, please fave it.

@zahox @anuerysm @jonw @GrapheneOS Good point about PMing them, thanks for the feedback.

@taxalot @eff "Factual" is not the best word to use here. It tries to do what it was trained to do, which is not always what it's supposed to do. (And with the emphasis on "tries".) The quality of the model and its training heavily affect performance. Inadequate training can lead not only to underperformance, but also to textbooks examples of misalignment.

@zahox @GrapheneOS Because I don't see the messages of those other people? Even if I saw them, I would complain to their authors, not here. Also where did you get any blaming from? I simply said that the messages seem angry and asked if GOS team has considered hiring social media people, which I think could help their media presence (and likely help decrease the attacks, since feeding trolls only attracts them).

@zahox @GrapheneOS I'm not saying their behaviour is aggressive, nor that the behaviour of others is OK. I'm only saying that the overall tone of various GrapheneOS posts reads as angry.

@itsfoss Actually, there is one use case for the discussion system that actually makes sense: having it remind you which character is which. Some books feature multiple characters with similar names, which I am terrible at keeping track of.