Tired Bunny

@alice @noboilerplate Copilot reportedly can appear in participant list for PRs despite never partucipating in which may cause false alarms with that approach.

@LukefromDC@kolektiva.social @aral@mastodon.ar.al @EUCommission@ec.social-network.europa.eu Nobody even in Russia itself wants to install Russian government controlled CA on their devices.

@LukefromDC@kolektiva.social @EUCommission@ec.social-network.europa.eu @aral@mastodon.ar.al I would generally treat any CA hosted in Russia as potentially controlled by government, due to it being a big and noticeable piece of public infrastructure that requires a lot of trust. (From my understanding, having CA root certificate installed to device or browser means it can potentially do MITM on any site browsed on it, not just ones actually signed by that CA.) It may be a "third party" but actually a front for a big company with close ties to government (happened with third party telegram client). Other possibility, hardware running it may be seized in a police raid at any moment, to avoid being raided company behind CA would have to answer every polite police request to give any data and potentially to do MITM attacks on their behalf. Also, Russian government may shut down any company that works with "extremists" and in general whatever is forbidden by law (a lot, and in very vague way), meaning this CA will only be useful only to sqeaky clean sites that avoid politics or ones specifically licking boots of current russian government. With that, most people in Russia itself (knowledgeable enough about CA) would prefer any foreign one over any local one. Though I guess, for most people not in Russia it may actually be not as bad, depending on their threat model and whether they want to make everyone using their site to have to install a third party root CA certificate that will likely never come preinstalled.

@kaiiak That looks pretty similar to astronaut tail concept (mechanical tail to position yourself in zero gravity, though it was designer student concept, so I am not sure if it would be effective)

@billgoats Are there some good resources on which models can be jailbroken?

@yura@udongein.xyz Because of how specific this scenario is, one pretty simple principle accounts for it - When rescuing people, don't create new victims Most rescue teams of various kind operate under this principle

@LunaDragofelis @lordofthemoon I am kinda wondering if trademarks have something simmilar to fair use exception for parody. If they don't, that's my wishful thinking, but they should get one. This is clearly a parody and not a competing product.

@LexYeen CC: @yura It seems that there are more details and more people are talking about it.

@zsmb13@androiddev.social And a lot of usecases people say, about this thing protecting from literal spyware/stalkerware, are assuming that it is not google-certified already, and that it cannot get certified either because ones who made it are affiliated with governments google must abide law of, or with by paying google under a different identity every once in a while and hiding app functionality better.

@fedorchib@mastodon.ml @drq@mastodon.ml @th3rdsergeevich@mastodon.ml Я уже это написал, но подписываюсь про инди, и добавляю, что обычно инди с онлайном полагаются либо на самохостовые сервера пользователей, либо на провайдера вроде Steam API (конкретно его легко заменить на опенсорсный эмулятор работающий в локалке)

@th3rdsergeevich@mastodon.ml

• Будет больше Minecraft/Terraria и меньше "Игра анонсирована в один день с анонсом закрытия игры через пол года"
  • Кучу инди игр пользующихся Steam API могут просто патчем встроить что-то вроде Goldberg Steam Emulator если Steam API закроется (и Steam сами могут сделать/будут должны это), и мультиплеер будет работать в локалке
• Моды на мультиплеер не платные и не продаваемый продукт, и даже так им почти ничего не стоит сделать бинарник серверов (большинство модов на мультиплеер и так самохост серверов или локалку делают).
• Коллизии на лицензируемый subcontent - по сути единственная реальная проблема, разработчикам придётся лицензировать всё с учётом этого закона и возможно платить немного больше за это.
• Законы не имеют обратного действия, разработчики уже сейчас могут работать над тем чтобы в ранних стадиях игры делать всё как лучше для всех, и делать оффлайн режим/мультиплеер по локалке вместо того чтобы делать planned obsolecence

@Njord

Ethical issues are one thing, but there are also issues that aren't ethical per say, but come from the same "big bag of issues":

• Requires a lot of compute for minimum usability. Because of that, it's always a service, one that could be taken away, become worse, become unavailable temporarily or even forever.
• Factual mistakes it makes may be harmful in cases where they look believable and someone relies on LLMs.
• It still can and will be used to generate stuff of questionable quality/factuality in bulk, adding even more noise to social media and web as whole, both of which were already justifiably called "a big junkyard" before LLM and image generators were popular.

If solutions for those issues are somehow included in theoretical solution for "ethical issues", I would choose between "Basically just a toy" and "Useful in limited cases". But I think that while those issues are connected to ethical ones, they are still separate, so I'll say "Worse than useless".

@yura@udongein.xyz (with exception of armed forces during rescuing captives, they are risking themselves, and during captive trades where sometimes volunteers are stepping in as temporary captives to guarantee that their side doesn't betray and start firing once captives are traded)

@inex@pony.social @kafazen@lamp.leemoon.network monkey paw реализация этой мечты потому что с шансом 60% советует другую проблему вместо решения ("а почему не [название совсем другого софта]?")

@lumi @Lugrim @adzetko I tried making my own recently, though with caveats*, so I don't think it's exactly what you need, but I'll throw it in anyway if maybe someone would use it as base or inspiration. https://codeberg.org/mailbun/ai-contribution-guideline *it's a contribution guideline instead of pledge, it is a bit wordy, stuff not related to llms is not even applicable to most projects, and most importantly not a blanket ban on use of LLM, and selection of allowed uses is opinionated

@davidrevoy@framapiaf.org @voxel@infosec.space I am kinda wondering if she will keep the "dumbed down" parrot after whatever the ending to the avian intelligence story arc will be.

@aburka@hachyderm.io @caseynewton@mastodon.social Specifically with Gemini chatbot, asking to identify Gemini pictures works, as it redirects image to some software that identifies hidden "watermark" that Google's image generation models add to image. It's really scummy of Google to tie this software to their chatbot, since it will cause confusion among less tech-savvy people that will attempt doing this with other chatbots and non-Google-generated images. (And it will futher the confusion about what features are actually added to chatbot and what is LLM just lying about doing something)

If you got a follow request right now, it means I didn't follow you from this account but followed from @untsuki@udongein.xyz

And I am importing follows.

I think I'll try being more active on this account, for various reasons.

@TobyBartels I heard that they weight the same amount in the atmosphere of the earth, but their objective weight is slightly different in vacuum, as air pressure changes weight ever so slightly (like water pressure allows objects with high volume to float), so feathers are a bit heavier if measured in earth.

@LukefromDC@kolektiva.social @EUCommission@ec.social-network.europa.eu @aral@mastodon.ar.al Good point, though to my knowledge, US doesn't do MITMs and police raids on behalf of Russia YET, and US companies in general are at less pressure from US government, so it's more of potential threat than one I could assume is used in practice right now, like with Russian certificates. That's a whole can of worms with how SSL root of trust is right now, where current default security features like root CAs are potentially worse than self-signed certificates depending on threat model.

@nixCraft@mastodon.social It's such a shame that California lawmakers had a pretty good idea (their "age verification" is practically parental controls, it doesn't actually verify anything, it just requires OS storing age group set by user during setup and telling it to app), but decided to go the worst way around it, imposing a law on OS devs instead of working with them to create a standard(s)/protocol(s) that would be recommended for OSes but required for age-restricted apps to respect. It could be a great alternative to privacy invasive age verification, and because of that, OS devs would be compelled to implement it, while enforcement on children's devices would be a parent's responsibility, as it should be. But the way it's implemented right now is, of course, either very ignorant, or malicious with intent to use it as precedent for further changes, and I am genuinely unsure which it is.

@sjvn@mastodon.social ipc vs oled screen comparisons be like

@eevee@slime.global @hailey@hails.org Good is not unix