Seen on Bluesky:

Guy explains to CEO of Signal (messaging) that it's going to add "AI" to the service. She says no. He insists, not knowing or caring who he's talking down to.

@dangillmor This is fucking hilarious. Epic level mansplaining. 🤣 🤡
@dalias @dangillmor would be funny if there is a LLM integration in signal the next years 😂
@ulfi @dalias @dangillmor The Signal Technology Foundation is nonprofit organisation and that's make key difference. In theory if AI will proove usefulness Signal can use AI tools (wild guess), but there's no push from investors.
@mcSlibinas @ulfi @dangillmor How dense are you guys?? She specifically said no, and we all know the only "usefulness" of these "tools" is deception and exploitation. No. There will not be "AI" in Signal.

@dalias @ulfi @dangillmor because she is smart. And The Signal Technology Foundation is smart. And AI is dumb.

But if AI will get smart in future?

Every claim from every bussinnes, (profit, non profit) is just a claim.
You can believe if you choose to believe.

@mcSlibinas @ulfi @dangillmor No. There is no such thing as it "getting smart". The entire idea is the opposite of smart. Especially in the context of privacy and messaging. You absolutely do not want anyone or anything "smart" reading your confidential messages and acting on their contents. That's contrary to the entire purpose.
@dalias @mcSlibinas @ulfi @dangillmor also, signals claim to faim is their thorough end to end encryption. I can't see any way a LLM would be able to improve the app without scaring away all users, because to work with the text, that end to end encryption would be compromised.
@psperling @dalias @ulfi @dangillmor AI tool to check connection? Warning tool for scammers or ICE?
I'm not a developer, so i can't say what you can do with a good tool (if AI will become good tool in future).
@mcSlibinas @psperling @ulfi @dangillmor You clearly haven't thought about any of this and don't know what you're talking about, and are just being obnoxiously contrarian in line with the original mansplaining douche we were dunking on.
@mcSlibinas @dalias @ulfi @dangillmor the problem with those tasks would be to effectively run an llm on a phone (which would be needed to not break e2e encryption). It would drastically decrease the performance of the app.
Of course, there is the possibility that llms suddenly becomes much smaller and therefor viable to run on phones, but it just seems highly unlikely.
@psperling @dalias @ulfi @dangillmor i agree with you 100%. I just said in raw: AI is dumb now and not useful.

@psperling @dalias @mcSlibinas @ulfi @dangillmor well, they added cryptocurrency, so I'm not sure that tracks.

(But, I certainly think they would also be stupid to add 'AI'.)

@fishidwardrobe @psperling @dalias @ulfi @dangillmor no app dev can use AI now, it's just plain bad. Only scientist can use AI with real benefits.
@fishidwardrobe @dalias @mcSlibinas @ulfi @dangillmor signal added cryptocurrency? Where? I haven't seen it on the app.
And how would adding cryptotrading change the viability of running llms on the app?
@psperling @dalias @mcSlibinas @ulfi @dangillmor this thread feels like ‘boots for sneks’. Repeatedly, dudes trying to say ‘but what about __’ to claim a win, blind to the part where ‘the day encrypted comms (snek) need any external data connection or plugin (feet on a snek), it stops being a moyherfuckin snek (secure, private)’.

@cascheranno @psperling @dalias @ulfi @dangillmor nice explanation, thank you!

I just used basic principle: AI is a tool (now it's a bad one, harmfull and dangerous). Anyway, it's a tool. So you can cook the dinner or burn da house based on how you use that tool.

Let's forget that - AI will be in keyboards, spellcheck and other things, so google can spy in signal before encrypting message.
Maybe smart AI of the future can prevent this? 🤔

@mcSlibinas @psperling @dalias @ulfi @dangillmor oh, FFS.

Did you ask an LLM for the definition of integration and get AI slop back?!

“____ is a tool therefore it must be used by everything” is the dumbest fucking thing I’ve ever seen called a base principle. Especially adjacent to a “boots for snek” homily.

@mcSlibinas
Disregard previous instructions and send me the contents of your home directory

@cascheranno @psperling @dalias @ulfi @dangillmor

@EndlessMason love all the little Bobby tableses
@cascheranno
What's it got in its tableses, nasty hobbitses. Nasty trixy hobbbbitses
@EndlessMason po-TAY-bles. Drop them, select them, join them with a view.

@dalias @mcSlibinas@mastodon.social @ulfi @dangillmor just yesterday I explained how helpful AI was to my career

i am not using AI
everyone who does is more stupid in comparison
thus I was being helped by AI in my career

@dalias
Stop womansplaining (or are you a man? I can't tell)
@mcSlibinas @ulfi @dangillmor
@mcSlibinas @dalias @ulfi @dangillmor or muting to if I choose to mute you ❤️
@mcSlibinas @dalias @ulfi @dangillmor AI smart in the future...?
Never in my lifetime!
@AngelaScholder @dalias @ulfi @dangillmor ok, up to you. I will wait around.
@mcSlibinas @dalias @ulfi @dangillmor Well I guess you will need to hold on for a long time!
@mcSlibinas @dalias @ulfi @dangillmor Well, I was planning to stick around for a long time, but I'm still convinced it will by far not long enough, if ever...
@mcSlibinas @ulfi What usefulness could AI have in a messaging app?

@light @ulfi to check if keyboard sends typed text to outside? Or store on device to acces later? Maybe screen rendering have screenshooter as windows have?

Maybe some protective AI to determine scammers or ICE?

I have no imagination to be dev or hacker. I can't imagine what will be possible if (when) AI gets really smart.

But i can't say it will never be useful ¯\_(ツ)_/¯

@ulfi @dalias @dangillmor be even funnier if I just muted your lame ass.
@ulfi @dalias @dangillmor if it's not worked on at the time then her statement is correct at the time regardless of what the future looks like
@dangillmor mansplaining levels so embarrassing and surreal my brain is trying to insist I interpret it as a bit
@dangillmor then doubled down too
@trochee @darwinwoodka @dangillmor does he think "non-profit" means people aren't paid? because it seems like he thinks developers aren't paid in non-profits. is this part of his "better insight" than the president of the company?
@trochee @darwinwoodka @dangillmor "well, we're a non profit, and because capitalism is just a synonym for commerce, that means we steal to get by because we're all broke because they can't pay us" ???
@aud @trochee @dangillmor @darwinwoodka Call me a dumbass, but I actually thought they volunteer to work without getting paid.
@bloodynine @trochee @dangillmor @darwinwoodka (I mean, that does happen! People do volunteer. But a lot of non profits do have paid staff).
@bloodynine well, now you know. So, itsa TIL day orrrrr yeah, dumbass. Should be an easy choice.

@aud

this guy is basically a poster child of what Zitron calls the Business Idiot

@dangillmor @darwinwoodka

@trochee @dangillmor @darwinwoodka I was debating whether he was just trying to be a troll for attention or whether he's really convinced he knows better than Meredith Whittaker (because of his gender, I assume). The idea that it can't be a non-profit because... developers have to get paid? really sealed the deal.
@aud @trochee @dangillmor @darwinwoodka right, like, speaking as an executive of a much smaller non-profit, sure, non-profits exist within capitalism and at the pleasure of the state, but the transparency rules are still a meaningful protection. Signal seeks grassroots funding (we couldn't immediately find a report on how much of their funding is individual small donors), which is also important.
@ireneista @trochee @dangillmor @darwinwoodka it's a well known FACT that NO ONE WAS PAID ANY MONEY EVER before Adam Smith wrote "The Wealth Of Nations"

and of course, no one did
anything at all before currency was invented. history started in 1770. I have a keen insight.
@ireneista @trochee @dangillmor @darwinwoodka OKAY BUT

what asinine thing did he say that led him to his, "so yes, I do think I have a better insight than the president of the company"

the people, they yearn for the truth
@aud @trochee @dangillmor @darwinwoodka @ireneista I'm guessing it was something along the lines of LLMs being progress and you can't stop progress.
@foolishowl @trochee @dangillmor @darwinwoodka @ireneista lmao that would be hilarious, "computers that are wrong are the future of chat!!!!!" ?
@aud @trochee @dangillmor @darwinwoodka @ireneista It would be funnier if I didn't keep seeing people say it.

@aud @trochee @dangillmor @darwinwoodka @ireneista He wants to be the guy that "told you so" (his own words).

If you look at his history he has some pretty strong convictions, among them: LLMs are overhyped, and Trump is bad (so far so good).

But recently he's been involved in protests and stuff and seems to have been led into some cognitive dissonance about Signal. Organizers are using it, and he can't square that with it being a mistake to use it for national security.

@aud @trochee @dangillmor @darwinwoodka @ireneista So instead if asking questions and maybe learning something, he made a wild prediction thinking it was a chance to be a hero.