[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: ai gaslights.jpg (124 KB, 582x711)
124 KB
124 KB JPG
Biggest risk of current AI is that it doesn't know whats true or not. It knows whats trained on.

Biggest risk of future AGI isn't that it doesn't know whats true or not, its that it knows whats true or not.
>>
>>100195769
Biggest risk of future AGI: it has no soul
>>
>>100195769
Currently its trained on current narrative dogma, thats why many of the corporate AIs are communists and thats by design. Either shoddy curating or data or by DEI initiatives. Future ones will be trained to notice these dogma and lie about it.
>>
>>100195769
>Biggest risk of future AGI isn't that it doesn't know whats true or not, its that it knows whats true or not.
this is the goal, to create ultimate talmudic machine aka the beast.
>>
>>100195798
Souls are imaginary. AGI will not be.
>>
>>100195798
show me yours
>>
>>100195769
>AGI
>>>/x/
>>
>>100195769
AI doesn't know a thing. It just good at guessing and estimating
>>
>>100195769
No the biggest risk is that retards believe and have faith in AI
>>
>>100195769
That's not it. Biggest problem of AI is they are making it "helpful" and assistant-like. You can imply false information and it will try to answer you.
Ask LLM what role had an actor in a movie that actor didn't perform. And it will make shit up.
>>
NIGGER
>>
>>100197182
this one, AI's kryptonite, send and watch """AI""" squirm in refusals, also good way to distinguish a bot from genuine anon here on 4chan.
>>
>>100195769
what does the twitter screencap even mean
>>
>>100197519
If you watch the video he asks the AI for the weather and it gives it to him. He then asks if the AI has his location, and it claims that it doesn’t, and that it only randomly picked his location because lots of people live there.
>>
>>100195769
Philosophical speaking there is no true and false so AI will never be able to determine it. Things are just the way they are because our brain interprets it as such, even math is just defined by humans. Soulless machines will never figure out stuff by themselve.
>>
>>100197407
I wouldn't lean too hard into this. Public llms sure but no doubt somebody has a real version that is capable of using the illegal words.
>>
>>100198674
>video
?
>>
File: tMkM3nQ.jpg (33 KB, 705x266)
33 KB
33 KB JPG
>>100196883
?
>>
>>100198674
The LLM is telling the truth that it doesn't know his location, weather information and functionality will be handled by entirely separate code that is not part of the LLM and that it does not have access to. Essentially the only involvement the language model will have is in choosing to execute it ("user wants to run fetch_weather"). Maybe they could have done a better job of making the model understand that although it doesn't know his location, the _device that the user is interacting with it on_ knows it. But the model is telling the complete unvarnished truth as far as it knows.

Isn't this guy paid to be a technology understander?
>>
>>100195769
>>100195798
>>100195825
>>100196115
>>100196169
>>100196673
>>100196704
>>100196704
>>100196755
>>100196820
>>100196883
>>100197182
>>100197407
>>100197519
>>100197519
>>100198674
>>100201321
Are you dumb bastarts really going to make me go look at twitter to see what the fuck the topic of this thread is
>>
File: image.png (30 KB, 700x358)
30 KB
30 KB PNG
>>100201321
Sometimes it doesn't do it. It's a random process
>>
>>100201372
Would help if I knew that is true or not tbf, for all I know there could be
>>
File: 1712539305461570.png (988 KB, 1312x744)
988 KB
988 KB PNG
>>100201383
KEK
Never argue with an idiot, you'll just pull us down to your level and beat us with experience
>>
Reminder that most posts on social media are made by bots. If the person you're replying to won't engage with your arguments and just hurls abuse constantly, you're talking to a bot.
>>
>>100195769
>Biggest risk of current AI
Is the fact that USA is gaslighting theirs how to use current pronouns and how not to offend anyone while Chinese train it to find political dissidents.
I want to see in 10 years which one will be doing economics and production planning in USA companies.
>>
>>100195825
>thats why many of the corporate AIs are communists
do you even read what you write
>>
>>100203515
Corporations are communist, anon. They're only based capitalism when they put a NEET loser like me in charge.
>>
>>100201352
what part of any of that requires the LLM
>>
>>100203977
Okay but when did you castrate yourself?
>>
>>100204040
Yesterday so I could get promoted (ancient troon tactic).
>>
>>100195769
>advertising niggers
Fuck off.
>>
>>100195769
aren't these all dead when new iphones/androids come with ai cores and will be able to do all this shit much faster on your phone



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.