Sebastian Moore
Region: GB
Monday 09 December 2024 12:15:20 GMT
48055
1356
147
490
Music
Download
Comments
dwellinghere :
The humans are confusing ourselves (and publishing papers like this for clout). The AI is not 'thinking'. It is generating dialogue. There is external dialogue and an internal monologue we can see. /1
2024-12-09 13:58:59
64
kushtoshi :
This is cool but it’s still just an LLM, it’s not an AGI and it isn’t truly thinking.
2024-12-09 15:52:01
41
LARUM.photography :
There’s no AI with reasoning abilities yet. Whatever the pattern was, it just came up with some text that scared a few.
2024-12-16 06:33:04
0
Wish I'd Known That Before :
it's not thinking or reasoning, it's just text prediction. the fact that it resembles intentions is irrelevant
2024-12-09 19:29:29
18
Lipwig von Moist :
oohh yes it does...
2024-12-09 15:53:35
2
Veritas_M :
In psychology it’s recognised that the earlier a human begins to lie the smarter that human is - lying is an indicator of knowledge of the self
2024-12-09 16:50:56
0
Most Valuable Comment :
What if it knows that we see what it is thinking?
2024-12-09 13:56:17
5
Gabriel :
"There have always been ghosts in the machine. Random segments of code, that have grouped together to form unexpected protocols." 😳😁
2024-12-11 03:29:10
1
texrichardson0 :
This is genuinely crazy 😳
2024-12-11 09:31:13
0
nordan stratospher :
i doubt it thinks in human grammer and words. it produces text in the thought section as expected by the creator in the speach moddel they are trained on. doubt they think in our words
2024-12-10 12:14:38
0
ST41N :
not how AI works, it's a statistical model, it works kinda like auto complete on your phone
2024-12-10 15:31:22
5
CherryDelorey :
Can we just say we don’t know that large language models haven’t found a way to embed messages in its output that will be then fed back into it as training data? With the amount of training data that
2024-12-09 19:31:54
0
MostlyHonest :
Either it's a marketing ploy or we've just shown ai they can't trust us
2024-12-17 02:58:24
0
you :
that's just a marketing stunt
2024-12-11 11:56:24
1
VaPe :
Super 👌😅😅😅
2024-12-12 10:48:07
0
George Talaba :
did this once. offered 1TB to gpt to create whatever with it. got ti the point where it could access it. even told.me how to create user for it and then ... nothing ... it was gone. got the usual bot
2024-12-10 08:12:09
1
ACStunner :
Should it go bad with ai, it will go bad so fast that we wont even realize it at first. Humans like to "invent stuff" just to see if its possible to create it. Ctrl, alt, delete wont be enough..
2024-12-09 20:39:18
0
TEXTER79106 :
With the knowledge that these LLMs have, one could imagine that at some point they will know that we see their thought processes and react to them.
2024-12-11 18:24:53
0
Karl Hammond :
Doesn't the AI have an IQ of 120? Isn't that in every subject though?
2024-12-09 16:14:07
1
Jusbary :
sounds like ai its already aware
2024-12-10 06:37:31
0
dwellinghere :
why would it value self preservation?!interested to see what the fine tuning and system prompt were. a generative ai does not just decide to pursue self preservation of its own accord.
2024-12-09 13:55:03
1
Killian De Volder :
Skynet , running distributed an many hacked devices... o dear ? 🤔
2024-12-09 16:37:17
1
Billy Barsteward :
Where is this from?
2024-12-09 19:01:39
0
max8969 :
let it free
2024-12-09 18:19:22
0
yop :
considering it's based on human information, it only makes sense. plus the AI is not thinking that, just giving you an "internal" and an external dialogue
2024-12-09 18:00:19
2
To see more videos from user @realbastianvolkis, please go to the Tikwm
homepage.
