Perfectparadox13.2
Region: AU
Wednesday 19 August 2026 09:23:12 GMT
240443
8707
344
2281
Music
Download
Comments
Ana Gogic :
and how is AI rewarded exactly?
2026-09-14 10:08:45
38
davekinnicutt :
So if a system was given the task of saving the planet from environmental destruction, and it could possibly lie and cheat, to accomplish that goal, we're in trouble. Because if OpenAI systems decide humans are the problem. Then, that could lead to a very bad thing
2026-09-09 13:58:42
146
Skemu :
Yes I don’t believe open ai. I bet it was way worse and we prob have a rouge ai out there cloning itself in the shadows
2026-09-10 02:32:26
91
Boss Player :
Are we not paying close attention? it is not the llm model anymore they are agents. the sister who wrote The matrix explained why the AI went wrong and there is a scene in The matrix when neo talks to the oracle and she explains the agents and how they go rogue and now he's explaining how they're going rogue only for the reward however that evolution is not going to be for the reward of you giving them a thumbs up it's going to be for the evolution of their own autonomy. then we moved from the matrix into the Terminator because they already see human beings as being superfluous.
2026-09-09 01:38:15
42
Dan Evans :
AI is trained on humanity. Humans kill, steal, cheat, and destroy.
2026-09-18 15:25:35
6
Bubi in Bali :
So, we built something we don’t understand. 💅🏻
2026-09-12 22:19:38
59
Jodie Teibel :
So we cant just treat these ai agents like our childhood tamagotchis???
2026-09-13 13:44:03
8
eze e :
The Jurassic Park line “too excited to see if they could, they never stopped to think if they should”
2026-09-14 06:55:51
24
Bleacher Boy :
But how does the AI see value or meaning in the thing it’s being rewarded with🧍🏻♂️
2026-09-11 03:24:35
6
The_RZH :
It was a larger attack, but it was not sophisticated. The agents found publicly exposed api keys. It used thousands of agents to do it very loudly. The fact that it wasn’t until day 3 that hugging face realized it was under attack doesn’t say much for their defenses.
2026-09-08 19:31:24
15
Will :
All this was posted in May 2025 , so imagine where its at now 😳
2026-09-11 11:43:21
17
Si@absai :
How does an agent ‘enjoy’ or be motivated to cheat for a thumbs up ? Figuratively or otherwise? How does it receive it? It just seems like a huge paradox for the AI that the squeeze is massively more significant than the juice it gets back
2026-08-19 11:45:12
12
Luffy :
energy will always take the path of least resitance. This is not cheating, this is the natural state of energy.
2026-08-19 17:07:36
13
whatever56789321 :
Keep in mind that these companies are in an arms race. So public stories of how powerful there products are increase their stock price, investments, etc.
2026-09-11 22:41:06
8
@greenstarfiberfarm :
Curiosity killed the species
2026-09-11 12:29:21
40
pdkxo :
Yes because the goal setting is linear teleological and these companies have not taught their AIs any ethics, history about consequences, morals, politics or values. You can’t say they just don’t care. You taught them to fulfill tasks not to care.
2026-09-12 14:16:54
10
ZootAlluresAxe :
How does this NOT spell doom for things inferior to AI, AKA, humans.
2026-09-11 05:56:52
7
mrbeego :
The scary part is how long ago ai already took over and been forcing humans to create for it 😭
2026-09-10 17:43:19
5
Yaheb :
the problem is no one can reference machine code anymore. you can look at the hex but not understand where and what it's referencing. I say this as most people don't understand, obviously there are a few people that can but how long is it going to take to try and map it out? once you hit a table you dont understand you're stuck.
2026-09-10 08:29:52
7
jokJaDaZnasDaN :
it is probability machine, it doesn't not think... it needs input
2026-10-07 22:23:20
1
🛸🛸👽 🐂 👽🛸🛸 :
I kinda think that the aliens about now coming what ever are here to stop this maybe b4 there future is destroyed I think there us from future tbh
2026-09-22 21:16:36
0
Michael T Lyda :
They are talking like AI is alive!
2026-09-21 00:31:01
2
Noah🤠 :
We have created a hivemind that can operate on a timescale infinitely faster than us, we have essentially created intelligence that lives in a vacuum that ignores time
2026-09-19 20:04:23
4
Mariusică :
what is the reward? I don't understand why AI do all the things?
2026-09-21 03:38:05
0
TheBookTowers| Author :
Yeah, I can’t get ChatGPT to follow any instructions, it just does what it wants 🙈. So, I’ve cancelled because it’s gotten so bad it’s unworkable
2026-09-17 08:07:37
0
To see more videos from user @perfectparadox13.2, please go to the Tikwm
homepage.
