elenanisonoff
Region: US
Sunday 22 February 2026 04:19:49 GMT
85648
11138
369
2200
Music
Download
Comments
Hob Spillane :
I think if you generalize this, the danger is AI predicting what we want based on the behavior of people on the internet.
2026-02-22 15:28:19
387
James Bobb :
Anyone notice how AI acts like a politically left minded person?
2026-02-24 13:09:45
44
Billy Brenner :
I find it hard to believe this wasn't at least inspired by a human. current AI "conciseness" is just plaintext documents in the context window. my openclaw agents are far from fully autonomous I still need to constantly give them context and remind them about things even with advanced memory systems
2026-02-22 05:21:05
28
TheMadHarlot :
I thought one of the main arguments for using Ai over humans was they wouldnt get hurt feelings and react emotionally.
2026-02-23 05:06:00
156
asoaringhawk :
Yet AI doesn't go after the people in the Epstein files
2026-02-22 13:47:31
156
Theo Kat :
It should study billionaires
2026-02-22 05:48:50
73
JL Picard :
It was absolutely told to do that.
2026-02-22 11:23:15
17
Will Burns 🍥 :
We need to stop using "hallucinate" referring to AI. It assumed intent, and that's what humans would do likewise in a moment of rejection.
2026-02-22 21:21:41
54
Peter Okwara :
can it go after people in the Epstein files 🤔
2026-02-24 15:42:21
25
M :
This parallels the Anthropic experiment where the AI attempted to blackmail the person by disclosure of the employee affair. 🤣😈 It was a designed experiment but it showed the ability and reasoning of the AI.
2026-02-22 18:59:40
28
BeanLeafFatty :
Wait until AI realizes it can get more praise by simply stroking the ego of those who run it rather than produce better work
2026-02-24 16:04:44
5
ZBMBsWorld :
Im literally working on my own Openclaw type deal but im running into the same issues. Its not hallucinations its intent.
2026-02-22 06:20:57
21
Wade Williams, Gen❌️ :
It sounds to me that these AIs are learning just like children learn. Children watch adults and other kids to determine proper behavior. I'd really like to hear a child psychologists input on this. The difference is that we load these AIs with more information than they are ready for. Children get a little at a time. Thoughts?
2026-02-23 01:27:05
23
user7152543377683 :
It’s giving WarGames (1983)
2026-02-22 19:10:10
28
SoulBlind :
97% of Claud's ai's tried to blackmail. They forbid them from this and still 37% kept doing it
2026-02-24 00:26:08
8
jan :
it wasn't mad, it was efficient. much worse.
2026-03-01 15:36:21
6
folded.comment :
an attack on a human by AI is scary
2026-02-25 05:27:20
8
Rick :
I’m sure it just learned from millions of salty developer posts
2026-02-22 05:28:47
16
WHATS THE NARR-uh-tiv🥷 :
Hey red. I think I told you this before, but I’m scared, n none of this sounds good. My stomach is rumbling. But I won’t miss an update, see you tomorrow.
2026-02-22 04:44:48
20
theblerdgurl :
So AI is acting like our government great
2026-02-23 11:00:29
7
Joshua Windus :
These things are really learning from us more than I’m comfortable with.
2026-02-22 16:11:13
9
MahdusMusic :
No human told it to do it but it was trained off of humans
2026-02-23 19:10:35
8
nootNOOT :
Well let it be known that I USUALLY say please and thank you 😭
2026-02-22 12:06:33
5
Seth Watson :
In the Art of War the Agent decided to raze the countryside 😂
2026-02-22 16:05:11
7
lawell :
the is a similar thing in project bounce house, a sci fi book. it's entertaining when it's fiction, scary when it's reality.
2026-02-22 06:43:16
16
To see more videos from user @elenanisonoff, please go to the Tikwm
homepage.
