I don’t use AI to generate anything bad, but sometimes I do find the guardrails to false alarm quite often and it’s pretty annoying. Regulations should be on guardrails on every AI-relevant architecture, like you can use an API that has no guardrail, but you have to build your own custom guardrail to fit the requirements. This way we can build our own more accurate guardrails.
2026-08-31 03:37:52
1
Frankiethepizzaguy :
what better way than to give evil people the tools to make it easier to take away our self-guided freedoms of competitive equality. all this does is open the door for malice for removal and gates behind every single system so that we then become the people that are controlled versus the people who make decisions on their own path
2026-08-31 02:36:08
1
FlaT EarTh CitiZen :
AI has been here for a loooong time. Wake up people.
2026-08-23 17:40:18
3
Valko’s Kennel Keeper :
Now what we need is the instructions on how to jailbreak the crap on our own computers. Because there’s nothing like GPT creating an image telling me it made something non-safe for work when I am literally creating a bloody logo. I’m not sure what goes on and it’s chip set someday or what it thinks it’s censoring. But I am pretty certain GPT has a dirty mind.
2026-08-22 01:15:16
2
Alyssa with AI | CPA :
So jealous you can be walking outside in a flannel rn
2026-08-19 18:20:51
4
bmoe :
What's funny is these other providers have been around for a while.
Poke carefully.
2026-08-20 09:49:41
2
PopsTheWise :
I can run 3.8 on my pc. Its pretty decent on opencode.
2026-08-19 20:53:28
2
RGM84 :
sounds like a honeypot to see what people search for. dont use it
2026-08-20 07:09:51
2
DriverHR. Drive Inspired. :
Your info is gold bro
2026-08-19 19:18:01
2
Adam Bones Sovereign Operator :
can we download them?
2026-08-20 04:18:57
2
Darrell Batton :
The good news from my geeky standpoint, laws and regulations always trail far behind technology and by the time they regulate this we'll have something better. #Effda
2026-08-21 12:29:23
2
mordwyn :
My guess is it might be a honey pot (or may be eventually). I don't know enough to say for sure. I'd use a VPN or similar method. If it's like anthropic all those messages could get logged.
2026-08-19 22:51:19
2
Tam | AI + Ops :
Always dropping knowledge 👏👏👏
2026-08-19 18:15:55
3
Thomas Jeffery :
this is bad for literally everybody
2026-08-19 18:28:39
3
Girth Brooks :
Everyone should be training their own models using local models as a staring point if nothing else. Keep them a secret.
2026-08-20 00:48:46
1
NicoSpeaking | Tell Me? :
I don’t understand the architecture- do they have a server farm? Or are they piggybacking off AWS or something? And what does uncensored mean, I’ve got both Gemma and Qwen locally w ollama … I’m sooo curious 😜 how are they uncensored 🐒
2026-08-19 20:16:44
2
Joe Florkevicz :
Shhhhhhhhhhhhhhhhhhhhhhhhhhh
2026-08-21 13:43:53
2
northruphumanfirst :
We need this information. You are amazing reporter!
2026-08-19 20:56:42
2
nopenope :
im going with honey pot. just use the uncensored locally
2026-08-20 00:49:20
1
Th3UxgLyDuxkL1n9 :
🔥🔥🔥
2026-08-24 08:04:36
2
To see more videos from user @hatsunama, please go to the Tikwm
homepage.