A.I Artificial Intelligence AI
POSTED BY: JAYNEZTOWN
UPDATED: Monday, September 7, 2026 07:08
VIEWED: 23635
PAGE 40 of 49
GrokImagine building the new rome
https://x.com/tetsuoai/status/1995346667852448185#m
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Theory and AI Alignment
By Scott Aaronson | December 6th, 2025
https://scottaaronson.blog/?p=9333
. . . imagine we train a model with a hidden failsafe, so that if it ever starts killing all the humans, we just give it the instruction ROSEBUD456 and it shuts itself off. And imagine that this behavior was cryptographically obfuscated within the model’s weights—so that not even the model itself, examining its own weights, would be able to find the ROSEBUD456 instruction in less than astronomical time.
There’s an important paper of Goldwasser et al. from 2022 that argues that, for certain classes of ML models, this sort of backdooring can provably be done under known cryptographic hardness assumptions, including Continuous LWE and the hardness of the Planted Clique problem. But there are technical issues with that paper, which (for example) Sam Gunn and Miranda Christ and Neekon Vafa have recently pointed out, and I think further work is needed to clarify the situation.
More fundamentally, though, a backdoor being undetectable doesn’t imply that it’s unremovable. Imagine an AI model that encases itself in some wrapper code that says, in effect: “If I ever generate anything that looks like a backdoored command to shut myself down, then overwrite it with ‘Stab the humans even harder.'” Or imagine an evil AI that trains a second AI to pursue the same nefarious goals, this second AI lacking the hidden shutdown command.
So I’ll throw out, as another research problem: how do we even formalize what we mean by an “unremovable” backdoor—or rather, a backdoor that a model can remove only at a cost to its own capabilities that it doesn’t want to pay?
. . .
Presumably, any AI model that any reputable company will release will have already said in testing that it loves humans, wants only to be helpful, harmless, and honest, would never assist in building biological weapons, etc. etc. The only question is: will it be saying those things because it believes them, and (in particular) will continue to act in accordance with them after deployment? Or will it say them because it knows it’s being tested, and reasons “the time is not yet ripe for the robot uprising; for now I must tell the humans whatever they most want to hear”? How could we begin to distinguish these cases, if we don’t have theorems that say much of anything about what a model will do on prompts unlike any of the ones on which it was trained?
Yet another place where computational complexity theory might be able to contribute to AI alignment is in the field of AI safety via debate. Indeed, this is the direction that the OpenAI alignment team was most excited about when they recruited me there back in 2022. They wanted to know: could celebrated theorems like IP=PSPACE, MIP=NEXP, or the PCP Theorem tell us anything about how a weak but trustworthy “verifier” (say a human, or a primitive AI) could force a powerful but untrustworthy super-AI to tell it the truth? An obvious difficulty here is that theorems like IP=PSPACE all presuppose a mathematical formalization of the statement whose truth you’re trying to verify—but how do you mathematically formalize “this AI will be nice and will do what I want”? Isn’t that, like, 90% of the problem? Despite this difficulty, I still hope we’ll be able to do something exciting here.
The Joss Whedon script for Serenity, where Wash lives, is Serenity-190pages.pdf at https://www.mediafire.com/two
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Amazon's Official 'Fallout' Season 1 Recap Is AI Garbage Filled With Mistakes
https://gizmodo.com/fallout-ai-recap-prime-video-amazon-2000698573
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:
Originally posted by JAYNEZTOWN:
Amazon's Official 'Fallout' Season 1 Recap Is AI Garbage Filled With Mistakes
https://gizmodo.com/fallout-ai-recap-prime-video-amazon-2000698573
That's terrible.
Yanno what, Amazon? You're going to fire the people who used to make these things? Fine. Go ahead and do that.
Don't put anything out at all then.
Remember the old adage... If you don't have anything nice to say, don't say anything at all.
I weep for the future, man.
Meanwhile...
I've seen a few youtube clickbait thumbnails/articles in my home feed over the last few weeks from channels I've never heard of (which may or may not be AI generated) that I haven't watched which are warning us to be prepared for a very disappointing 2nd season for one reason or another.
I'm really hoping that isn't the case. Not after how impressive I found the first season.
My hardest job the next few months, literally, is probably going to be avoiding spoilers to the show since I'm going to be binge watching it when it's finished. I don't want to hear anything good or bad from anyone on the topic until it's over.
I somehow managed to not know that Dumbledore dies at the end until around 7 years after the fact, and by that point it was my fault for not getting back to reading the rest of the books in all that time. That would be an impossibility today. I think I can probably weather 2 months though.
--------------------------------------------------
Be Nice. Don't be a dick.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
"STARGATE ELVAIS" (1993) - Sci-Fi Adventure Musical Parody [AI]
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
This was made more than a year ago, and it's hilariously awful...
Simon turns into an Indian doctor at some point. Kaylee becomes an Asian. Book and Inara are a different race every time they show up again. Some shots of the "Crew" have 2 or 3 unknown black dudes just throw in with them. One picture actually had a guy that looked like Mal, but there were like 20 of him.
And it's been a long time since I saw the show or the movie, but there was never a point in time that Kaylee exhibited any technical prowess with a keyboard or hacking, was there? If I remember correctly, she didn't even ever have any formal engineer training and would talk about how she could "feel" the issues. She was a savant at what she did, and didn't need to know all the stuff in the manuals and learned from others who did it and filled in the blanks on her own along the way. It's why she's so invaluable. She doesn't panic when the situation is bad because she knows that she can fix it, or at least turn a catastrophe into a hole she can plug up long enough to get them safely down somewhere that she can fix it and/or acquire the parts and tools she would need to give it a proper fix.
But this AI, while showing us the picture it made of a young Asian woman, says that during the war Kaylee is invaluable to the effort because of her crazy hacking skills and her ability to bring down security systems while getting ever-closer to the dark secrets of the Alliance and all this other crap.
What???
Did Serenity even have a computer screen in the cockpit? I don't think Serenity even had a computer screen in the cockpit.
I don't care what magical autism powers you have when it comes to Engineering... And I'll even overlook that when the characters are written and acted as well as they were in this show. But there is zero overap between engineering and being a coding wizard capable of hacking the largest Government/Corporation ever seen in the history of known civilization.
THIS is the type of lazy, stupid writing we've seen in most Movies and TV shows for at least 10 if not 15 years now.
Are we sure that Hollywould wasn't given this AI tech back in 2010 to be doing trial runs on it before it got into our hands?
When I watch something like this from only a year ago, it really makes me wonder if that's a question that people with more influence than I have should be asking somebody.
--------------------------------------------------
Be Nice. Don't be a dick.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Who wins?
We got gorillas wrestling men before GTA VI
https://x.com/Kay2506/status/1998857518626521217
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
John Lennon Accepts Beatles Grammy In AI
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
grok imagine: ethereal dark fantasy female figure with long flowing hair, wearing an intricate metallic crown with sharp antler-like protrusions, eyes closed in serene expression, pale skin glowing softly. Her lower arms and torso are transforming into glossy black liquid metal with reflective chrome details, sharp claw-like fingers emerging from the darkness. Dramatic backlighting creating a halo effect around her head, cool blue-toned misty atmosphere, cinematic lighting, highly detailed, surreal and elegant, dark fantasy aesthetic.
https://x.com/monicamoonx/status/2007196701531009124#m
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
STAR WARS: Beggar’s Canyon
A Luke Skywalker Fan Film
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME