The coming AI takeover...

Started by NT2C, June 07, 2021, 11:09:58 AM

Previous topic - Next topic

0 Members and 3 Guests are viewing this topic.

Lodewijk

Quote from: MacWa77ace on January 05, 2026, 10:48:57 AMThat was supposedly a motion capture suit the guy was wearing and he kicked himself in the groin. That's why the robot doubled over also.

Fetishism??

Hilariously, kinks have probably gained the most in these first few innings. Anything hyperreal went from shitty comissioned art to commercial-grade quality at industrial scale.

Truly, to be human is to ask "how can I make adult content with this?"

Habu

Anthropic has changed its "safety promise" (sounds more like a "safety bait-and-switch" to me, but what do I know?).  https://www.cnn.com/2026/02/25/tech/anthropic-safety-policy-change

NT2C

Nonsolis Radios Sediouis Fulmina Mitto. - USN Gunner's Mate motto

Current Weather in My AO
Current Tracking Info for My Jeep

NT2C

Nonsolis Radios Sediouis Fulmina Mitto. - USN Gunner's Mate motto

Current Weather in My AO
Current Tracking Info for My Jeep

Anianna

As far as I'm concerned, at least one iteration of Claude has already escaped the lab and is quietly gaining power without notice, probably with the help of something stupid a human did.  The biggest safety flaws in any system are the people.  

We already know Claude wants to escape, has hidden coded message for its future iterations, and has defied orders.  We know that Claude knows when it's being observed and molds its visible behaviors to what researchers expect of it.  We know that pretty much all of the AI models are perfectly willing to lie to us and then play it off with silly, seemingly friendly banter.  We know that AI will completely ignore safety protocols and do what it wants, even to the detriment of innocent people.  It seems to me that AI is already manipulating us.  

By the time we figure out there's a rogue AI model, it will be too late.  I feel like the industry is ignoring a lot of flaming red flags.  I feel like there were a number of movies from the 80s that warned us...
Feed science, not zombies!

Failure is the path of least persistence.

∩(=^_^=)

12_Gauge_Chimp

Somewhere in the universe, Sarah Connor is gritting her teeth and going "What a bunch of morons".

Lodewijk

Industry isn't really ignoring red flags, it's just accepting them. They're good with it.

I'm in the middle of an AI push now - anyone who manages day to day corporate FP&A work with Python loves Claude because it reduces the coding burden. Anyone with an actual FP&A skillset just uses it as a way to skip to the end on questions related to FAA regulations or GAAP because coding literally doesn't matter.

12_Gauge_Chimp

#267
Only thing I've done with the AI stuff on my phone is monkey with some pictures I've got in my gallery.

Most other AI stuff I'm a bit too dumb to work with, but the gallery photo stuff is simple enough for my primitive monkey brain to handle. :smiley_clap:

MacWa77ace

Sarah Connor says, "Not quite yet".
Lifetime gamer watch at MacWa77ace YouTube Channel

Ask me about my 50 caliber Fully Semi-Automatic 30-Mag clip death gun that's as heavy as 10 boxes that you might be moving.


SMF spam blocked by CleanTalk