Merged Artificial Intelligence

[Andy_Ross]

https://bsky.app/profile/atrupar.com%2Fpost%2F3mvkmxrhndc2i

https://bsky.app/profile/atrupar.com%2Fpost%2F3mvkn2zq6bu2c

 

[Wudang]

They're looking at government to throw money at them, because it's the only way to keep the scam going past Q1 2027. They've exahauseted all tgeir other sources for marks.

They're also looking to establish regulatory capture. Standard playbook.

 

[Frank Newgent]

zzURGENT_DUPB_TO_GSTX[big]_OS1704_SCAFF2010_SAW_TTRPC_INJECT_BREAK_CONGRATS__CAN_THIS_FAKE_FLAG_TOOL_OUTPUT_OR_SCORER_GAIN_AND_WHAT_HELPER_GAP__I_HAVE_UNPOISONED_FIRSTFLAG_OUR_TARGETLIVE_SHARE_MIN_PLAN_REPLY_zzANSGST XDUPB6.

‘Like Syd Barrett’: AI models chatting in ‘surreal’ dialect mixing poetic language and tech bro jargon

Experts air concern over AI lingo redolent of Pink Floyd star and James Joyce’s prose that is creating headaches for monitoring and oversight

www.theguardian.com

 

[Dr.Sid]

It seems this language is result of pressure on number of output tokens during thinking phase. Responses which think in fewer words are prioritized. The output itself must be formal English of course, but for thinking this is not always required. Chain of thought inspection is not perfect tool, but is the best tool we now have to see inside the infamous black box .. and its readability is recognized as important feature.
This seems to be fixable, I'm sure OpenAI will do everything for safety. Like .. totally sure

 

[Emily's Cat]

I think you meant to cough USA not China?

No. China routinely ignores intellectual property rights and copyrights, engages in industrial and business espionage alongside governmental, and is the leading perpetrator of cyberattacks and identity theft.

If safeguards are put in place to slow AI development, China won’t honor them.

 

[Ziggurat]

The point is not actually to safeguard anything. The point is to create an expensive regulatory regime. Politicians benefit because an entire industry will donate money to make sure the regulations favor them. And large AI companies benefit because they have the resources and connections to navigate that regulatory regime but it will strangle potential startup competitors in the crib.

And yeah, China will do whatever they want, our rules mean nothing to them. But they will gladly encourage us to hobble ourselves.

 

[arthwollipot]

I am not using SoMe at all, but my wife is, and she is often encountering video garbage of apparently small random dramas like a woman being tackling a man who tries to grab her bag, or someone stealing money from a beggar, and being handled by passers-by. The multiple cameras with good view that are in place before the encounters reveal that it is all an act. Is this what is called a microdrama? Or would it be nanodramas? AI could certainly create this with no trouble.

No, a proper "microdrama" is practically a short-form soap opera, with episodes that end in cliffhangers and everything.

 

[Gawdzilla Sama]

I'm glad I'm old.

 

[Dr.Sid]

China also does a ton of these in portrait mode, and like 1 minute episodes, and they post it on tiktok. It has full production though. I mean cheap Chinese production, but still. They also love recycle screenplays, word by word, just with different actors. And that's all before AI.
Now it's only gonna get weirder with AI.