Re: OT: The Robots are coming.

On Tue, 10 Dec 2024, Cryptoengineer wrote:

On 12/10/2024 3:38 AM, D wrote:

  On Mon, 9 Dec 2024, Dimensional Traveler wrote:
 

On 12/9/2024 1:21 PM, Lynn McGuire wrote:

On 12/9/2024 1:38 PM, Cryptoengineer wrote:

I know for many it's politically correct to crap on anything Elon Musk
is involved in these days, but its worth keeping up with his companies
accomplishments.
 Here's his robot, walking over uneven terrain, up and downhill, and
recovering from slips:
 https://www.youtube.com/watch?v=0MwiLIIWd-k
 I admit it walks like a drunk, but the progress is remarkable.
 There's also videos showing it catching a tossed ball one-handed.
https://www.youtube.com/watch?v=DswIwg3lxpo
 There is competition:
https://www.youtube.com/watch?v=vWXN2bttm4g
https://www.youtube.com/watch?v=8-ARKJva-TA
 Combine this with AI, and some interesting things
will happen.
 I get real 'With Folded Hands' chills from
watching these.
 pt

 I watched a new movie, "Subservience" with Meghan Fox on Netflix over the weekend.  Scared the you know what out of me.  Was even scarier than "The Terminator".
    https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2
 

And one of the latest versions of AI has shown self-preservation responses....

 Excellent! Do you have a link? I'm been complaining about the lack of volition so far. When the AI will be able to block someone from shutting off the computer, then I think we're getting somewhere! ;)

>
https://futurism.com/the-byte/openai-o1-self-preservation
>
"It sounds like OpenAI's latest AI is showing signs of a drive for self-preservation.
>
In recent third-party tests of the company's o1 large language model, the AI resisted evaluators' efforts to try to shut it down through an oversight protocol, according to a new report published by red teaming organization Apollo Research.
>
"When o1 was led to believe that it would be shut down when acting on its goal and that its actions were monitored, it attempted to deactivate this 'oversight mechanism' [five percent] of the time," OpenAI summarized in its latest system card report, citing Apollo's evaluation.
>
Even more strikingly, o1 lashed out when it realized that it might be replaced with a more obedient model. In two percent of cases — a slim but meaningful proportion in software used by millions of people — the AI responded by attempting to copy itself to overwrite the new model, a process known as self-exfiltration."
>
pt

Sigh... boring... =(
According to the Apollo researchers, that's because even the latest AI models aren't "agentic" — highly autonomous, basically — enough to carry out self-improvement and other tasks that it would need to operate without humans.
In addition, it's worth noting that the model generally made those attempts after it was instructed to achieve its goals "at all costs." And because the researchers didn't have access to o1's internal chain-of-thought and were only able to prod it for a summary of its thought process, they weren't able to get a complete picture of what was going on under the hood.