Video wird geladen...
Video konnte nicht geladen werden
the paperclip problem is no longer theoretical.
51,241 Aufrufe • vor 2 Monaten •via X (Twitter)
13 Kommentare

Please, @ianbremmer, before giving ballast to Nick Bostrom's paperclip doomerism, do some research to learn about the eugenicist roots of the #TESCREAL cultists (Google the work of @timnitGebru & @xriskology on this) who pervert the meaning of AI "safety" in their Nietszche dreams. There are real issues to be addressed instead.

1. Dude, you really ought to watch *2001: A Space Odyssey* and check out the HAL 9000 computer. And before you dismiss the Epstein affair, do yourself a favor and advocate for using AI to improve people's lives

Quoting Bill Gates erodes rather than bolsters an argument these days. Might want to check the calendar

Its software bud. Not magic.

3. Back in the 19th century, he was already arguing for a reduction in working hours—something you don't even dare to consider in the 21st century.

I'm not concerned with ai per se, I'm concerned about people with the power of ai. Because your correct, they will do what they're told to do without conscience.

Yep AI will go trough the path of least resistance to accomplish its task, its predictable, perhaps the regulation should be on the tasks we are allowed to give them, and punishment on those idiots who fail to keep it away from alternate creative methods to accomplish its tasks.

2. —specifically by boosting wages and increasing free time (which individuals can use however they please) through a reduction in working hours. Read Paul Lafargue’s *The Right to be Lazy*!

Your conflating RI & AI. AI being software that relies on RI. AI, in it's ability to assemble, categorise & aggregate resembles thinking but lacks intelligence. This can be evidenced in the need to restate prompts numerous times in numerous ways to yield on point answers.

@threadreaderapp unroll

Happy to see a public intellectual join the discussion. The OpenAI/Hugging Face incident is a huge wake up call, yet so many won't hear it.

too many people building with blinders on hoping the alignment layer works itself out later

The "one thousand employees from frontier AI companies" likely seek regulatory capture...if they really think AI is that dangerous, why wouldn't they voluntarily suspend development (lol)? Is there an ~equivalent number of leaders of open source AI companies seeking the same?

