Open AI, Q*, Qualia and AGI rumours

[Mon0Zer0]

Rumours coming from the Open AI CEO firing (and subsequent rehiring) fallout continue to circulate concerning the development of Q*, which some, including Elon Musk speculate is a *better* way of decoding or solving math problems, including ones it had never seen before (been trained on). One frightening application is its potential ability to quickly solve 256 bit hashes used in cybersecurity using maths beyond the skills of human beings.

One other application is its ability to create a flexibly structured "brain" network to solve other issues, which could directly lead to Artificial General Intelligence.

David Shapiro goes into the rumours and their possible impact here:

One other fallout, not covered, are the potential regulations that companies like Open AI are drawing up which they claim may be necessary to protect "against the risks from frontier AI models which others claim are exceptionally totalitarian (I've only briefly scanned it, so can't comment on it personally).

 

[Brian G Turner]

It's funny how this has been labelled "AI superintelligence". You'd think they'd reserve the term "AI intelligence" for something actually resembling intelligence from an AI. ChatGPT so far basically seems more like a copy/paste productivity tool than anything else - a super efficient way to summarize data more than anything else.