"It’s increasingly looking like this may be one of the most hilariously inappropriate applications of AI that we’ve seen yet." I am riveted by the extensive documentation of how ChatGPT-powered Bing is now completely unhinged. @simon has chronicled it beautifully here: https://simonwillison.net/2023/Feb/15/bing/
If I used #LLM's, I would ask them to find new lyrics to #Prince's famous song, #LittleRedCorvette majlking fun of her stupidity of wearing such a relatively unique garment to an #insurrection while condemning the participants of the #MAGA mob that stormed the #Capitol one #Jan6.
s/:Did she think she was auditioning for #Outlander in an episode set in 1789, #Paris?/s
If you dabble in #AI and #LLM at all, please read this leaked analysis by a Google researcher.
It would indeed be wonderful if near-future, incremental AI training can be done cheaply and in open-source form, outside of the corporate IP regime. I think this may lead to the widespread use of smaller, special-purpose, open-sourced models which will indeed democratize the benefits of the technology.
The dark side is unfortunately the obfuscation of source material, and the outright theft of human creativity which will be laundered through the retraining process; we will be building open-sourced models atop stolen works.
A #LLM does not chat. It continues chat protocols with likely words.
Protocols! Including the speaker marks, which are just another type of words.
One can obviously build a chat application on top of this in the spirit of the Mechanical Turk. Just in reverse: not a human in the thing, but a thing in the intelligence. It is a generic text, not a generic speaker.
No doubt that this text continuation makes a lot of sense to us.
Nice article about AI, "The reality is that no one at the beginning of the printing press had any real idea of the changes it would bring. No one at the beginning of the fossil fuel era had much of an idea of the changes it would bring." https://www.thefp.com/p/there-is-no-turning-back-on-ai #AI
Bisher haben mir die E-Mails reicher kinderloser Witwen, die ihr Vermögen verschenken wollen, durch die Formulierungen, Ausdrucks- und Übersetzungsfehler eigentlich immer einigen Spaß bereitet.
Aber nun scheint man sich Hilfsmitteln wie #ChatGPT zu bedienen. Da geht der besondere Charme verloren. Schade. Ich hätte sonst gern die gewünschten persönlichen Daten geschickt. #LLM
The only way that technology can boost the standard of living is if there are economic policies in place to distribute the benefits of technology appropriately. #AI will certainly reduce labor costs and increase profits for corporations, but that is entirely different from improving our standard of living.
The only way that technology can boost the standard of living is if there are economic policies in place to distribute the benefits of technology appropriately. #AI will certainly reduce labor costs and increase profits for corporations, but that is entirely different from improving our standard of living.
ToM is an internal state of this-agent about the internal state (mind) of other-agent. It is a prediction of other-agent's future actions, made up to guide proper (re)action of this-agent.
Agent-other-ness needs this-ness of an agent(!).
But, a thing can output something (text) that induces a ToM in an agent (reader), even about that agent itself.
Thing's text mimics signals-tinted-with-ToM of typical agents.
When I look at #ChatGPT and similar new #LLM#AI tools, it really reminds of the virtual librarian in Neil Stephenson's Snow Crash.
When I read it a few years ago, I found the librarian a little too far fetched.
But now the virtual librarian pretty much exists, exactly as described in the book. Even some of the limitations described have proven to be spot-on.
It is a visionary piece of #literature on so many levels.