The first 30 minutes of this is an interesting interview with Stephen Wolfram (of Wolfram Alpha fame). TL;DW: He makes some good points. He says that there's two primary types of AI: LLMs and Computational AI. Neither of which think like humans, or will ever be capable of thinking like humans. LLMs, he says, have basically cracked a previously unknown law of grammar and know how to spit out sentences that make sense. Computational AIs are able to sort out things like complex mathematical problems that we don't know how to solve (or can't solve easily). Wolfram basically likens all forms of AI to tools that we use to help us do things, and gives reasons why they'll never fully be able to replace us in every area because there's aspects of being human that just can't be replicated in software. He also thinks that we're headed for a plateau of the technology (if we haven't already hit it).
Google’s Sergey Brin Says Engineers Should Work 60-Hour Weeks in Office to Build AI That Could Replace Them How about he eats a bag of rancid whale dicks instead?
Reportedly, MicroSoft has canceled construction on a large number of planned data center projects. These data centers most likely would have been used for AI models...
Surprise! People don't want AI deciding who gets a kidney transplant and who dies or endures years of misery
So I've been avoiding AI stories because they make me weep for the future of humanity. But I saw this article, and thought it was a brilliant examination but only with what's wrong with AI generated fiction, but also with us as consumers of stories: https://countercraft.substack.com/p/mfa-vs-llm-is-openais-metafiction The best quote, I think, is here: And i think he's right. This is a subject I've been pushing on my students regarding literary interpretation. How do we read a story? And how do we try to understand it?
Ilya Sutskever, ex-OpenAI, gets $2b funding not to release anything until he has ‘super intelligence’
> has super intelligence > still thinks trebuchets and knockoff Batman costumes are the best way to catch a roadrunner Yeah, that sounds like machine learning levels of competence.
Teacher Used AI to Make Pornographic Videos of Students, Authorities Allege as Corinth Families Demand Accountability
https://www.mindprison.cc/p/studio-ghibli-style-ai-art-crisis-openai Thoughts on what is gained and lost when anyone can replicate anything you create with just a few words, and what that says about our society. Guided missiles and misguided men.
Just saw an interesting way to respond to a bot on FB. Dude had made a post and the bot replied, "How are you doing?" When seems innocuous enough. Dude replies with "Spell pickle for me." The bot goes, "Can I send you a friend request?" Dude again asks it to spell pickle for him. This time the bot replies with the pickle emoji. Given that I've seen some questionable posts in my timeline, I think that I'm going to ask them that question to see what the result is.
> Have a popular anime > Amazon wants to stream it on their service. Great! "Okay so here's the video files, here's the subtitles --" "No, no, we have a subtitle generator AI, we don't need your stupid professionally generated subtitles" > Amazon AI hears someone talk to a character named Nika and . . .