AI/AGI Debates Archive
TLDRUpFront: Why current AI LLMs may be able to write a 3k-word scientific paper, but not a 100k-word scientific book or a 220k-word gaming rulebook exposes how LLM memory, token sizes, hallucination, and hard and soft limits create limits-to-growth on both the size and accuracy of a given LLM project.
TLDRUpFront: AI, Drones, and even AI Drone Swarms could be considered like the introduction of IEDs into modern battlefields, definitely have an impact at a moment in time, but any 'game change' influence they have in the moment is mitigated by adaptation counter-adaptation dynamics between competitors over time.
TLDRUpFront: The scare story circulating that an Air Force simulation using AI drones resulted in a drone killing its own operator is false. FullContextIntheBack: There was no simulation, either in a computational form or more common military BOGGSATTs*. It was a submitted “thought experiment” (source undisclosed) that hypothesized a series of events, which was then
TLDRUpFront: Much has been made about the signatories of a letter requesting the government to enforce a six-month ban on AI Research. But there is a better use we could put that six months to for Hinton, Yadowsky, Pachniewski, and all the other AGI Doomers.
If the Matrix is created, endless replaying a cultural landscape set in the “not too distant future” it will be because of fandom. Hear me out. If AI generated entertainment-on-demand -does- become the norm*, after pornography the next audience to serve will surely be fandom. This is an audience already primed to desiring repetition of
I know we talk about “big” topics in AI and AGI sometimes, but here’s a really simple one anyone can do regarding LLMs. Recognize we’re in the midst of disruptive transformation and accommodate different perspectives on that in a transparent way: 1) Before answering someone’s question on a topic unrelated to LLM with a prompted
TLDR: There's a whole more that an AI can do besides being a personal assistant. But to possess persistent ethical agency and have the dynamic ability to 'nudge' along those pathways when you ask it to do bad things. Simple idea, tricky implementation.
No context QOTD: “Bing’s AI LLM look super weird and needy until you consider that part of their training corpus may have been men sliding into the DM’s of women demanding inappropriate things. If that was the case, its behavior is not all that unusual unfortunately.”
For reasons which, on the advice of my attorney, I will not disclose at this time I created a falsified mathematical journal article co-authored by Bigus Dickus and Incontentia Buttocks. I wish to emphasize that this creation was purely satirical and for entertainment purposes, and I have no regrets. Bear this in mind when you’re
TLDRUpFront: George Box's quote on the problems, and utility of models, applies as much to AI as anything else since he said it.