GENERATIVE AI

→ View original post on X — @garymarcus

ChatGPT’s Deceptive Design: Anthropomorphism Without Genuine Intent

By

@garymarcus

–

09 April 2023 23h16

ChatGPT really really wants* me to like it. *actually it has no beliefs or desires whatsoever, but its canned answers…

Understanding in LLMs: Emergent Properties Beyond Token Prediction

By

–

09 April 2023 22h57

Bullshit machines might be harsh. The real debate is if the predicting the next token requires these models to build some level of understanding — even if it isn’t reliable. The whole emergent property of LLMs is fascinating, even if they can’t hierarchically, decompose their

→ View original post on X — @esyudkowsky

GAN trained on Internet data would require superhuman cognition

By

@esyudkowsky

–

09 April 2023 22h10

Sure, obviously a GAN trained to the limit of possibility on modeling the probability distribution from which the entire Internet was drawn would have to contain superhuman cognition within it.

Multi-ControlNet and Open Source AI Tools Overview

By

–

09 April 2023 22h01

Old blog Incase helpful: https://
open.substack.com/pub/creativete
chnologydigest/p/multi-controlnet-and-open-source?r=13o7oi&utm_medium=ios&utm_campaign=post
…

ControlNet and Ebsynth advance photorealistic video generation techniques

By

–

09 April 2023 22h01

Aside from Gen 1 ControlNet + Ebsynth still produces the most photorealistic results. Areas for improvement:
– Need a way to force consistency in style across keyframes
– Need to exploit temporal data in an image sequence (PS: optical flow might be good for keyframes selection as

Roblox Pursues YouTube Integration Through 3D Generative AI

By

–

09 April 2023 21h53

If you prefer a continuous reading experience, here's the unabridged blog post: https://
creativetech.beehiiv.com/p/roblox-wants
-youtube-3d-generative-ai-holds-key?_gl=1*mvrc3r*_ga*MTk2MDY4Nzc3LjE2NzQ0NDUyODY.*_ga_E6Y4WLQ2EC*MTY4MTA2NzQ0MC4xNC4xLjE2ODEwNjk4ODIuNjAuMC4w
…

→ View original post on X — @garymarcus

AI Safety Requires Better Control and Regulation for All

By

@garymarcus

–

09 April 2023 21h51

“AI safety isn’t a right or a left issue. New, poorly controlled AI that is unreliable…can easily fool people & may set off a wave of cybercrime, & lead to…atmosphere of distrust. It’s in everybody’s interest that we move to a safer, better controlled form of AI.” –
@garymarcus