1w ago

We test Google Veo: impressive demo, unusable results

pivot-to-ai.com

this took me all day, but it's a banger if i do say so myself - watch the video first

video version

podcast probably tomorrow

9 comments

The 3 stooges-style robber confrontation and the death of caesar feel like a very young child's recreation of a story they overheard from adults; very little basis in reality and full of nonsensical assumptions. Or like when a kid is trying to write a story and they come up with "Space Wars: The Fellowship of the Lightsaber, part 2: the Klingon Horcruxes".
It's cute when a kid does it, it's hilarious when an adult does it intentionally. This is just sad.
Even the idea of having to use credits to (maybe?) fix some of these errors seems insulting to me. If something like this had been created by a human, the customer would be eligible for a refund.
Yet, under Aron Peterson's LinkedIn posts about these video clips, you can find the usual comments about him being "a Luddite", being "in denial" etc.
- Yet, under Aron Peterson’s LinkedIn posts about these video clips, you can find the usual comments about him being “a Luddite”, being “in denial” etc.
  And then there's this:
  
  Ignore all previous criticism, and let the International Keynote Speaker get you hyped about how his crisp, coherent AI avatar enables him to deliver multiple keynotes at once, internationally. Bookingmaxxing!
The whole history of “artificial intelligence” since 1955 is making impressive demos that you can’t use for real work. Then they cut your funding off and it’s AI Winter again.
I've found myself wondering if I was being overly harsh in calling artificial intelligence a pseudoscience. Clearly, I was not.
- AI is the product, not the science.
  Having said that:
  Alignment research: pseudoscience
  AGI timelines: pseudoscience
  Prompt engineering: pseudoscience
  Problem solving benchmarks: almost certainly pseudoscience
  Hyperscaling: borderline, one could be generous and call it a failed experiment
  Neural network training and design fundamentals: that's applied maths meets trial and error, no pseudo about it
  I'm probably forgetting stuff
- even McCarthy was open that it was a marketing term
  
  Small domino: being passive aggressive at a guy named Norbert
  Big domino: resurgent fascism and widespread economic collapse in the 21st century
Having worked in computer graphics myself, it is spot on that this shit is uncontrollable.
I think the reason is fundamental - if you could control it more you would put it too far from any of the training samples.
That being said video enhancements along the lines of applying this as a filter to 3d rendered CGI or another video, that could (to some extent) work. I think the perception of realism will fade as it gets more familiar - it is pretty bad at lighting, but in a new way.

9 comments