Vibe coding doesn't scale past Sunday. It's a weekend high, and a great one, but on Monday the codebase has no owner, the bugs have no origin, and "ask for random changes until it goes away" stops being amusing...
Vibe coding was fun because you forgot the code existed. Puzzle coding is fun for the opposite reason — the model is finally working with your mental model instead of around it.
Vibe coding doesn't scale past Sunday. It's a weekend high, and a great one, but on Monday the codebase has no owner, the bugs have no origin, and "ask for random changes until it goes away" stops being amusing...
Traceability falls out for free. When something goes wrong, the skill, the schema, the hand-off between two nodes are all there to point at. Swap one piece; don't re-roll the whole thing.
Seeing what ChatGPT can achieve with text content I wonder how long it still takes to have the "AI Producer" for computer graphics, and frankly I dont think that would be more than hand full of years.
There was this paper that had the idea of deforming a plane into any watertight mesh. I am missing some detail about the type of deformation, maybe a fold, and the title of it. Anybody has any clue?
Seeing what ChatGPT can achieve with text content I wonder how long it still takes to have the "AI Producer" for computer graphics, and frankly I dont think that would be more than hand full of years.
"Can we build an AI producer that can create a movie given any novel?" in 3D and with the ability for new content to be produced on demand. For my Master thesis I worked on generating human body movement given a target; say kick this position in space, with variable styles...
While presenting my results, to my advisor, Prof. Martin Giese, he asked what would an ultimate tool I want to have with such tech. My answer was a new movie by Charlie Chaplin, where his presence and style is preserved and reproduced for a new content.
Some of you may have noticed I like to smooth things with Savitzky–Golay filters rather than Gaussian filters - why? SavGol filters preserve the peaks and troughs in a signal whereas Gaussian filters with large sigmas produce a damping effect causing trajectories to shrink:
We're presenting our latest and greatest hits tomorrow (Wed, Oct 20, 2021) at the @3dbodytech conference!
I'm so excited to hear many other presenters already talking about Meshcapade :)
Find more deets here: https://t.co/HbuZIgaVqm
If you want to hear what @PerceivingSys has been doing, all you need to do is attend #ICCV2021 session 9A today, Wed, 4-5pm CEST (10am EDT). Come get to know PoP, SPEC, PARE, ROMP, ACTOR, SNARF, SPICE, DSR, SOMA, and SAMP. https://t.co/60s8nRqd0U