2 ms·
Some days I wish I could just run DFT on everything and have it spit out the ones with "interesting" results and then patent those. Unfortunately, this is not
by dekhn 3y ago
Some days I wish I could just run DFT on everything and have it spit out the ones with "interesting" results and then patent those. Unfortunately, this is not feasible (computationally) nor likely to produce very many true positives.
I wonder if you could troll theorists into doing simulations of just about any substance by publishing convincing-looking papers in arxiv. Scratch that: we know it's possible.
- carabiner 3y agoYou think it's over? Or are we back?
- dekhn 3y agoI don't think these results will replicate and it's still unclear whether the material has any useful properties. But I could be wrong. The theory sims say nothing; it's a bunch of noise right now.
- sbierwagen 3y agoBrute forcing the space of all chemicals isn't practical, so just do what humans do, generate plausible candidates with a neural net and then check with DFT. The patent system in its current form isn't going to survive the next few decades, on this I agree. First step is probably going to be chopping down the patent length from its current 15 year term.
- yreg 3y agoWe are starting to do things like that with some drugs afaik.