How exactly could science solve the problem of aging and death? Could AI help with this, or is it just an unsolveable problem of life.Personally i want to die, but it seems like a huge accomplishment if we ever manage to do it.
>>17086198Mind uploading
this gay board needs an "AI" general
>>17086198I am a computational lead at a gerontology series B startup. If you are actually interested in the field, you would have heard about this company sometimes last year for sure.Deep learning is very useful for all kinds of things in bio. The main problem is that none of the smart people want to touch gerontology due to a catch-22 problem: everything so far has been pure scams, so people are worried they'll kill their career if they're associated with this, and won't work on anything useful because it's a scam.Where I work, the main model we work on is amateurish because instead of people with a background in comp bio like me or applied bio, the company was built by mathematicians and physicists, so they think they can reduce complex processes to 2 probability distributions of choice and 4 variables. As a result I'm stuck leading a project that is inherently doomed and trying to steer it in a good direction by convincing them we actually do need to look at transcriptomics data and proteomics data and shit like that."AI" is completely worthless. Even things like pLMs and DNALMs, let alone ordinary language models. While there are endless scams about how great and magical and perfect they are, our peers at AZ and novartis and amgen confirm our findings: they have 0 correlation with biological reality, no matter what metric you try to track. That is, again, totally unlike bespoke deep learning models, where we have a handful of amazing tools (though most still fail as they overfit too much due to the complicated nature of bio data and batch effects).The key problem in gerontology is we don't have enough fine-grained longitudinal data over long periods of time, and we have not found any animal model that properly recapitulates human aging.Deeply integrating bio experiments with computational tools, as was key to making oxford nanopore tech work at all, is how this will be solved. But that requires understanding this need, which nobody does right now.
>>17087882>Deep learning is very useful for all kinds of things in bio. The main problem is that none of the smart people want to touch gerontology due to a catch-22 problem: everything so far has been pure scams, so people are worried they'll kill their career if they're associated with this, and won't work on anything useful because it's a scam.I think that's a big problem. As soon as you mention you want to cure aging, people think you're doctor Frankenstein or some weirdo who's afraid of death. But if you really question the arguments against longevity, none can be made that can't also be applied to every form of medicine. It's weird how this isn't treated like some prestigious final problem of biology like fusion power in physics or the millennium problems in mathematics, but instead as some tower of Babel that should not be built.
>>17086198
>>17086198Anti-aging Light
>>17087975>>17088005Big if true
