A generation of start-ups have failed to live up to the hype. Executives are now betting that more powerful tools will crack the complexities of human biology
Archive link: https://archive.ph/Hge9G
Because LLMs (where all the hype has been, the past few years) are the wrong AI tool for drug discovery.
Even if you end up making a new drug the entire shitshow that goes behind the equivalence of double checking it to make sure it doesnt do anything horrible prevents any of this from becoming instantly profitable. From my knowledge they probably came in expecting profit instantly not realizing this is a minimum of 5-25 years. You have to get really lucky or almost computer level of knowledge to make it and get it approved quickly. The real problem is the fact that you have to prove what every variant can do so a drug that has like 30+ variants wouldnt be worth it.