While these people are obviously professionals in their field, anyone who touched pharma/bio IT with a ten foot pole understands how much there is to improve there. And that's just the beginning - imagine how much there is to improve in understanding of the body overall?

It's not about "AI will discover a cure for X in 10 years".

But it's about having the applied intellect of a an engineer, chemist, biologist, and all mixtures of that, available on demand, in a second, at scale, enabling you to do everything around it faster, better and cheaper.

Every step of drug discovery, synthesis and testing can be improved.

Over 90% drugs tested in humans fail. And maybe 0.1% of overall discovered ones. While yes, testing on humans is the bottleneck, we can fix problems around the bottleneck easily.

If you had a data processor and 90% of data that goes into it is corrupt, the processor isn't the bottleneck yet - the data quality is.

If we can create better biological models to test these drugs, better interaction models, improve pharmacokinetics, improve yield, we can go from 1000 discovered, 1 works, to 1000 discovered, 100 works. We can predict some candidates will fail due to better models and not test these. We can improve how they are delivered to the body and make them more effective. We can catch potential side-effects early on and avoid going to trials at all.

This is what AI can do even nowadays, let alone 5 or 10 years from now.

And that is a huge improvement. It revamps the industry.