One thing I learned from working in high tech industry: don't trust any claim until you've had time to investigate it, understand it, and see it in work with your own eyes. Too many times I've been told about "gee whiz" capabilities of software only to discover that it was merely a gleam in the "inventor's eye" and not really present in the software. The software industry is probably the worst at making unsubstantiated claims for its products.
One thing that makes me suspicious about the above claims at 4:40 about medicinal chemistry using this genetic algorithm approach to "evolve new drugs" is the very fact that I don't see any major drug company successfully using this technology. I do know that big drug companies are desperate for new drugs and pay billions to their labs to develop new drugs. If a multi-million dollar software package could obviate the need to invest billions, drug companies would be beating down the door of Eric Bonabeau's Icosystem Corporation. This company is over ten years old and isn't a billion dollar company and on everybody's lips. That says to me it doesn't do the gee whiz developments that the above video claims.
The fact that the company web site names Nymbler as a "product" says to me that the gap between the claimed capabilities and the actual delivery is an awesome unbridgeable chasm:
The ominous panda generator and the addictive baby-naming site Nymbler are two examples of what we, at Icosystem, call the "hunch engine."To claim that you can discover new drugs and test complex engineered system but cite as your "products" a baby naming algorithm and panda expression generator apprears to me to be a monumental "bait-and-switch" con. You claim to be able to generate billion dollar values with new drugs and sophisticated testing, but you point me at rinky-dink applications like baby name generators and panda expression generators!
This article in The Atlantic doesn't change my mind.
No comments:
Post a Comment