Business Insider: AI Startup Valued at $2B
Probably. After all, most ventures fail. Tech investors, like bidders at an auction, are subject to the Winner’s Curse: those who err on the side of too much optimism are willing to pay more than those who calculate the risk/reward ratio correctly. They win the auction by overpaying.
From what I can glean from the BI article, the play here is to better evaluate the chances that a drug candidate will ultimately prove safe and effective. Given that the costs of failure can run from tens to hundreds of millions of dollars, even a modest ability to recognize and eliminate losers from the development pipeline would be very valuable indeed.
Of course, this is not a new problem. There are thousands of successful drugs and many more thousands of unsuccessful ones, and drug makers have developed a number of “rules” that distinguish winners from losers
From https://www.archie-west.ac.uk/pr…
Lipinski’s Rule is the most famous of these, stating that orally active drugs violate no more than one of the following criteria:
- No more than 5 hydrogen-bond donors
- No more than 10 hydrogen-bond acceptors
- Molecular mass < 500 Da
- A level of fat solubility (expressed as log octanol-water partition coefficient) less than 5
There are many variants and exceptions to this rule, but you get the idea.
Of course, incumbent drug companies are not blind to the notion that computerized analysis can give them an edge in identifying the best drug candidates. One report indicates that 18 of them have ongoing AI programs. Is BenevolentAI (the unicorn in question) going to beat them all? Maybe. Is that a good bet at a $2B valuation? Seems unlikely.
And how plausible is the underlying value proposition? The cost of drug development keeps going up logarithmically, and has for decades, a phenomenon known as Eroom’s Law (Moore’s Law in reverse).
From Eroom’s Law
Drug discovery technology is light-years more advanced now than it was 50 years ago; a number of technological revolutions have been implemented. But the cost of discovering drugs just keeps shooting up. This suggests to me that technology is not the limiting factor here, and thus better technology is unlikely to be the solution.
If there were a derivatives market in bad investments in biotech, I would go short on this one. But I would do the same with nearly all biotech investments. It’s a very tough game.