Great essay! - a comment on "Biological Insights Could Become More Valuable":
I think another big challenge here is the burden of proving the value of the tech. The time it usually takes is so long, and monetization is so hard that we didn't see any major insight platforms emerge yet in biotech. In silicon like Cadenes case it's much easier, validation is days to months aways mostly, and in many cases the simulations are as good approximation of the physics you need. However, in bio, to validate a GLP-1 hypothesis for eg , you would have to run clinical trials and years would pass until you know the answer.
I hope the big bets on companies like Xiara, Recursion, Insitro etc. would prove valuable enough to let the industry take this necessary leap of faith into insight platforms in general. Then we can debate if there's a real moat to be had here.
Another highly impactful essay, which I will re-read in a few days time. One thought: for the sake of argument let’s assume the “bull” case regarding ML & antibody generation, enabling the industry to generate mabs of exquisite sensitivity and affinity. Now, of the >10000 diseases (majority of which are orphan), how many do you think are sufficiently well understood at the molecular level to enable them to be addressed using our new, improved methodology? I honestly don’t know the answer here but suspect that it’s a minority (and a small one at that). Thus, the rate-limiting step is likely to be fundamental disease understanding rather than the challenges inherent in drug discovery/development. (I think Robert Plenge at BMS has written about this extensively). Furthermore, consider two very high profile (and prevalent) diseases; Alzheimer’s & Parkinson’s. Here, we’ve experienced either clear cut failure or minor success despite (apparently) identifying the causal agent & pathway. I suspect “better” mabs will only shift the dial a little at most.
Packy. Thanks for linking to this.. WOW... so much to digest... this is a serious loaf of information and logic
Great essay! - a comment on "Biological Insights Could Become More Valuable":
I think another big challenge here is the burden of proving the value of the tech. The time it usually takes is so long, and monetization is so hard that we didn't see any major insight platforms emerge yet in biotech. In silicon like Cadenes case it's much easier, validation is days to months aways mostly, and in many cases the simulations are as good approximation of the physics you need. However, in bio, to validate a GLP-1 hypothesis for eg , you would have to run clinical trials and years would pass until you know the answer.
I hope the big bets on companies like Xiara, Recursion, Insitro etc. would prove valuable enough to let the industry take this necessary leap of faith into insight platforms in general. Then we can debate if there's a real moat to be had here.
Another highly impactful essay, which I will re-read in a few days time. One thought: for the sake of argument let’s assume the “bull” case regarding ML & antibody generation, enabling the industry to generate mabs of exquisite sensitivity and affinity. Now, of the >10000 diseases (majority of which are orphan), how many do you think are sufficiently well understood at the molecular level to enable them to be addressed using our new, improved methodology? I honestly don’t know the answer here but suspect that it’s a minority (and a small one at that). Thus, the rate-limiting step is likely to be fundamental disease understanding rather than the challenges inherent in drug discovery/development. (I think Robert Plenge at BMS has written about this extensively). Furthermore, consider two very high profile (and prevalent) diseases; Alzheimer’s & Parkinson’s. Here, we’ve experienced either clear cut failure or minor success despite (apparently) identifying the causal agent & pathway. I suspect “better” mabs will only shift the dial a little at most.
Five stars for this write up!! Really really excellent. 90% new information for me. Awestruck...