This is a fascinating post – thanks for writing. I think there’s a lot of truth in the argument. It’s consistent with other pieces I’ve seen and my own prior experience as an academic. However, I think the argument still lacks power because of the difficulty in quantifying the extent of the problem.
You rightly point out the slow rise in the age at which PI’s get their first NIH R01 grant (rising from a mean of 37 in 1995 to 41 in 2020 according to my squinting at your graph) and the striking decline in new VC funding. But elsewhere you write things like: “the existence of breakthrough technologies and continued scientific progress is undeniable - from mRNA vaccines to large language models, from CRISPR to renewable energy advances. Individual scientists, engineers, and entrepreneurs continue to push boundaries and achieve remarkable breakthroughs. Yet this observation makes the systemic problem even more striking: these advances emerge despite our funding structures, not because of them.” I’m sure that’s true to some extent but how much? To take CRISPR as just one example, my guess would be that Jennifer Doudna, one of the leading lights in this field, has been well supported by the NIH.
And while I agree with you that “the drive to make processes more predictable and measurable can lead to a system that prioritizes the reproduction of past patterns over the emergence of genuine novelty”, again I wonder how do we quantify that effect? Can we? The risk of propagating this kind of stagnation seems to be likely to grow as funders start to apply AI to analyse the outcomes of prior decision making.
Finally, I also agree that we need to “innovate in how we see innovation”, but how do we do that exactly? Does it not take us back to the perennial problem of persuading funders and researchers to take risks – and then properly assessing the yield from that risk-taking?
Sorry for all the questions. I do think this is a genuinely important and difficult problem. While one can try to diagnose general problems with 'the system', I'm interested in how we get to the root of those problems (which likely operate in different parts of the system) with the clarity needed to plan and test some kinds of intervention.
Thank you for these thoughtful questions. You raise an important point about CRISPR and Jennifer Doudna's NIH support. However, I think this example actually reveals some deeper complexities in our funding systems. While Doudna did receive significant NIH funding, she's also had to navigate substantial challenges - including strategic funding decisions that led her to work with Gladstone Institutes specifically to avoid profit-driven constraints on rare disease research.
More importantly, for every Doudna who successfully navigates these challenges, we have cases like Katalin Kariko, whose revolutionary mRNA work was repeatedly rejected for funding, or Carol Greider, who received an NIH grant rejection the same day she won her Nobel Prize.
This points to what I believe is the real quantification challenge: how do we measure not just the successes, but the near-misses? The breakthroughs that almost didn't happen? You're absolutely right that we need better ways to quantify these dynamics - not just to identify what's broken, but to understand what kinds of innovation our current systems might be systematically overlooking.
I'd love to continue this discussion - I think your questions about measurement and evidence get at something crucial about how we evaluate and improve our innovation systems.
This is a fascinating post – thanks for writing. I think there’s a lot of truth in the argument. It’s consistent with other pieces I’ve seen and my own prior experience as an academic. However, I think the argument still lacks power because of the difficulty in quantifying the extent of the problem.
You rightly point out the slow rise in the age at which PI’s get their first NIH R01 grant (rising from a mean of 37 in 1995 to 41 in 2020 according to my squinting at your graph) and the striking decline in new VC funding. But elsewhere you write things like: “the existence of breakthrough technologies and continued scientific progress is undeniable - from mRNA vaccines to large language models, from CRISPR to renewable energy advances. Individual scientists, engineers, and entrepreneurs continue to push boundaries and achieve remarkable breakthroughs. Yet this observation makes the systemic problem even more striking: these advances emerge despite our funding structures, not because of them.” I’m sure that’s true to some extent but how much? To take CRISPR as just one example, my guess would be that Jennifer Doudna, one of the leading lights in this field, has been well supported by the NIH.
And while I agree with you that “the drive to make processes more predictable and measurable can lead to a system that prioritizes the reproduction of past patterns over the emergence of genuine novelty”, again I wonder how do we quantify that effect? Can we? The risk of propagating this kind of stagnation seems to be likely to grow as funders start to apply AI to analyse the outcomes of prior decision making.
Finally, I also agree that we need to “innovate in how we see innovation”, but how do we do that exactly? Does it not take us back to the perennial problem of persuading funders and researchers to take risks – and then properly assessing the yield from that risk-taking?
Sorry for all the questions. I do think this is a genuinely important and difficult problem. While one can try to diagnose general problems with 'the system', I'm interested in how we get to the root of those problems (which likely operate in different parts of the system) with the clarity needed to plan and test some kinds of intervention.
Thank you for these thoughtful questions. You raise an important point about CRISPR and Jennifer Doudna's NIH support. However, I think this example actually reveals some deeper complexities in our funding systems. While Doudna did receive significant NIH funding, she's also had to navigate substantial challenges - including strategic funding decisions that led her to work with Gladstone Institutes specifically to avoid profit-driven constraints on rare disease research.
More importantly, for every Doudna who successfully navigates these challenges, we have cases like Katalin Kariko, whose revolutionary mRNA work was repeatedly rejected for funding, or Carol Greider, who received an NIH grant rejection the same day she won her Nobel Prize.
This points to what I believe is the real quantification challenge: how do we measure not just the successes, but the near-misses? The breakthroughs that almost didn't happen? You're absolutely right that we need better ways to quantify these dynamics - not just to identify what's broken, but to understand what kinds of innovation our current systems might be systematically overlooking.
I'd love to continue this discussion - I think your questions about measurement and evidence get at something crucial about how we evaluate and improve our innovation systems.