It is the ultimate definition of drinking from a firehose.
While this friction is painful, it’s helpful to remember that it is a temporary structural mismatch. History shows us that during the Industrial Revolution, the explosion of production capabilities completely broke existing legal, social, and environmental frameworks before society finally adjusted. A classic economic example is Jevons Paradox: when the price of coal plummeted, industries didn't use less of it; consumption exploded, and the entire macroeconomic landscape scaled up to match the cheap input. Biopharma is on the exact same trajectory.
Instead of panic, this shift requires a major philosophical and operational upgrade:
-First, reframing the "loss". The industry is suffering from a massive fear of missing out, lamenting the "good" drugs left behind in spreadsheet hell. But are these unpursued molecules truly a tragedy? No. If not for AI, they wouldn't have been discovered in the first place. This isn't a tragic loss; it is simply an unrealized opportunity cost.
-We need implement system architecture upgrades. To survive in a world of radical abundance, companies must abandon linear, rigid asset management. Portfolios must be operated on highly sophisticated, dynamic, and probabilistic pathways.
When inputs become effectively infinite, the old ways of doing business die. The incentives, architectural design, and regulatory systems must be fundamentally re-engineered to match our new reality. This will be a painful transition for sure.
Great post. Loved the completely different take on the use of AI in drug discovery. Biotech and the pharma world in general can no longer think in only one direction when it comes to dud development - adopt the next best tech tool quickly. A holistic, future centric approach built on critical thinking will be the key differentiator. Thanks Matt!
I read the article “Biotech Is Entering Its Netflix Era” and, honestly, for me it was a huge waste of time.
The piece is built on a premise that simply does not reflect the reality of those on the other side—patients, families, and communities living every day with rare, incurable diseases. The central idea suggests a future where there will be an abundance of therapies and the challenge will be choosing between them—almost like a Netflix-style catalog.
But this is not, even remotely, the reality we live in.
I speak here not as someone from the market, but as the wife of a patient with SCA3 (Spinocerebellar Ataxia type 3) and as an active member of the ataxia community in Brazil.
Our reality is very different:
There is no cure.
There are no disease-modifying treatments.
There are very few initiatives in development.
And time, for us, is not theoretical—it is disease progression in real life.
When I read an article suggesting that biotech’s main problem will be “too many options,” it feels completely disconnected. Because in practice, the problem is exactly the opposite: scarcity.
And there is one specific point that stands out even more: the recurring narrative in the ecosystem that there are not enough patients for clinical trials.
This simply does not hold true when we look at the Brazilian reality.
In Brazil alone, in the case of SCA3:
We have a structured community of nearly 5,000 individuals living with ataxias.
Among them, around 2,000 patients with SCA3—a highly relevant number for clinical studies.
And importantly, we have a remarkably rich genetic heterogeneity, which could add significant scientific value to therapy development.
In other words, patients are not the problem.
The problem is lack of connection, lack of engagement, and lack of real interest in looking beyond traditional markets.
And I can give a concrete example.
We have made multiple attempts to establish contact with Vico Therapeutics, which is developing the VO659 gene therapy, including for SCA3. These efforts came from organized patient associations, with data, structure, and a genuine willingness to collaborate.
The response was zero.
No demonstrated interest in understanding the Brazilian landscape.
No openness to acknowledging that these patients exist.
No indication of willingness to expand clinical efforts or even map this population.
So when I read discussions about future bottlenecks—about how to manage multiple available therapies or how to select patients across different options—it again feels like an attempt to solve a problem that does not yet exist for us.
Because those living with a rare, progressive disease are not worried about choosing.
They are waiting:
for at least one therapy to exist,
for someone to acknowledge this population,
for clinical trials to include broader geographic and genetic diversity,
and for the pace of science to keep up with the pace of the disease.
In Brazil, there is an active, committed, and resilient ecosystem.
There are patients, data, organized associations, and real advocacy happening every single day.
There are also many people working behind the scenes—quietly, often without support—trying to build exactly the bridges that this article seems to assume already exist.
Innovation in biotech is essential, and we all hope to reach a future where there are many treatment options.
But today, for diseases like SCA3, the conversation still needs to be different:
it is not about choosing between many therapies.
it is about ensuring that at least one reaches us—and that no one is left behind.
I get what you are saying and I feel your pain. I am afraid the article is talking about a different side of the problem: from the standpoint of big pharma, not the patient. The patient will never get enough cures, even if we spend trillions of dollars. But the point of the article is that big pharma only has so much bandwidth. As we get better in AI drug discovery, big pharma will have too many choices of drugs to work on and not enough resources to do it. This has nothing to do with patients unfortunately. Patients will continue to suffer.
Very few biomedical “discoveries” actually advance all the way through the development pipeline to a safe and effective “treatment”. If we overload the opening of the pipeline, the output at the other end (treatments) might actually decrease. So I think what Matt is writing about is relevant to all of us who care about both science AND patient care. The good news is that AI is also speeding up the translational machine from discovery to treatment and reducing the cost, so I think (hope?) it will be a net win, Catiana.
Great piece! The problem of judgement isnt new, pharma and VC all converge on the same diseases (validated endpoint, accepted biomarker, clean regulatory path, paying market…), which is why 95% of rare diseases have no approved therapy and a billion people with NTDs get 0.2% of biomedical R&D. Abundance of AI generated drugs wont fix that bias since the “narratively convenient science” is an existing pathology. We have to do better for all of humanity.
I think this is a great example of how when the constraint is scarcity of resources capitalisms is an effective means of deploying capital, but when that scarcity diminishes the market allocation of capital breaks down. How can we find other means of allocating capital that really are a boon to people and outcomes. Where can we use the state and existing scientific and biotechnical institutions to help slog through this backload and doing the background scientific research needed to keep innovation flowing? Great article and thank you for now giving me a new thing to worry about with AI
Fan of all your posts … and especially this one! I agree with the analogy of Netflix and what has happened in the cinema industry. I think where AI can unlock value is to redefine the system of drug discovery, not "just" doing better biology or chemistry
Why couldn't we increase the number and supply of clinical trial participants and streamline the trials with technology and more modern approach? How many % of humanity are currently involved in clinical trials?
Great post. What makes a biotech company stand out now may no longer be who can discover something first, but who has the infrastructure, biobanks, and operational capability to validate, approve, and distribute something faster than everyone else.
Insightful narrative. The scarcity of patient population for clinical testing has always been an issue, agree it is going to get worse. It will also lead to further increases in cost of development at later stages, which is already high. Engineering success of biology in generating asset bulk does not really improve PTRS.
Terrific final line-the migration of scarcity. The contrast between AI driven drug discovery and 1950s style drug development is really striking. Hopefully the recent FDA pilot program on real time feed of cancer trials will kickstart some genuine innovation in the space.
Worth watching what it does to the financing more than the science. Two proof-of-concept trials are already live, and a continuous data feed collapses the phase-gate scaffolding that venture money and pharma licensing synchronized around for seventy years. Lab innovation may be the easy part. Rebuilding the capital plumbing around real-time results is where it gets messy.
I wonder if we will begin to see a rise of "pharma patent trolls" for 'discovered" molecules like in high tech...I'm off ask AI how to patent molecules, bye.
For me, the harder question is what happens when AI creates too many plausible options: more candidates, targets, more “promising” programs. But we keep the same limits: patients, trial sites, capital, regulatory capacity, and time. Maybe the real bottleneck moves from discovery to selection.
It is the ultimate definition of drinking from a firehose.
While this friction is painful, it’s helpful to remember that it is a temporary structural mismatch. History shows us that during the Industrial Revolution, the explosion of production capabilities completely broke existing legal, social, and environmental frameworks before society finally adjusted. A classic economic example is Jevons Paradox: when the price of coal plummeted, industries didn't use less of it; consumption exploded, and the entire macroeconomic landscape scaled up to match the cheap input. Biopharma is on the exact same trajectory.
Instead of panic, this shift requires a major philosophical and operational upgrade:
-First, reframing the "loss". The industry is suffering from a massive fear of missing out, lamenting the "good" drugs left behind in spreadsheet hell. But are these unpursued molecules truly a tragedy? No. If not for AI, they wouldn't have been discovered in the first place. This isn't a tragic loss; it is simply an unrealized opportunity cost.
-We need implement system architecture upgrades. To survive in a world of radical abundance, companies must abandon linear, rigid asset management. Portfolios must be operated on highly sophisticated, dynamic, and probabilistic pathways.
When inputs become effectively infinite, the old ways of doing business die. The incentives, architectural design, and regulatory systems must be fundamentally re-engineered to match our new reality. This will be a painful transition for sure.
Please use less AI in your writing. This hurt to read.
Great post. Loved the completely different take on the use of AI in drug discovery. Biotech and the pharma world in general can no longer think in only one direction when it comes to dud development - adopt the next best tech tool quickly. A holistic, future centric approach built on critical thinking will be the key differentiator. Thanks Matt!
I read the article “Biotech Is Entering Its Netflix Era” and, honestly, for me it was a huge waste of time.
The piece is built on a premise that simply does not reflect the reality of those on the other side—patients, families, and communities living every day with rare, incurable diseases. The central idea suggests a future where there will be an abundance of therapies and the challenge will be choosing between them—almost like a Netflix-style catalog.
But this is not, even remotely, the reality we live in.
I speak here not as someone from the market, but as the wife of a patient with SCA3 (Spinocerebellar Ataxia type 3) and as an active member of the ataxia community in Brazil.
Our reality is very different:
There is no cure.
There are no disease-modifying treatments.
There are very few initiatives in development.
And time, for us, is not theoretical—it is disease progression in real life.
When I read an article suggesting that biotech’s main problem will be “too many options,” it feels completely disconnected. Because in practice, the problem is exactly the opposite: scarcity.
And there is one specific point that stands out even more: the recurring narrative in the ecosystem that there are not enough patients for clinical trials.
This simply does not hold true when we look at the Brazilian reality.
In Brazil alone, in the case of SCA3:
We have a structured community of nearly 5,000 individuals living with ataxias.
Among them, around 2,000 patients with SCA3—a highly relevant number for clinical studies.
And importantly, we have a remarkably rich genetic heterogeneity, which could add significant scientific value to therapy development.
In other words, patients are not the problem.
The problem is lack of connection, lack of engagement, and lack of real interest in looking beyond traditional markets.
And I can give a concrete example.
We have made multiple attempts to establish contact with Vico Therapeutics, which is developing the VO659 gene therapy, including for SCA3. These efforts came from organized patient associations, with data, structure, and a genuine willingness to collaborate.
The response was zero.
No demonstrated interest in understanding the Brazilian landscape.
No openness to acknowledging that these patients exist.
No indication of willingness to expand clinical efforts or even map this population.
So when I read discussions about future bottlenecks—about how to manage multiple available therapies or how to select patients across different options—it again feels like an attempt to solve a problem that does not yet exist for us.
Because those living with a rare, progressive disease are not worried about choosing.
They are waiting:
for at least one therapy to exist,
for someone to acknowledge this population,
for clinical trials to include broader geographic and genetic diversity,
and for the pace of science to keep up with the pace of the disease.
In Brazil, there is an active, committed, and resilient ecosystem.
There are patients, data, organized associations, and real advocacy happening every single day.
There are also many people working behind the scenes—quietly, often without support—trying to build exactly the bridges that this article seems to assume already exist.
Innovation in biotech is essential, and we all hope to reach a future where there are many treatment options.
But today, for diseases like SCA3, the conversation still needs to be different:
it is not about choosing between many therapies.
it is about ensuring that at least one reaches us—and that no one is left behind.
I get what you are saying and I feel your pain. I am afraid the article is talking about a different side of the problem: from the standpoint of big pharma, not the patient. The patient will never get enough cures, even if we spend trillions of dollars. But the point of the article is that big pharma only has so much bandwidth. As we get better in AI drug discovery, big pharma will have too many choices of drugs to work on and not enough resources to do it. This has nothing to do with patients unfortunately. Patients will continue to suffer.
Your systemic assessment of biomedical science’s big new challenge (too much discovery!) is spot on, Matt. Thanks for your perceptive writing. 👏
Include that the day when so-called “excess discoveries” actually lead to cures for more than 7,000 rare diseases, affecting over 300 million people worldwide—and when more than 5% of them have approved treatments—then I will agree that we are living in an era of excess discoveries. ( https://www.rarediseasesinternational.org/wp-content/uploads/2024/06/Briefing-Note-WHA-Resolution-on-Rare-Diseases_June-2024.pdf)
Because for millions of people, this is not about efficiency.
It’s about possibility.
Very few biomedical “discoveries” actually advance all the way through the development pipeline to a safe and effective “treatment”. If we overload the opening of the pipeline, the output at the other end (treatments) might actually decrease. So I think what Matt is writing about is relevant to all of us who care about both science AND patient care. The good news is that AI is also speeding up the translational machine from discovery to treatment and reducing the cost, so I think (hope?) it will be a net win, Catiana.
Great piece! The problem of judgement isnt new, pharma and VC all converge on the same diseases (validated endpoint, accepted biomarker, clean regulatory path, paying market…), which is why 95% of rare diseases have no approved therapy and a billion people with NTDs get 0.2% of biomedical R&D. Abundance of AI generated drugs wont fix that bias since the “narratively convenient science” is an existing pathology. We have to do better for all of humanity.
I think this is a great example of how when the constraint is scarcity of resources capitalisms is an effective means of deploying capital, but when that scarcity diminishes the market allocation of capital breaks down. How can we find other means of allocating capital that really are a boon to people and outcomes. Where can we use the state and existing scientific and biotechnical institutions to help slog through this backload and doing the background scientific research needed to keep innovation flowing? Great article and thank you for now giving me a new thing to worry about with AI
Fan of all your posts … and especially this one! I agree with the analogy of Netflix and what has happened in the cinema industry. I think where AI can unlock value is to redefine the system of drug discovery, not "just" doing better biology or chemistry
would love if you go through my article https://arcpath.substack.com/p/mastering-the-bone-marrow-assessment?r=8153x6&utm_medium=ios
Love this analogy! Interesting read.
I’m a high school senior and I write SPLICED. - weekly illustrated syn-bio breakdowns for non-scientists.
Why couldn't we increase the number and supply of clinical trial participants and streamline the trials with technology and more modern approach? How many % of humanity are currently involved in clinical trials?
Great post. What makes a biotech company stand out now may no longer be who can discover something first, but who has the infrastructure, biobanks, and operational capability to validate, approve, and distribute something faster than everyone else.
Insightful narrative. The scarcity of patient population for clinical testing has always been an issue, agree it is going to get worse. It will also lead to further increases in cost of development at later stages, which is already high. Engineering success of biology in generating asset bulk does not really improve PTRS.
Terrific final line-the migration of scarcity. The contrast between AI driven drug discovery and 1950s style drug development is really striking. Hopefully the recent FDA pilot program on real time feed of cancer trials will kickstart some genuine innovation in the space.
Worth watching what it does to the financing more than the science. Two proof-of-concept trials are already live, and a continuous data feed collapses the phase-gate scaffolding that venture money and pharma licensing synchronized around for seventy years. Lab innovation may be the easy part. Rebuilding the capital plumbing around real-time results is where it gets messy.
I wonder if we will begin to see a rise of "pharma patent trolls" for 'discovered" molecules like in high tech...I'm off ask AI how to patent molecules, bye.
For me, the harder question is what happens when AI creates too many plausible options: more candidates, targets, more “promising” programs. But we keep the same limits: patients, trial sites, capital, regulatory capacity, and time. Maybe the real bottleneck moves from discovery to selection.
Check out my profile/post, I post monthly!
https://adamorogers.substack.com/p/the-great-biotech-divide-venture