Alex Telford is the founder of Convoke. We cover the ins and outs of developing and pricing new drugs, what can be done to shorten the time it takes to bring drugs to market, and why AI won't be a silver bullet for the industry.
Principles & Lessons:
1) Ideas from Basic Science Require Considerable Translation Alex highlights that “it starts with academic research,” mostly funded by the NIH in universities. But once a target is discovered, “you then have to do a lot of work to convert it into a drug,” which means turning an initial hypothesis into something safe, manufacturable and clinically tested. He stresses that the “steadily escalating sequence of trials” – Phase I through Phase III – involves many years and significant cost before the FDA (or other regulators) will allow any product to reach patients.
2) We Are Running Out of “Easier” Small-Molecule Discoveries Reflecting on Paul Janssen’s era in the 1950s, Alex explains there used to be “a lot of opportunity” to modify or create new compounds from nature. Now, “the kind of low-hanging fruit have...been exhausted.” This is especially clear in older drug classes such as antibiotics or classical small molecules. He contrasts that with “the rise in biotech” from the 1980s onward, which introduced new methods (like recombinant DNA) to unlock different therapeutic possibilities.
3) Complex Modalities Are the New Frontier Alex describes how modern drug development increasingly focuses on advanced therapies such as “gene therapies,” “cell therapies,” and “radiopharmaceuticals.” He notes that “CAR T-cell” therapy involves extracting a patient’s own immune cells, gene editing them “to have this sequence protein inserted...then reinfusing them.” He mentions how it creates “a super complex manufacturing process,” making it harder to copy but also offering new avenues to treat diseases that don’t respond to older methods.
4) Regulatory Tension Between Speed and Safety Comparing the process to the AIDS epidemic, Alex highlights that “there’s no one-size-fits-all solution” in deciding how quickly to approve drugs. With HIV in the 1980s, “patients were quite rightly saying that we’re going to die anyway...you should let us try these drugs.” That pressure helped create “the accelerated approval pathway,” but he points out that this “regulatory arbitrage” now steers investment into diseases with fewer or no existing treatments, because “the requirement for evidence is lower and the need is greater.”
5) Randomized Controlled Trials Are Essential but Not Always Feasible When discussing vitamin D, Alex notes that observational data alone was misleading: “people with low vitamin D levels have generally worse health outcomes,” but trials showed that supplementation may not help. He calls RCTs “the gold standard” because they randomize patients, avoiding hidden biases. Still, he acknowledges that “if you have a condition with too few patients,” single-arm studies and “digital twins” might be a better route than demanding a large, randomized approach.
6) Profit Motives Shape Which Drugs Get Made Alex likens pharma economics to “a lottery-type model” where “a few really huge winners” pay for the many failures. A “blockbuster” is a drug that earns at least $1 billion per year, and it’s these blockbusters (like Humira) that often generate most of an entire company’s returns. Because so much hinges on finding a rare mega-winner, firms often ignore smaller patient populations: “it’s just not worth big pharma’s time...if it makes a few tens of millions a year.”
7) Balancing Drug Pricing and Innovation Is Tricky Drawing attention to U.S. pricing, Alex remarks that “the U.S. accounts for an outsized share of the revenue” in new drugs, in part because “you can charge whatever you want.” This can feel egregious, but he warns that “if you put pressure too much on drug prices, you can...remove the incentives for people to invest.” The system relies on occasional extreme blockbusters to support the “many losers who don’t recoup the cost of development,” so undercutting that big payoff might erode overall research.
8) Integrating New Technologies and Improving Processes Offer Hope While discussing AI, Alex says he is “short-term pessimistic, long-term optimistic,” clarifying that alpha in the near term may come from “automating some of the processes along with drug development,” such as speeding up trial enrollment or drafting regulatory documents. Over time, AI might help in “scalable tinkering,” reminiscent of how “millions and millions of Marios” tested jumps in a video game. He suggests that bringing faster feedback loops into drug development, combined with flexible regulation and better biomarkers, could lower costs, expand which diseases are targeted and eventually spur more innovative drugs.
Transcript