MBI Daily Dose (July 08, 2025)
Companies or topics mentioned in today's Daily Dose: Therapeutic abundance, Agent Archetypes, OpenAI's SBC, Rent growth
I was admittedly a bit skeptical about AI revolutionizing therapeutics. While I understood the AI’s potentially profound impact on drug discovery process, I wasn’t sure if it can have a sustained impact given the regulatory process can act as a bottleneck. So, even if some part of the overall drug value chain speeds up materially thanks to AI, I thought it would still be difficult to accelerate the overall process.
However, Jacob Kimmel’s piece “Creating therapeutic abundance” has prompted me to change my mind a bit. I overestimated the regulatory bottleneck and underestimated scientific bottlenecks in medicine in the status quo and if AI can help us with the latter, perhaps therapeutic abundance is still on the cards.
It is bit of a long excerpt, but that only shows my appreciation of the overall piece:
Most of the knowledge of drug program lifecycles remains locked within drug companies. Nonetheless, we can bucket the failures into a two broad categories of safety and efficacy and make informed estimates.
- Safety failures – ~20-30% of all candidates
A molecule was developed, but proved unsafe in patients. These are typically detected as failures in Phase 1 trials. - Efficacy failures – 70-80% of all candidates
The remainder of all drug candidates that fail – 63% of all drugs placed into trials period – fail due to a lack of efficacy. Even though the drugs are safe, they don't provide benefit to the patients by treating their disease.
From these coarse numbers, it's clear that the highest leverage point in our drug development process is increasing the efficacy rate of new candidate medicines
…our main challenges are scientific. We simply don't know how to make effective drugs that preserve health or reverse disease! If we want more medicines, we need to understand why they don't work and fix it.
Concretely, a scientist sits and thinks hard about the problem, makes a guess at the responsible molecular players based on their intuition, prior art, and their new data, then tests to see if the molecule is causal. The vast majority of these hypotheses are wrong! The few that prove to be correct often become the basis of our modern target-based drug discovery process and several companies quickly launch programs to prosecute them.
Is it possible to build a more deterministic, less constrained discovery process? Can we discover target biologies with a complexity matching the origins of disease?
Two technological revolutions argue in the affirmative. Functional genomics methods now enable us to test far more hypotheses than ever before. From the resulting data corpuses, artificial intelligence models can search otherwise intractably large hypothesis spaces, like the space of possible genetic circuits or combinatorial therapies. By performing most experiments in the world of bits rather than atoms, it’s possible to address questions that were inaccessible to a previous generation of scientists.
In practice, this allows researchers to treat the cell as the unit of experimentation, increasing the throughput of many target discovery questions by 100-1000X. These methods aren't applicable to every target discovery problem (e.g. some pathologies only manifest across tissue systems), but they nonetheless unlock a class of putative interventions that were previously too numerous to search effectively.
Artificial intelligence models can learn general models from the data generated in functional genomics experiments of many flavors, predicting outcomes for the experiments we haven't yet run. If we manage to construct a performant model for a given class of target biologies, we may be able to increase the efficiency of target discovery by many orders-of-magnitude. The cost of discovering a target could conceivably go from >$1B to <$1M.
Agents are the topic du jour these days. The Information yesterday published a helpful piece on agents, segmenting them in different categories, relevant products, and common use cases:
On a different piece, The Information published about OpenAI’s Stock Based Compensation (SBC) which apparently was ~120% of their revenue in 2024. Of course, OpenAI expects SBC to have tremendous leverage over time as they expect the revenue (the denominator here) to increase much faster than SBC. Meta’s hiring spree will almost certainly force OpenAI to be even more generous with their SBC program. Is this some 4D chess by Zuck to make life difficult for a company that is legitimately threatening to join the big tech club? I am not a big believer in that. I think the surge in compensation for AI “super talent” is the natural market outcome, especially after Meta hit some road blocks in their existing effort. If building compelling AI products requires abundant compute and AI talent, Meta has a lot of compute but it is lot less useful without the latter.
If OpenAI continues to march forward without much of a hiccup, it will give us a clear indication that the talent density may run pretty deep in OpenAI. At the same time, if Meta’s strategy proves successful in building SOTA models and AI products, that may stamp the primacy of AI talent even more, increasing compensation for AI super talents even higher. People talk about how AI may enable a billion dollar company run by one individual, but we may see a billion dollar “employee(s)” first!
In addition to "Daily Dose" like this, MBI Deep Dives publishes one Deep Dive on a publicly listed company every month. You can find all the 60 Deep Dives here. I would greatly appreciate if you share MBI content with anyone who might find it useful!
US apartment rent growth YoY was below 1% in 2Q’25. One anecdotal experience that I can share is I moved to Sacramento in December 2023, and recently moved out to rent in a different neighborhood. I was talking to my previous landlord a couple of weeks ago and it turns out he had to rent the place at the same rent he was renting me back in December 2023. Considering it’s an annual lease for the next tenant, my landlord effectively “froze” rent for 30 months!
Current Portfolio:
Please note that these are NOT my recommendation to buy/sell these securities, but just disclosure from my end so that you can assess potential biases that I may have because of my own personal portfolio holdings. Always consider my write-up my personal investing journal and never forget my objectives, risk tolerance, and constraints may have no resemblance to yours.