Presumably one manifestation of this strategy is a phenomenal podcast series the firm produces, featuring senior partners, distinguished guests, and frequently both, discussing with some granularity a topical issue in technology or entrepreneurship.

While the podcast often features a portfolio company (which the hosts are not shy about mentioning), the value of the podcast lies in the quality and (apparent) authenticity of the conversation – it’s pretty remarkable to hear Marc Andreessen discuss the cloud with Salesforce founder Marc Benioff, for instance, or to hear Benedict Evans comparing the perspectives of Apple and Google. While many VC and other investment and advisory firms have started to offer promotional podcasts, these often come across as guarded, scripted, and corporate; the a16z podcasts, in contrast, tend to project immediacy and refreshing candor.

In this context, I was thrilled to discover a recent episode entitled “When Bio Meets Computer Science.” As expected, it was articulate, thoughtful, and relatively grounded. Nevertheless, I realized at the end I didn’t entirely agree with all the assertions, and I thought it might be worth some discussion.

A16z’s high level thesis – with which I strongly agree– is that there are profound opportunities at the confluence of biology and technology, especially as more undergraduates emerge who are trained in both (apparently the majority of Stanford students take at least one computer course, even though most don’t wind up in computer science or engineering). It’s also true that the new opportunities aren’t simply being able to do existing activities faster or at larger scale – rather, it’s the opportunity to ask and pursue questions you couldn’t have even conceptualized in an earlier era.

The a16z view of this space, as a whole, can be divided into three parts: (1) Digital Therapeutics; (2) Cloud Biology (not what you think); and (3) Computational Medicine.

(My usual disclosure: I work at a cloud genomics company; I don’t believe we either do business with or directly compete with companies backed by Andreessen-Horowitz, nor have we received funding from this firm. It is true that one of the most perceptive articles I’ve read recently about genomics was written by Mark Kaganovich, the CEO and co-founder of SolveBio, an a16z-backed company.)

The central arguments of a16z:

(1) Digital therapeutics. Many healthcare problems are behavioral, and digital health companies might address these challenges more effectively, and less expensively, than drug companies. Representative (and portfolio) company: Omada Health (my 2012 post on the company is here; our recent Tech Tonics interview with founder Sean Duffy is here).

(2) Cloud Biology. The arrival of highly automated labs will revolutionize biology startups in the same way the arrival of cloud computing revolutionized technology startups. These lab facilities will enable biotech startups to do the experiments they need with greater reliability and without the capex spend; scalable research facilities will be available when needed, and you only pay for what you use (again, like cloud computing). Representative companies in this space: Emerald Cloud Labs; Transcriptic.

(3) Computational Medicine. Physicians and researchers must contend with an overwhelming and ever-increasing amount of data. For instance, a key challenge in oncology is matching many potential cancer drugs to the exact characteristics of the tumor in question. These sorts of problems can be solved with software, which is continuing to get better and cheaper.  Representative companies: Foundation Medicine cited as company in the oncology diagnostics space.

Collectively, a16z says, these three trends will lead to an “explosion” of experimental biology and digital health startups, offering the opportunity to develop clinically impactful products for a fraction of the cost of a traditional life science startup, and representing a pointed contrast to the ever-increasing cost of traditional biotech drug development, which seems to follow so-called Eroom’s Law (Moore’s Law in reverse).

I fervently hope these three focus areas revolutionize biology and health the way a16z envisions, but I’d place my bets deliberately at the moment.

First, I worry a16z profoundly underestimates the difficulty of developing a behavioral intervention that’s actually clinically effective, and also underestimates the cost of rigorously proving this (and payors are going to demand credible data before agreeing to cover the cost of an intervention such as Omada’s or Propeller Health’s, which both, by design, target a relatively large number of covered lives). (Addendum: A thoughtful response from Omada notes that they “take on meaningful financial risk,” collecting a “significant portion” of their revenue from customers “only as individual patients achieve clinically-meaningful milestones” — an approach they acknowledge makes for a “nerve-wracking business model.”)

Second, I worry a16z misunderstands why drug development is so expensive: it’s largely because it’s so difficult to figure out early on what is and isn’t going to work, and so you wind up spending millions of dollars on Phase 2 and Phase 3 studies that ultimately fail. Even successful clinical development programs are fairly expensive (i.e. even if you don’t add in the cost of failures). Thus, taking some of the cost out of early research may not be particularly transformative – although, if it somehow increases the probability of success (i.e. due to increased reliability of the data, as the founders typically argue), that could be significant. The idea of enabling more researchers do to more early studies and better studies with the same amount of money is also appealing (if true); perhaps it will take a visionary entrepreneur and a compelling use case (analogous to what Benioff did with customer relationship management using the cloud) to demonstrate the power and potential of this emerging technology.

Third, I’m not sure that the key problem we face in areas such as oncology is the algorithm challenge a16z describes – i.e. matching the right patient to the right drug at the right time. Rather, the more substantial challenge, I’d argue, is the lack of any effective drug for many cancers, and the need to come up with far better treatment options for most. (The recent progress in immuno-oncology seems especially promising.) If fancy algorithms could help us identify powerful new drugs (or new uses for existing drugs) – that would be a substantial advance, and many Silicon Valley companies (Capella Biosciences, led by Pek Lum, and NuMedii, led by Gini Deshpande) are intensively working on these two respective challenges.

Fourth, a16z seems to have a somewhat outdated view of biotech entrepreneurship and life science venture. For instance, we’ve already seen the rise of ultra lean, asset-focused drug companies (as David Grainger and I have discussed here, here, here), and giving rise to the sort of structural considerations Bruce Booth describes here. Moreover, the pitying tone used to describe the poor life science VC contemplating a massive capital investment in a new biotech company also contrasts with the data presented in this post by Booth and Bijan Salehizadeh, debunking a number of myths about life science investing.

Fifth, and perhaps most importantly: I worry that in applying their “software eats the world” model so insistently, a16z might overlook key discrepancies, areas where the analogies they reach for don’t quite capture the challenges they’re describing. The concern is that in looking so intently to replicate successful technology structures (like the cloud) in new domains (like biology), they may lean too heavily on superficial similarities, and miss substantive distinctions.

I was also surprised by the a16z perception that doctors view medicine as “here’s a problem and we’re going to fix it.” In my experience – and as I argued explicitly in one of my earliest blog posts – it’s more often the technologists who tend to bring this solutionist mindset to healthcare; providers and researchers who are living in, and mired in, the wonderful, frustrating, chaotic mess of healthcare delivery and medical science often have more nuanced, less binary view of their role and their challenges.

Even with these reservations, I emphatically share a16z’s optimism that emerging technologies – and the creative thinking that often accompanies them – can have a profoundly positive impact on healthcare and medicine. Key hurdles we face in achieving this include: (a) figuring out where best to apply emerging technologies; (b) retaining a healthy respect for the difficulties of modifying behavior to the point of achieving meaningful impact; (c) retaining a similarly healthy respect for the difficulties of understanding biology to the point of identifying a promising opportunity for intervention; (d) ensuring our primary focus — and ultimate measure of success — remains the lives we’re able to impact, and not the technology we develop in effort to achieve this. I’d point to Counsyl (no financial relationship to disclose) – a diagnostics company not in the a16z portfolio, but co-founded by Balaji Srinivasan before he joined a16z as a partner – as an example of a company that seems to get all this exactly right.