There aren't enough smart people in biology doing something boring
Author: Abhishaike Mahajan and Eryney Marrogi
Abhishaike Mahajan is a senior ML engineer at Dyno Therapeutics, a biotech startup working to create better adeno-associated viral vectors using ML. He also writes a blog (http://owlposting.com) focused on the intersection of biology and ML.
Eryney Marrogi is a 2nd-year medical student at the University of Vermont, and is currently working on AI protein design for continuous monitoring devices at Caltech. Before medical school, he was in the startup world, working primarily on making new viral vectors for gene therapy as one of the earliest employees at Dyno Therapeutics. He came to Dyno from Flaminia Catteruccia and George Church's labs at Harvard, where he assisted in deploying CRISPR-Cas9 for the development of gene drives in mosquitoes. He likes writing about biology (on his Substack), and loves hearing input about what topics he should consider!
Help Seeds sprout this spring with a 33% discount on our subscription!
There aren't enough smart people in biology doing something boring. At least in industry.
If you work in biology for long enough, you’ll eventually realize that most decent or ambitious companies in this field are run by exactly one type of person. They are often deeply curious, hard working to the point of near pathology, and will almost always end up pursuing some sort of crazy pie-in-the-sky mission. Like curing aging or making de-novo proteins in a zero-shot manner or trying to usher in entirely new dogmas in biology. In other words, something where immense intellectual output leads to outsized market payoff.
The companies they start will usually have this thesis. In this pursuit, they will spend millions, sometimes billions, of dollars’ worth of venture-capital and government grants and philanthropic subsidy dollars. They live and breathe biology, and their ultimate goal in life is to have some sort of fundamental impact on the field at large. The people underneath them will usually not be too dissimilar.
Now, most decent companies in any other field are run by a similar type of person, with one important distinction: they don’t demand as much intellectual satisfaction.
Stripe is a decent example of this. Stripe is a fundamentally boring business on the surface — you’re making it easier for people to send money to each other through the internet. It’s not exciting in the same way that, say, Google was, with their much more grandiose vision of ‘indexing the world's knowledge’. The interesting bits of Stripe are perhaps found in how you build such a payment system and the potential second/third/fourth order effects that easier money exchange has on the world. But the header line is boring. And I have no doubt that Patrick Collison — the CEO of Stripe — is enormously intelligent and could’ve easily pursued something with a higher level of intellectual ‘taste’.
But he didn’t. He did payment processing.
And he did it well enough to turn Stripe into the largest private fintech company in the world, with a valuation of $65 billion and over $1 trillion in payment volume over its lifetime. Alongside making Patrick a billionaire, the democratization of online payments — boring as it sounds — almost certainly changed the world for the better. Improved efficiency of businesses, more small businesses being launched, and reduction of financial crime. Even more impressive is that the best engineers in the world aim to land a job at Stripe, even today, 15 years after they launched.
Patrick entered MIT majoring in math and, later on, physics. Would it have been better for him to have become a math researcher? Or perhaps, more pertinent to this essay, a biology researcher, given his current interest in it? Maybe society would’ve got some wonderful things from a mind like his being focused on scientific subjects, and perhaps Stripe genuinely was a waste in the grander scheme of things.
But…I think there’s already a lot of very smart people in biology trying to do crazy things. There are relatively few smart people in this field trying to do boring things and do them well. Put another way, everyone of note in biotech wants to be George Church (arguably one of the greatest biology researchers alive), no one wants to be Don Combs (the founder of New England Biolabs, the reagent manufacturing company fueling America’s biological research engine).
But that’s not the case in software.
Stripe isn’t unique here. Facebook is another one, Shopify is another one, Zoom is another one, DocuSign is another one. All of these are run by outwardly smart people who are doing something that feels very not smart! Social media, easy set-up of e-commerce, video chatting, document signing. All of them are taking one relatively boring idea and doing it at scale to benefit hundreds of millions of people. And doing it in a way that doesn’t want to make people tear their hair out, which is surprisingly hard.
I think, fairly, we could quibble as to whether these companies truly are ‘not smart’. Doing anything at scale very quickly runs into ‘you need very smart people to coordinate the mess’, even Docusign, who people often point to as an example of needlessly large companies (which isn’t true!). But I feel pretty confident that none of these companies’ missions could be phrased in a way that a snobbish intellectual would find convincing.
Software is filled with people who will happily do the boring, but deeply impactful, thing. People may well point out that pure software may struggle with a sense of grander ambition, but I don’t think even that is true. Yes, there is a glut of yet-another-payroll-software startups, but there’s also software startups like Exa.ai, Replit and Haize Labs who have decently insane missions. Smart people creating software companies have somehow stumbled across a perfect mix of ‘ambitious and impactful’ and ‘boring and impactful’.
Biology has overly indexed on the former, despite it being far, far harder. Why?
Here’s one answer: the historical role that for-profit biology has played is basically a single thing: developing drugs. Or developing a platform for drugs or developing a tool to do lead optimization of drugs or some other drug-related [thing]. Everything drug-related is usually expensive, which means you usually need someone else’s money to do it. And if you’re using someone else’s money, you need to convince them to bet on you amongst the dozens of other companies also trying to develop drugs.
And the best way to do that is to say you’re doing something radical.
Something no one has ever done before. Revolutionary even. A new assay, a new way of thinking, a new way of interrogating biology. After all, many biotech startups are academic spinouts, which are a great testing ground for galaxy-brained ideas. Each time it happens, the founders do it knowing that the regulatory hurdles are massive, but they do it regardless — because they ultimately think their approach will be so good it just blows away the regulators.
Of course, almost all of the time, this isn’t smoke and mirrors! Most of these life-sciences companies purporting radically new ideas are, indeed, pursuing those radically new ideas.
But what may very well be smoke and mirrors are how necessary those radically new ideas actually are.
Thinking purely about drugs for a second, consider that almost 24% of all drugs that enter clinical trials are abandoned due to ‘strategic business decisions’. Given this, it feels deeply unlikely that the medical industry lacks interesting leads. You could explain this away by citing potential efficacy concerns with these drugs, but given the success that people have had in reviving shelved drug assets — a practice called drug repositioning — the shelving feels much more closely related to random economic winds. People have found a degree more success in drug repurposing, which is just picking up already-FDA-approved drugs for one condition and seeing if they work for another condition.
Funnily, the semantic difference between these two terms (repurposing versus repositioning) is still a bit fuzzy, so papers may use them interchangeably, but both are astonishingly efficacious and cheap. And also, boring to pursue. Fellow biology writer, Trevor Klee, has written about one such drug repurposing success story in the past, I’d highly recommend reading it.
Now, fairly enough, there are lots of non-scientific issues with doing drug repurposing/repositioning. Enforcing patents on repurposed drugs is challenging since off-label prescription of generic drugs is hard to prevent. Regulatory bodies require the same level of safety evidence for each new drug indication, even if there’s already prior evidence for its safety. And, of course, there are immense organizational hurdles in pharmaceutical companies pursuing indications for shelved drugs outside of their core competency. As this Nature paper discusses, all of these have fixes, but are understandably challenging to address.
Let’s forget drugs entirely. Creating drugs is hard, maybe the boring ideas like drug repurposing and drug repositioning are genuinely insufficient. What if we look outside of that?
To offer one example, I think better CRO’s are an excellent candidate for ‘boring but impactful’. If scientists had access to extremely high quality CRO’s, I have zero doubt that a fair bit of biology research would be more ambitious, more accessible, and more replicable. This is especially true as more and more computational people leak into the life-sciences field, many of whom don’t understand the finer details of the wet lab (me included!).
Yet, few dependable CRO’s exist.
I’ve written about Plasmidsaurus before as one such good one, given their combination of valuable service rendered + cheap + fast + easy to work with. Are there others? Charles River Labs? Maybe Twist Biosciences? But, past those, the average experience of working with a CRO is dealing with costs that are nearly price-gouging, months of waiting time, tons of back-and-forth emails, and shoddy results.
Typically, the service that CROs provide is bad enough that an exceptional scientist will still trust their own capability to outperform a CRO on most wet-lab tasks, even in areas outside of that scientists core competency. Biologists will make their own viral vectors instead of outsource, or clone a ton of individual plasmids in the most painful way possible, all because they don’t trust the existing marketplace to adequately serve their needs, or provide a decent one at a reasonable cost.
Why is the state of most CRO’s so dismal?
Here, it feels challenging to say anything that isn’t ‘the best people don’t want to start a CRO’. CRO’s are boring and mostly monotonous work. Their innovations are largely going to be logistical ones, not scientific. The most skilled wet-lab people, the people most suited to dramatically change this space, would much rather work on something far more ambitious.
I understand this mentality and empathize with it. I just think it’s a shame!
There are people trying to alter this. I think Adaptyv Bio is a fantastic example of someone recognizing this market need amongst computational scientists and addressing it. I hope they succeed! But I’d ideally want more — a flood of talented bench lab biologists who have some deep expertise in some set of widely-used techniques and create companies hyper-focused on democratizing those to others.
There’s reason to think the time is ripe for better CRO’s in general.
CRO’s have a reputation of being a grueling place to work at, given the thin margins and speed demands. In turn, this means the workforce is typically inexperienced, freshly graduated researchers (who typically hope to jump to a better company after a few years or head off to grad school in an attempt to avoid future bad employers altogether), which contributes to the often low-quality results produced by them. Experience can have a surprisingly large impact in the result of any moderately complex lab experiment — even when hyper-detailed protocols are available.
It’s hard to fix this problem outright. If you hire better scientists, you need to pay more, which means you need to increase your prices, which means you’ll be undercut by people charging less. And, unfortunately, a ‘high quality results means a high price point’ corporate thesis is as hard in biology as it is hard for every single other non-luxury industry. You need to ensure high quality while keeping prices low if you want to have any real hope of acquiring life-sciences customers.
Better lab automation may be the way to achieve this. This is how Plasmidsaurus keeps their prices low and speed high; almost all of their sequencing results are performed by a fleet of Opentron machines. Currently, lab automation excels at extremely routine experiments/assays that don’t deviate much from one run to the next. Sequencing is often a great fit for this, given both how routine and how mechanically simple it is. But there are many other actively-used wet-lab procedures that don’t quite fit that bill. Perhaps better lab robotics could help, or better programming tooling assisted via english-to-code LLM’s.
But, as Josie Zayner states in her article, the value add of further lab automation isn’t clear-cut. From the piece:
So, while lab automation may help for some routine tasks that aren’t particularly physically complex, going far beyond that may not be worth the effort. She goes on to state that the truly needed innovation in the lab is in better remote access. Again, from the piece:
Science really needs Remote Access and Remote Control. Say on my way to work in the morning I figure I need turn on a couple of incubators or change the temperature in some because I had a culture expressing proteins. Maybe I wanted to equilibrate my H(F)PLC column? What if I could do this from my cell phone? The thing that needs to be optimized in Science is not the "work time" it is the "down time". When I need to wait 30 minutes for my H(F)PLC column to equilibrate and cannot use it or when I need to wait for a centrifuge to cool down or an incubator to heat up or many of the other things. This is where time is wasted. Transferring data files between computers or restarting a experiment on a piece of hardware remotely this would drastically help Scientific productivity.
Figuring all this out sounds boring. Building and testing all of it sounds boring. Finding out how to slightly improve the efficiency of a lab to ensure high quality results sounds boring. Ensuring that your customer base is consistently happy with your services (which account for a very tiny step in the grand picture of science) sounds boring.
But it’s needed, and your prize for engaging in it and doing it well will be a near guarantee of personal wealth and (positive) impact on how science is done. The only cost is silencing your internal snob for what a ‘real’ biology company should be doing. And, in time, I hope biology as a field begins to care more about impact than about doing the thing that sounds the most interesting. Or, at least, weigh it similarly to how software does it.
Obviously, I’m grateful for the people doing ambitious things in biology, because ambitious things do have outsized potential payoffs that boring companies often don’t have. I think the company I work at, Dyno Therapeutics, is doing something ambitious, as is Gordian Bio, Cradle, and many others. But more boring, but useful, companies run by smart people would be a fantastic addition to the ecosystem. There really aren’t enough of them.
If you consider yourself a boring biotech company, reach out to me! I’d love to hear your story and potentially write about you. People going down that path deserve more attention.
That’s technically all I have to say, but there’s some immediate responses that people may have to this essay that I think would be worth responding to.
Pure software is different. In biology, boring companies usually aren’t going to become Stripe-sized companies. Yeah, probably. Vertical business models in biology, which most boring non-therapeutic biotechs will be, are pretty profit-capped. But like…I don’t get the obsession with wanting to become a $1B+ dollar company. You could personally become extraordinarily rich by doing the boring stuff! You could personally have a massive impact on science by doing the boring stuff! Achieving a multi-billion dollar valuation is something venture capitalists and public market investors should care about, and I think it’s bad for anyone who doesn’t fall into those camps to be psyoped into personally caring about that too.
Making money at all in biology requires being a therapeutics company, which requires you to do something exciting. I get that it’s a popular trend amongst more boring biotechs to eventually dip their toes into therapeutics to make more money, like Schrodinger. But Schrodinger did fine for 30 years on just making already-made molecular dynamics software easier to work with — a pretty boring mission. I think they underperformed relative to their own sense of ambition, given their decision to IPO a few years back, but they had become a staple piece of software to have in many industry and academic labs prior to that. Widespread industry penetration, stickiness amongst its user-base, and surviving for decades is a success, and I’m sure made the original founders very rich. Moreover, as Trevor Klee has previously written about, Viking Therapeutics had an $8B dollar market cap and also had the aforementioned boring thesis of drug repurposing. Being exciting is not a prerequisite to doing something useful, even in the therapeutics landscape.
If a startup doesn’t have an exciting/disruptive scientific thesis, it’s not going to be funded by VC’s. One, I don’t think that’s true at all, there are plenty of boring biotech companies that have gotten funding. Kaleidoscope focuses on just making software for collaboration in biotech R&D; boring, and are funded by Dimension and Hummingbird. Culture Biosciences takes bioreactors, connects them to the cloud, and allows you to rent space in them to iterate on your ideas cheaply; boring, and are funded by Craft Ventures and Northpond Ventures. And so on. Two, a sufficiently boring idea shouldn’t need that much VC funding, or maybe even any at all. Just like…build it, ask people to try it, and iterate. If you aren’t throwing stuff into clinical trials or doing large-scale R&D, you probably don’t need much funding.
If you enjoyed this post, please consider supporting us with a one-time donation or a paid subscription. All proceeds go to the authors we feature here on the Best of Science Blogging feed (myself not included; see: Bucks for Science Blogs: Announcing the Subscription Revenue Sharing Program).
Help Seeds sprout this spring with a 33% discount on our subscription!
And when they ask what we want? We answer: A world where mercy isn’t passport-bound. https://thehiddenclinic.substack.com/p/the-price-of-a-pulse