Interview with Curie.Bio
AMA-style interview with Founder/CEO, Zach Weinberg, and Head of Founder Strategy and Operations, George Voren
This is reposted from a Q&A we did in February 2023. Check out our Slack for more.
We were lucky to have the opportunity to interview co-founder & CEO, Zach Weinberg, and Head of Founder Strategy and Operations, George Voren, from the new biotech venture fund Curie.Bio!
Read more to learn about:
💶 What Curie.Bio looks for in their investments
🎯 Platform vs. program ventures
💻 Role of software in the biotech industry
🍕Pizza Fridays
and more!
Question 1:
Lukas: Already starting with a question for @zach
what are the current challenges for the field of ehr analysis? which tooling is missing the most?
Zach: I'm not sure "ehr analysis" is really a field
Zach: That's a tactic
Zach: I'd always ask the same question: what problem are you solving? why is ehr data necessary? why is it better than alternatives? etc etc
Zach: So many people start with "i want to use ehr data" but have no idea why
Zach: The problems depends on what you're trying to do with it
Question 2:
Nicholas: Huge thanks to @zach for participating in the q&a today. i’ll get us started with a few topics, but as always, please jump in with your own questions! thanks for kicking it off, lukas
Nicholas: You’ve now started 3 different companies. your first was in digital ads, the second was in healthtech, and now you’re fully into early stage drug discovery. talk us through each transition – what was most surprising and why did you make the leap each time?
Zach: Hi all. i also have @george from our team joining me. he runs our founder experience group and thinks about how we interact with founders from first contact through the entire process.
Zach: Transition from ad tech -healthtech -biotech...
Zach: The why is simple - i got bored at each one and wanted a new challenge that was closer and closer to what i consider important innovation.
Zach: But that's more of a personal preference. i would say the challenge is always the same...there's a lot of context you have to learn and you have to learn it from really smart people. so the only shortcut is there is no shortcut. you go and find the absolute smartest people in the industry, build a relationship and ask 1000 questions. it takes time. took me 2 years for each to finally feel comfortable (we started researching curie in early 2021).
Question 3:
Nicholas: Congrats on the recent launch of curie! how do you see what you offer as distinct from existing modes of funding and why do you think now is the right time for this?
Zach: we believe that early stage therapeutics companies need 2 things: (1) money. (2) in the weeds help with extremely complex drug discovery challenges, from target selection to experiment design to vendor selection/contracting to expert calls to assay development etc etc
Zach: Right now, it's hard to get both at the seed without basically handing over control to an incubator
Zach: You can get money, but you get very limited help from most vc funds
Zach: Or you can go to a flagship style money, with a ton of people around the table, but that's not your business
Zach: That's theirs
Zach: So we are trying to create a model that gives you both. seed capital where you are in charge of your own business and an insanely experienced team who works with you in the trenches
Zach: Not some bullshit "help" aka we make email intros
Zach: Actual, detailed work
Zach: So that's what we're designed to do. we are not meant to be the cheapest money to be clear. we are meant to increase your chances of developing a successful drug.
Question 4:
Nicholas: On your website you mention an explicit interest in programs over platforms: “however, value inflection in biotech is driven by drug discovery, so we only invest if our funds can go towards progressing a lead program versus solely platform development.” this community has a lot of biotech platforms and i’m curious if you can expand on your thoughts here.
George: to be clear, we aren’t anti-platform. we just require that the investment into the company at least partially goes towards making progress on a lead program or two, and not *only* in the platform.
Zach: (2 of our first 4 portco's are platforms btw)
George: Historically in biotech, the value of the platforms is really based on the success of the first few drugs that come out of it. and that’s what we help with - directing where to apply the platform to engineer best-in-class therapeutics.
George: A common pitfall we see when evaluating companies that are quite a bit down the road in their development is that they spend a large portion of their resources in building out a platform and then do not have any assets to demonstrate the utility of that platform (yet). at that point, it’s on the founders to convince the investors there’s a valuable drug that can come out of that. our goal is to help founders convert resources from platform to asset as early as possible so they can be most efficient in generating their first high quality drug.
Nitya: What would you say are some current examples of platform companies that embody this approach well?
George: fun fact… if you define a successful platform company as one that has created 4+ fda approved drugs… out of the thousands of biotech companies that have been created, you can count the number of “successful” platform companies on one or two hands. i think companies like relay therapeutics on well on their way there as they have three clinical stage molecules in a relatively short timespan that came out of their platform within their first 6 years… and a pipeline of many more. however, that took significant amounts of capital to get there and they’ll be judged on the success of their first couple drugs.
Nicholas: This focus on the initial drug makes sense to prove out the platform, but aren’t there potential pitfalls in being so focused on a lead asset? one thing that i’ve heard is that the pressure to show proof points forces the company to push something out before it might otherwise be ready and that it’s often the second asset that demonstrates the true power of the platform
George: Good point. the big value of platforms is the opportunity for multiple assets, so it’s always a multiple shots on goal approach, multiple program(s). it’s true that it is often not the first program that ends up being the one the company is known for but it’s important that you believe in each asset to put forward. this is what gets the future investors excited - material progress on multiple assets with an opportunity for more down the line.
Zach: The key in the seed is giving series a investors a reason to believe this will work as a therapeutics. we think of it not as a single program, but a set of top programs where the platform can really shine. in a sense - "programs" is the better term. once you get to an a, you can raise more money at higher prices and actively pursue a few programs at once (maybe it's the second or third asset that shines). you probably don't get a 4th unless #1-#3 shows material promise.
Zach: So identifying the top 1-3 is absolutely critical
Jack: Seems like this approach can lead to a lot of tech debt that you then have to fund the 'burn down' of later. do you find that to be the case?
taking a platform approach first can incentivize more scalable science...
Zach: Re: tech debt, i think the theme is "there is no later if you can't prove the platform can actually work in making a drug"
Jack: Very fair reply :wink:
Jack: I worry that we leave a trail of debris in pursuit of the singular goal for the first 1-3 programs.
Luis: I think a lot of the super optimistic cs folks in here really need to hear this now, tomorrow and every day.
Zach: Imo tech debt is not a bad thing if you understand the tradeoffs
Zach: Startups are about hitting milestones...you have to trade off having the most scalable infrastructure with time/progress on applying the tools.
Zach: So that's part of what we're helping with...how can we cost effectively apply the seed dollars to show real progress. hit the milestone. money gets cheaper. now you can invest some of the larger $ back in the tech. make more progress. money gets cheaper again. repeat repeat repeat.
George: To the above… so then when you consider all of the other companies that have been successful… it’s based on the success of their first 1-2 drugs… and the value of their platform is in retrospect only tied to the success of those drugs.
Question 5:
Nicholas: Are there comparables in the tech world that you look to? i’ve heard the comparison to yc a bunch, but i don’t know if that’s actually a good description of yc
Zach: It's hard to compare to tech because tech is so much easier for founders. let's be honest...starting a software company is not that difficult relative to biotech. i know, i've started 2 of them. it's a walk in the park compared to therapeutics.
Zach: So i think yc is a fine comparison in spirit, in the sense that we believe in motivated, founder led teams who have the right incentives and support
Zach: But the details are massively different
Zach: And that's where the comparison really ends
Zach: We are therapeutics experts - this is all we do
Question 6:
Nicholas: Most of your career has been in software-native companies – where do you see the role of software in early stage drug discovery?
Zach: increasingly valuable in many places but by no means a silver bullet. we love great software when appropriately complemented with the right strategy, the right wet lab partners, the right experiment design etc.
Zach: In a sense, software is an increasingly important component of a very complex set of steps
Zach: Key is just understanding the details of what value the software is actually providing
Zach: Faster? cheaper? more accurate? better?
Zach: I know it's kind of a bullshit high level answer but that's the truth
Zach: Software doesn't really mean anything unless it's evaluated in the context of the problem it's supposed to solve
Question 7:
Nicholas: Let’s dive in here! this community loves nothing more than to dive into the details of software applied to biological problems :slightly_smiling_face:
more specifically, what parts of the drug discovery process do you see as being the most amenable to acceleration via software? both internal (e.g. as part of a platform) and external (e.g. you’re looking for a software company to solve this problem for all of your portfolio companies)
George: a. internal - anything that can identify patterns from large or multidimensional datasets that are otherwise challenging for a human to understand. gathering insights from screening of a large number of molecules and leveraging the data to learn and design the next iteration will yield major improvements in efficiencies by reducing the design-make-test cycle. as an example, one area we’re excited about is machine learning applied to dna-encoded library screening and using information from screening a large diverse set of molecules to generate an improved library for follow-up screening.
a. external - biotech still suffers from a highly fragmented data management problem. integrating internally generated experimental data with data generated from various external vendors is a challenge. we have seen and are beginning to explore solutions with our founders.
Jack: really interested to learn more about this.
we don't have a good git for science (and git is for code, not data). everyone says they are working on it – but it is such a fragmented ecosystem right now.
Zach: We're also excited about software that can make any parts of the pre-clinical / discovery work *cheaper*.
Zach: Cheaper = more shots on goal = more wins over the long run
Nicholas: How does one prove that it makes work cheaper?
Zach: Use it in a real world setting :slightly_smiling_face:
Zach: It's not easy
Zach: But any tool that can take a person and make that person 2x+ more efficient ultimately has potential to make a process cheaper. just a matter of showing it can work in the real world.
Jack: +1 to above. would like thoughts on the post-discovery part too.
development, tech transfer, and commercialization are slow and expensive. why don't we talk about that enough?
Question 8:
Robert: What are your thoughts on desci/daos in science? i'm still not sure i fully understand what they actually are concretely :sweat_smile:, but my feeling is that the problems they're tackling are better solved without crypto?
Nicholas: :trollface:
Jack: Adjacent use case wise – i've seen some compelling use cases for chain of custody for single patient therapies.
(not using desci/dao afaik, but mentioned such could be used to great successes.)
Zach: ^ "kickstarter for science projects" is my takeaway
Zach: Sounds fine, let's not pretend this is something more than that
Question 9:
Nicholas: Do you believe lims/elns make work cheaper?
George: Re: lims/elns: only if it makes you not have to repeat experiments or design your next one better :slightly_smiling_face:
Jack: I don't see lims/eln being the solution with thoughtful design, fair data, and integrating design/outcomes into a unified data set (i.e. not have to repeat experiments or design your next one better). hot take but not a lot of tools i've seen used in labs take that full stack and deliver a solution.
what does curie see in that space?
Zach: We don't really focus in that space.
Jack: Would that not fall into your message? seeking to learn more about how others see the difference. always seeking a better solution and education here :pray:
we're also excited about software that can make any parts of the pre-clinical / discovery work *cheaper*.
cheaper = more shots on goal = more wins over the long run
Zach: We're excited about it, but we don't spend time on it.
Jack: Ah, i think i mis-understand. you like/support but don't invest in that space?
Zach: We're happy to be customers once things are cheaper tho!
Zach: Correct
Nicholas: I want to drill in here because i think there’s an intersection of a few points we discussed above. i would consider not having an eln/lims as taking on some amount of tech debt, but i think it also likely speeds you up in the near term. so i’m curious if you encourage your portfolio companies not to take on an eln/lims until they have initial proof points/raise a series a/etc.
Zach: Just depends on the company. you have to fit the right operating model to the problem you're working on.
Zach: Otherwise you're a hammer searching for a nail
Question 10:
Jay: @zach. congratulations on the amazing launch yesterday. quick question: what is your policy on pizza fridays and has it evolved over the years? :slightly_smiling_face:
Zach: @jay i would say i took a very hard line on the value of pizza fridays while at flatiron. over time i've gotten more certain i was right.
Zach: High end weekday catering for free was a low interest rate phenomenon
Jay: :slightly_smiling_face: you're an innovator in more ways than one. in seriousness, amazing what you all are doing with curie, and stoked to be working together again.
Question 11:
Nicholas: Zach you have a lot of experience in the development/clinical side of the space. curious if you see increasing value (and opportunity) in bringing patient data earlier in the discovery process
Zach: Yes, if it's the right data for the problem at hand. "fit for purpose" as they say.
Zach: For example, one of the first 4 companies we funded is leveraging an established retrospective biobank for some of the seed work.
Zach: Without that biobank, the seed plan doesn't work, we don't fund it
Zach: Focused in ms
Zach: So in that scenario, the patient data here was robust: clinical record + imaging + banked blood
Zach: And we designed the seed plan w/ the founders to leverage this dataset
Zach: We also use patient data sets where appropriate to evaluate market size but that's less exciting
Question 12:
Nicholas: That brings up the question of publicly available data — there’s always the question of how clean/reliable those datasets are. when someone comes to you with a proposal to leverage that data, how do you evaluate how useful a resource it will be?
Zach: I know it's not a satisfying answer but it depends on what they're using the dataset for. if everything is publicly available, consider me a skeptic there's anything unique in their approach that won't be immediately competed away.
Zach: That's not to say you need a unique data set to build a company, you don't
Zach: You need a unique idea, a unique insight
Zach: What is it that you know that others don't? why have others missed this? what gives you confidence you're right? what data would make you change your mind?
Zach: We ask these questions (and more)
Question 13:
Nitya: Out of curiosity, if an internal data integration platform was built within a biopharma co, what would a solution that counteracts privacy concerns as a result of that easier access include? most preclinical info/ internal protein data sets etc remain siloed within specific teams within the org (wasted opportunity given that data is often not used beyond beyond that team for other analysis but i’ve found a lot of hesitancy in making that process easier)
Jack: I've chatted with a few companies and lawyers about this. short version: pharma risk averse and slow regulatory maturation.
dm me to chat more.
Jack: There is also an element of "i'm a unique snowflake so there is no value to unifying data outside of my silo".
Erik: Jack's answer resonates with my experience, and i bet others in #c03qqs39z6f|embedded-data-teams would concur.
Jack: ^ and my second message isn't experience – i did that in my lab roles until i learned better. :melting_face:
Zach: In my experience in pharma, privacy is not the blocker
Zach: Competitive dynamics are
George: One thing we often see in this vein is the “novel target for disease x”. more often than not, if a company is based on the potential value of a novel therapeutic target, and that insight was derived from publicly available data sources, it becomes much easier to get behind that data if there’s wet lab derived mechanistic data in a relevant model to go along with it.
Question 14:
Nicholas: You mentioned competitive dynamics in pharma being challenging. any advice for people selling into biotechs or large pharma companies? or fun (anonymized) stories that you can share?
Zach: Run!
Zach: I mean so many challenges. i would say if your product only works in large pharma, it better be insanely valuable because there's only so many of them and they move slowly (that was the flatiron story).
Zach: If i were investing in a company that sold into life sciences, i would only really consider it if the product was applicable across the entire spectrum (early stage biotech through pharma)
Zach: Otherwise market size is going to destroy you
Zach: That's the challenge, most of these "i'll sell data or xyz software to pharma" startups end up running out of customers quickly.
Zach: One item i want to share before i forget. we're going to open up a referral program that's fairly simple: if you send us a therapeutics company that we end up funding (it's a very small % to just manage expectations), we'll send you $50k in cash. no strings attached.
Josh: What about a bug bounty? that is if someone sends peer reviewed literature showing clear competing evidence against a potential target?
Question 15:
Nicholas: Any advice for future founders or people interested in starting a biotech company?
Zach: ^ the best ideas tend to come from the smartest people. if you want to have a real shot, find ways to work with the smartest people in the industry first (either in academia or in biotech)
George: ^and don’t be afraid to ask those smart people a ton of questions and be open to being wrong or having your mind changed.
Nicholas: Thanks so much to @zach and @george for their time today! it’s been really interesting to hear about the genesis and philosophy of curie and i’m sure this community will overlap heavily with curie’s target companies
George: Thanks everyone for your great questions. don’t hesitate to reach out if we can be helpful.
Zach: Thanks all, ping anytime.
Have a company you are interested in being interviewed? Or have feedback about our #q-and-a blog posts? Ping us on the Bits In Bio Slack channel!