00:00welcome to the a 16z podcast today we're
00:03having another of our hallway style
00:04conversations these episodes are based
00:07on videos that are also available on our
00:08YouTube channel youtube.com slash c /a
00:1116z videos what does it mean for biology
00:15to move from the high-risk painstaking
00:17realm of the laboratory bench to the
00:18lower risk get it done world of
00:20engineering bio team general partners
00:22Vijay Pandey and Jorge Conde discussed
00:24this shift including how entrepreneurs
00:26can apply principles from one discipline
00:28to another and how it affects the
00:30healthcare startups code to market hi
00:32I'm Vijay Pandey a drama partner and the
00:33bio Fund hi I'm Jorge Conde also general
00:37partner on the bio Fund and so today
00:38we've got one of my favorite topics to
00:39go into so you know we've seen companies
00:42that are built with science risk and
00:44companies especially tech companies that
00:46are engineering companies you know how
00:48can we take things from the science
00:50curve which is stochastic and high risk
00:52towards something that's more like
00:54engineering more like grind it out get
00:57it done methodical a great deal less
00:59risk yeah it's funny because we talked
01:02about this concept all the time and and
01:04you know reading a children's book to my
01:06kid it was really interesting because
01:08science was defined as this concept you
01:11have a hypothesis and then you test it
01:13and then you revise your hypothesis
01:15that's some real work actually yeah and
01:17then in the night of you know I picked
01:18up an engineering book there at the
01:19merit of science and they defined
01:21engineering as this concept of you
01:22design something then you build it and
01:25then you test it and then you refine it
01:27and refine it and refine it so is that
01:29is that what you have in mind you talk
01:30about these differences yeah I think so
01:32I mean I think there's different ways to
01:34get there so it's not gonna be like a
01:35magic silver bullet so let me just start
01:38with one and and see what you think so
01:39like one things I think we've been
01:41seeing a lot in biology is so what I
01:43would call Legos that you buy I was just
01:45big and complicated you got but you got
01:47a natural hierarchy there you got atoms
01:49you got molecules got proteins you got
01:51membranes you got cells you got tissues
01:53you got organs got people you got
01:55organisms you got ecosystems we can we
01:57keep on going to universes if we need to
01:59you know then the question is can we
02:01find the parts and if you can find the
02:03parts then you can actually maybe mix a
02:05match but do we have an example where
02:07that's already working in biology yeah
02:09the thing with Legos is you can line up
02:12to the circles and can make them click
02:13yep yep you know that's actually really
02:16important part of it so yeah you can
02:17think of Carty is like a simple example
02:18right you know that you've got two
02:20different pieces that you're putting
02:21together you know companies like Asimov
02:23is developing first the Legos and then
02:26putting those legs with the helping you
02:28put the LEGO pieces together and I think
02:29we're gonna see more of that it's hard
02:32though I mean because it's not always
02:33obvious what the Legos are and so that
02:35might be the science part of it but then
02:37once you actually have the Legos then
02:39you get to build stuff almost like um
02:41you know when people build a bridge
02:43they're not researching steel you know
02:45they're given the girders and them and
02:47all the materials and then they put a
02:49bridge together so I think if people can
02:51come up with parts people can engineer
02:53from the parts do you think do you think
02:56that so much of what has driven the
02:59industry and specifically the sort of
03:01the biotech industry we look at
03:02traditional space is so much money and
03:05spend and therefore risk goes into
03:07trying to discover Legos is that yeah I
03:09don't even know if they're thinking
03:10about that way right I mean because
03:11they're just trying to come up with a
03:12small molecule for the disease and maybe
03:14what we're starting to see is like this
03:16shifting and this something you know you
03:17thought a lot about is what is a drug
03:20here and as drugs become cells now the
03:24Lego the Legos become really important
03:26before if drugs are molecules your Legos
03:28are like phenyl rings and things like
03:29that and again so there's not a lot to
03:31do there but like now they have to speak
03:33complexity I think now maybe it starts
03:34becoming more important yeah cuz I I
03:36agree with you I think I think one of
03:37the things that's interesting is if we
03:39weren't if we want to take this lens and
03:40I'd love to get your take on this we
03:42want to take this lens of sort of
03:43science shifting to engineering you know
03:45how do we think about high throughput
03:47biology within that context because one
03:48could make the argument that if you were
03:50gonna do biology at massive scale at
03:53some point do you sort of stumble into
03:55engineering yeah yeah I think point in
03:57so one version of it is actually you
04:00could imagine two extremes or wanted to
04:01say now we're not gonna do that at all
04:02because you don't do massive screening
04:04of bridges right you know you design
04:06your bridge done so on the other hand I
04:09mean in coming up with the parts there's
04:12probably a lot of experimentation and so
04:13I could see roles for that I I think in
04:17this sort of parts analogy I think the
04:19exciting thing is just just once we've
04:21had once we got in there of what we can
04:24in many cases who actually kind of
04:25gotten there yeah I think that's right
04:27yeah and so if we look at engineering
04:29you know one of the neat things about
04:31this space as its applied to my oh and I
04:34know you spent a lot of time thinking
04:35about this and the firm and the fund
04:37increasingly spending more and more time
04:38looking at opportunities here is what
04:40disciplines in engineering are
04:42applicable to biology yeah yeah that's
04:44that's actually something really cool
04:46because I think it's a great point that
04:48you know when you typically think about
04:50mechanical engineering you're thinking
04:51about literally bridges but I think we
04:53can considering electrical engineering
04:55material science computer science all
04:57these disciplines are pushing into
04:59biology and so instead of steel it's
05:01it's it's bone or it's a muscle but the
05:05same principles actually carry over
05:07really nicely and if you look at on the
05:09academic side academic departments and
05:11there are really remaning themselves
05:15into these areas and then the fruit of
05:17that's becoming interesting fodder for
05:19these companies actually if we could
05:21pause on that point for a second how is
05:23academia responding to this world this
05:25sort of shift in the world
05:26yeah the industry will follow presumably
05:28yeah I think you know one of the
05:29interesting things is that you don't
05:31have new departments creative very
05:33frequently like you know physics
05:34departments are thousands of years old
05:36or something like at least the Cambridge
05:38Wood is probably at least a thousand
05:39something like that so you know it's
05:43interesting that they're in there's a
05:44new department that was created in the
05:46last 10 ish years a department of
05:47bioengineering sometimes called
05:49biological engineering and the creation
05:53of these departments I think has really
05:54accelerated this because you could have
05:57sort of inject people that are both
06:00engineers and biologists into
06:01engineering schools and from there I
06:03think a nucleus out and that by
06:05engineers sort of are having so much fun
06:07that I think other disciplines don't
06:09want to be left out and that it kind of
06:11makes it easier to it's a facilitator
06:13straight mechanical engineering ticket
06:15engineered to get into space yeah I you
06:18know it's funny because I think one of
06:19the areas that I've been long interested
06:21in it's been genetic engineering yeah
06:24and if you look at genetic engineering
06:26as a field and as a practice and I want
06:28to get your take on this it was
06:29historically almost like it was genetic
06:32scientists right weren't actually
06:34deciding what was the engineering in
06:35genetic engineering well it was kind of
06:38right if you would just mix up all the
06:39letters and see what worked when a word
06:41actually came out yeah yeah and and now
06:43I think with the advent that things like
06:45crisper and things that companies like
06:46azimoff are doing where they're actually
06:47making DNA it does a design medium yeah
06:50I think it's we're actually starting to
06:51major netic engineering be an engineer
06:53discipline but I think that's a
06:54relatively recent yeah there's an
06:56interesting point that which is
06:58basically just because you put the word
06:59engineering there doesn't think it's an
07:01aspiration more than a success and but
07:04actually just even interesting that we
07:06have long had this aspiration and that
07:08we're starting to see it I mean um you
07:10know we've talked about these
07:11disciplines I think you know we talk
07:12about computer science I think machine
07:14learning itself is becoming its own
07:15thing and it's maybe a final a sort of
07:18third way to sort of connect up with uh
07:20with engineering and that so much of
07:23science when I think about is very
07:24bespoke that you come up with a
07:27biomarker you've got some team of
07:28scientists and spending years from maybe
07:31sometimes decades to find this one
07:32marker and to repeat that would be to
07:34repeat all that process and it's
07:36stochasticity whereas something like
07:38we're she learning is did you sort of
07:39engineer process and then you just put
07:41new data and you roll through it and so
07:43I think that might be a serve yet
07:44another Avenue towards taking something
07:47that's normally in the science world and
07:48shifting it in towards the engineering
07:49world that's actually a really
07:51interesting point because if I if I'm
07:52thinking about this from the perspective
07:54of industry and I'm a you know I'm I'm
07:56at a biotech truck company and I look at
07:59my pipeline of drugs you know
08:00historically at least the the most
08:03advanced program is the most valuable
08:04one because that's the one that's
08:06closest to hitting an inflection point
08:08whether it's proof-of-concept or having
08:10data in a human clinical trial and the
08:12second one is the second most valuable
08:13one and so on and so forth so chronology
08:15determines value and a big part of that
08:17I think was because of this concept of
08:18bespoke Ness yes that you developed this
08:20first drug and it doesn't necessarily
08:22educate you and what you're going to do
08:24with the second drug unless you're going
08:25after the same target or the silo yes
08:27but in your world in the world you're
08:30describing this engineering world it's
08:32actually the reverse is true yes that
08:34the second drug is more valuable in the
08:35first if you're using engineering
08:36principles because what you learn from
08:38example one sort of in views valued an
08:41example to and so on and so yeah and on
08:43top of that you know it's commonly asked
08:45why Pharma doesn't look at their
08:46failures and I think the common answer
08:48is that while they the feeling is that
08:50there would not be value in doing
08:51and that's why they don't put my into
08:53and be very expensive and so on but if
08:55you're in this engineering curve the the
08:56the false positives are actually as
08:58important to learning as the true
09:01positives and so I think both of those
09:03get integrated in which is just a new
09:05way of thinking about things and it
09:06leads to a shift for anything we're
09:07seeing this more and more where farmer
09:08companies will start to view themselves
09:10more as data generating companies and
09:13data science companies um as machine
09:15learning gets in and yeah machine
09:16learning often is like scary with AI and
09:18all this stuff but if you think of it as
09:19just the best statistical use of the
09:21data I mean that's what farm has been
09:23wanting to do and trying to do for four
09:24years if you were to roll forward
09:27pick a number ten years from now yeah do
09:30you think a former company will have as
09:31big a dry lab ie people on computers as
09:34they do a wet lab by people at lab bench
09:36with my pets yes that's a really
09:38interesting thought and question I think
09:40I'm I think we're already seeing a
09:42little bit of that and we're just the
09:43shift to see ah rows where there's like
09:46not a purely medicinal chemist Javal
09:49sort of drug designer job and medicinal
09:50chemists have so much great intuition
09:52and experience design drugs that they've
09:54been sort of the tip of the spear that
09:55but as ml starts being competitive and
09:59and and hopefully really helping them
10:01Ford there may be a new server job which
10:03is not doing the synthesis themselves
10:05but I should just be a drug designer or
10:07a drug engineer throwing that
10:09aspirational term in there and and so
10:12that may be that may be what's happening
10:14especially if synthesis is done at a CRO
10:16somewhere else then that design job that
10:18engineering job actually becomes the
10:19real one and it serves shifts the tools
10:21you look for in the end though the ml
10:24has to work right I mean if it doesn't
10:25work good and this is not gonna then
10:27that future doesn't exist or is it a
10:29different timeline but so it's
10:31interesting cuz it occurs to me as
10:32you're talking that you know a lot of
10:33times when we talk about a company that
10:35has a platform technology yeah
10:36oftentimes that platform technology
10:38really is a combination of some
10:41technology but really a lot of just
10:43know-how and expertise around a specific
10:45area of biology yeah yeah but what
10:48you're describing in terms of this shift
10:49engineering is platforms then can
10:51increasingly become companies or
10:53companies that have platforms I should
10:54say increasingly become companies that
10:56have the ability to move drug discovery
10:59development from being a very bespoke
11:01thing yep to being a very sort of
11:05that's a good point like ml is a
11:06platform amongst others and then there's
11:09the other generating platforms there's a
11:11analysis platforms and at that platform
11:13I think really could be a really
11:15interesting shift and we're seeing more
11:16and more companies that way that are you
11:18know before it used to be that the
11:19platform was never really valued very
11:21much it's all about the assets and now
11:22we're getting to the point where people
11:23are very curious if platforms can
11:25reproducibly create multiple assets let
11:29me ask you an unfair off-the-cuff
11:30question you've looked you've seen a lot
11:32of things what's the most surprising
11:34application of engineering in biology
11:36that you've come across whether it's an
11:38academia or whether it's here on the
11:40investing side yeah you know there's
11:43it's hard to tell because there's so
11:44many different surprises like you know
11:45three one I always love to talk about is
11:47like where we talked about this at that
11:48tree that glows at night you know that
11:50lucifer's tree or something like that
11:52and the idea that that reason why I find
11:54that so compelling is that it's a
11:56combination of sort of the the technical
11:59engineering and inside to make it happen
12:01but also this idea that the future will
12:03not be steel and metal it will be this
12:07sort of engineered biological thing that
12:09just grows that has this function
12:10something that actually helps fight
12:13against global warming not contribute to
12:14it something that a really sustainable
12:16I'm easily shippable because you ship
12:18them in little guys and then they grow
12:20or something like that it's just all of
12:22it is just a very different vision for
12:24what our world will be like and it's
12:27something that I'm it's just the
12:29beginning and maybe just starts with one
12:30tree but then you know there's lots of
12:31different things no I I it's funny you
12:33say that because I think one of the
12:35things that's so fun about this concept
12:36of engineered biology is that you know
12:39at least in my sort of limited view is
12:41you know historically when you think
12:42about engineering it's making things
12:43better yeah making things more efficient
12:46yep making things more quickly but when
12:49you have biology as a design medium it's
12:52about making things possible that you
12:54didn't even know we're possible yes like
12:55glowing trees yes well we should
12:56probably talk about a little bit about
12:58nuts and bolts for how to get things
12:59done you know I think you know in in you
13:03know because it's nice to have the
13:04philosophy but like how do you actually
13:05do it and you know towards that I think
13:08when I think about it like my favorite
13:10paradigm is something like the Apollo
13:12mission so President Kennedy says we're
13:14going to the moon everyone the engineers
13:16we're going to moon you know I can only
13:18imagine that moment because that sounds
13:20crazy right I mean to do something like
13:22that and how you take something crazy
13:24like going the moon or like curing
13:26cancer or increasing long human
13:28longevity by 50% how do you do that and
13:31the Apollo mission actually did Apollo
13:33mission was even just the last one you
13:34have mercury and Gemini then Apollo uh
13:36uh you know as a kid I really loved
13:38space and just you know Powell Evan got
13:41to the moon so what was like one through
13:42ten you know so one gets into orbit to
13:45actually forget all the details but they
13:47have to practice docking in space after
13:48practice all these things actually went
13:50around the moon before landing on the
13:51moon and if you break it up into little
13:54any little bit isn't so bad and and can
13:57be engineered and then you sort of do it
13:59step by step by step by step and I think
14:02that's for me the inspiration for how to
14:03take some big crazy thing at like going
14:06to the moon that if you did it from like
14:08a screening perspective the screen
14:09rockets they have to get over there
14:11that's good you know one a million
14:13chance but if you engineer little bits
14:15bits by bits with you know
14:18okay ours you know your your key
14:20milestones with with kpi's your key
14:22metrics I can see how that starts to
14:24make a little more sense but you know
14:26it's still easier said than done but I
14:27think now it takes a big crazy thing
14:29until like lots of not so bad things and
14:32that's important because that's where
14:33your lego concept comes into play
14:35because you couldn't you couldn't build
14:36a rudimentary rocket if you didn't know
14:38yeah where to put the screws yep yep
14:40literally quite literally pull it
14:42together yeah that's a good point yeah
14:43these these what they learn from from
14:45you know the previous mercury in Gemini
14:46was to get the pieces and then they
14:49started mixing them together and even
14:50look at like the Soyuz Rockets they're
14:51like that's the Rockets put together and
14:53so on so you know my mind that is one of
14:55those hallmarks sort of just landmark
14:58like wow moments in human engineering
15:01and I'm always curious to see how we can
15:03learn from that process to bring it over
15:05you know what that said there was a lot
15:07of people a lot of effort in there and
15:08so I think this is again not something
15:10that will be done by one company or one
15:12group but you know I think collectively
15:14the ecosystem strengths be built so just
15:17to put this from the perspective of
15:19Industry and from entrepreneurship how
15:22does entrepreneur a get value for great
15:25value for creating Apollo one yeah yes
15:27actually that that's a really tough one
15:30I think I think there has to be this
15:31ecosystem from academia and research
15:34institutions to startups to big startups
15:37to big companies and I think you know in
15:41Asimov's case you know all the parts
15:42they came up with in academia for e.coli
15:45and then now they start to look at
15:47creating parts for other things they say
15:49they create the process mmm and now they
15:51could repeat the process for other types
15:52of cells I think that might be part of
15:54is that they got to like at least to
15:56let's say the Apollo missions maybe they
15:59did Gemini you know at MIT so something
16:02where you can get part of the way there
16:03we're you know I think you and I can
16:06have some sense that oh we know where
16:07this is going and that this is now
16:09engineering not like kids playing with
16:11rockets and in the backyard you know
16:13where maybe was gonna blow up or not and
16:14who knows it's gonna work
16:15something where we had the feeling like
16:17it's it's now receive the steps it will
16:20also be interesting cuz I think you know
16:22there's lots of big incumbent companies
16:23Google and IBM doing all this research
16:26and they bring an engineering mentality
16:28and engineering sort of zeal and
16:31excitement to other areas and so I
16:33suspect we'll start to see that I mean
16:34okay ours are like I think part parcel
16:37of how Google runs and I would think
16:38that would come over yeah it's a good
16:40point because III think for for
16:42companies that are trying to innovate in
16:44this space and using engineering as a
16:45way to do so in bio the big challenge
16:49from a business development perspective
16:50has historically been proved to me
16:53you know if I'm the buyer about yea
16:54large company proved to me that what
16:56you're doing is real yes and it's going
16:59to work well yeah so so how does that
17:00work I mean how would you answer the
17:02question I mean what do they need to do
17:04what would the buyer need to see well so
17:05I think I think what's interesting and
17:07we can take specific cases azimoff is a
17:10great example where to use them one of
17:11the beauties of what Asimov does is
17:13there's high predictability in what they
17:15design is going to work yeah and so if I
17:19you know if I'm on the on the receiving
17:21end of looking at Asimov's technology as
17:23a potential partner you know business
17:25development collaborator if I say you
17:28know can you design something to do X
17:30and it's significant percentage of the
17:32time it does as intended yeah that's
17:35very different than historical biology I
17:37think one of the big things in synthetic
17:38biology has a store has really been you
17:40try ten thousand things together
17:42and if somehow that ratio is you know
17:44nearly inverted then all of a sudden it
17:46becomes a much more it's much easier to
17:48get people to believe that you
17:50understand where the screws need to go
17:51if you understand where the screws need
17:53to go then as you pointed out it starts
17:55to become clearer that you'll eventually
17:56get to the moon yeah so it sounds like
17:58you're saying that if you can
17:59demonstrate that you've sort of shifted
18:01on to this engineering curve that would
18:03be sort of the key part for the
18:04go-to-market is it you I think it's
18:06proving that something works as intended
18:09yeah he's historically been the big sort
18:12of activation energy you need to
18:14overcome for a buyer and the business
18:17development context and so when you're
18:19doing the sarcastic risk that you're
18:21that means there's a long period of time
18:23that has to go until you get this proof
18:24of code so what does that proof look
18:26like presuming it's more than just like
18:27a paper and science and nature or
18:29something like that no it's usually a
18:30I'm gonna show you X you know step wide
18:33stepwise improvement over a short period
18:36of time or at least over a predicted
18:38period of time that what I can do works
18:40yeah and then it's not a one-off that is
18:42reproducible yeah and that it works in
18:43various contexts cuz the various context
18:45usually represents some sort of a
18:46fundamental if not universal truth yeah
18:48so that's proof of concept that's like a
18:51small deal yeah so normally the way it
18:54plays out it's it's it's it's usually a
18:56proof of concept that leads to a small
18:58deal yeah and a lot of times I think one
19:01of the challenges that early stage
19:02companies have is you don't want to run
19:03into this risk you die of die from
19:05pilots yeah yeah right because I think
19:07his stories either either ish together
19:09it's easy to get into a pilot it's hard
19:12to go from a pilot of an actual
19:13agreement yeah and so this is where it's
19:16certainly the traditional biotech
19:17companies have had a very big challenge
19:19right because every project is a science
19:20project and they're long and you don't
19:22know how they're gonna turn out if
19:24you're doing something that's more on
19:25this engineering curve you actually can
19:27use in your concept of okay okay ours in
19:29kpi's you can actually say that this is
19:31going to be the project plan and this is
19:33what we're gonna deliver over a
19:35relatively short period of time because
19:36we can essentially iterate quickly yeah
19:38and if you can do that
19:40then that leads to an initial project
19:41and one of the things that certainly has
19:43been shown in tech time and time again
19:45is if you can demonstrate value at a
19:47small scale and prove that you can
19:48actually scale then you have this you
19:50know the concept of the famous concepts
19:52of land and expand on the enterprise
19:53side you can actually start to see some
19:55that in biology the reason why land and
19:56expand historically hasn't existed and
19:58biotech yeah it's because the expand
20:00part was really hard yeah things were so
20:01bespoke well yeah you get your one drug
20:03and then you go go to town on that one
20:06and then a platform was useful for
20:08exactly but then like you're there
20:10exactly so so how do we focus on the
20:12drug asset yes how does it change I mean
20:13I think it changes to this concept that
20:15you know asset number one is actually
20:18less valuable than asset number two
20:20which is less valuable asset number
20:21three because you're learning from each
20:23one yeah to the next one so you're
20:25getting better over time because you're
20:26moving from bespoke to the end you're
20:28getting faster you're getting like
20:30higher efficacy lower talks or all of
20:32those things or presumably yeah so
20:34you're getting better faster and cheaper
20:35I mean that's that's obviously the the
20:37sweet spot yeah or you're in an ideal
20:39world you're going from impossible to
20:40possible yeah um but that obviously so
20:43there's never that use doesn't happen in
20:44one step that's right exactly yeah yeah
20:46but I do think that's interesting that
20:48when we see companies that have very
20:50powerful platforms and there have been
20:51analogies for this in biotech the ones
20:54that have very powerful platforms that
20:56can improve scale ibly and
20:58systematically on sort of an engineering
20:59like curve tend to go from you know hey
21:03they take something that isn't possible
21:04to possible over a longer period of time
21:05yes and they tend to sort of grow and
21:08gain traction very predictably because
21:10they can improve on a very predictable
21:12curve and I know I you know we use this
21:14example all the time but it's what
21:16Illumina did for sequencing yep right
21:18and there are other examples where that
21:19has happened where you've shown that you
21:21can demonstrate something in one context
21:22you demonstrate it in a second context
21:24and then by the time you get to the
21:26third people just accept it as a Jenny
21:27EMA is that famous quarry it's a tribute
21:29to the most powerful forcing universes
21:30compounding interest and and that's what
21:36this is right it's not money compounding
21:38it's technology like getting 30% better
21:40every year is or 30-ish percent better
21:43is like doubling every two years it's
21:44exact doubling every two years that's
21:46like we start with children's books I
21:47well my favorite kids books is that
21:49Indian story about the grains of rice
21:51where like as as as a gift or as a
21:55reward the peasant asks for two then
21:59four and doubling every day over a month
22:01and the Raja thinks you know this is not
22:03that big of a deal right two grains for
22:05grains is not a big deal of course in
22:06the end it gets to two to the 32
22:08four million grains of rice which was
22:09like all of it and it just it sneaks up
22:12on you if you can get to that that is
22:15how you make impossible possible that's