Hacker News Reader

Comment by biohax2015 | Hacker News Reader

biohax2015

Makes me regret abandoning biotech/computational biology for the safer shores of software engineering :(

I'm a bioinformatics scientist. Salaries are notoriously horrible in this field and I get paid better than most wet lab scientists who have more education and experience than me.

I think biotech may be exacerbating inequality more than most industries, as the profits are most definitely not going to the employees like they do in other STEM fields.

You likely made a 2x salary decision as long as you’re a ground level employee in either field.

I would think software engineers are still very much needed in almost every biotech startup :)

There are some that rely very heavily on software engineers but most don't, although that is starting to change I think. The biotech startups that the tech press covers represent a tiny fraction (under 10%) of biotech startup activity. The separation between the biotech and tech startup worlds is almost at a global maximum. It used to be that major tech VCs were also major biotech VCs. Starting about 5-10 years ago tech VCs gave up on biotech. Coincidentally that's about when all the exciting stuff started happening

Pretty much all of the big biotech exits are mostly not even covered in the tech press (with the exception of stemcentrx, unity and maybe a few others)

AI/ML seem to be leading to some new companies that have a big software component. But in general the “tech” crowd doesn’t think much about pharmaceutical, biotech, personalized medicine etc.

Biotech requires several very different foundational assumptions than consumer-tech. You can't "move fast and break things" in biotech. You have to budget a lot more for R&D, and you have to be okay with dozens of experiments outright not working and not leading to anything useful. Having world-class scientific talent on your team is actually really important, while it's overkill for most software. Your iterations are usually about whether the science works, not whether the market will accept it. Technical risk dominates market risk. There's a large regulatory barrier to even getting a product to market. Your per-unit profits are often significantly higher than in consumer software, but your per-unit costs are higher, and your unit volumes are much lower.

[source: I'm in consumer tech and my brother-in-law is a cancer researcher.]

A VC's core competency is typically in knowing & applying the core industry dynamics to their investments, so if the core industry dynamics are different, it makes sense that VCs will eventually specialize in subjects that they actually know something about. Trying to apply consumer software "best practices" to biotech will get people killed and you sent to prison.

I agree, i think the biggest issue right now with tech VCs starting to get back into biotech is that they don't appreciate the technical / scientific risk, or think they can magic the risk away with AI. The tech VCs are renewing their interest bc they've seen that its possible to get $10B+ exits in under 5 years in biotech (KITE and JUNO got to $10B valuations faster than I think dropbox, snapchat and twitter), but i think some of them misunderstand the risk, as you mention, and will get burned

Cycles times are costly / more expensive, and companies usually only get a handful of chances to get it right. And it's too expensive to try to manage this risk through diversification alone -- cost of failure is too high. That's why biotech VCs spend so much time on diligence -- mostly scientific diligence -- compared to tech counterparts

Somewhere in the multiverse, all software is developed with the care that biotech requires, and there's much less of it, but people don't have to put up with crap.

We would still be in the mid-1980s software-wise if that was the case.

Once upon a time, software was developed with that care. Read up on the Multics security model, or Spark ADA, or the multiply-redundant systems that Tandem and Stratus put out. It really wasn't until the microcomputer era that software became "test it until it works" rather than "rigorously prove that it meets the specification".

And that was absolutely the right move, economically, because for most of the stuff we use consumer software for, it's really not that important. If Reddit shows the wrong point total for a story, nobody really cares. If YouTube occasionally hiccups while playing back, it's not a big deal. I'd rather have YouTube at all rather than insist that they get all the bugs out.

Someday, probably, people will insist on ironing out all the kinks in software, the same way we consider plane crashes unacceptable now (if you read the history of aviation, there was a time period where virtually all the key aviation innovators died in plane crashes or accidents). That period comes after all the major markets have been discovered and filled, though. Once computers have spread into every niche that they can, with clear market leaders for each industry, we'll see them devote their energy toward making things bug-free.

they are, but they're usually considered "IT" It's a secondary role compared to science. The computational action is for math, stats, and informatics specialists (like computational genomics).

This often translates into lower status and lower pay.

You really need some chemistry and biology to get love in a biotech firm.

Reply on news.ycombinator.com