Imagine every time you needed to troubleshoot something in a computer, you had to break out the multimeter. But the multimeter only works for one specific thing, which is an indirect measurement of the phenomena you’re actually interested in and trying to troubleshoot. Also the multimeter has some serious flaws and biases that make it seem like it’s working, but everything you get out of it is wrong and leads you to misinterpreting your results. That’s what its like to do biology.
A lot of innovation and progress is made when new tools to probe what’s going on in biology are invented. For example, cell viability assays used to be conducted by literally counting the number of live cells vs dead cells. Now, you can use a photometric test that’s extremely easy and takes a lot of guesswork out of the assay and makes it routine.
Researchers fundamentally need more of those types of assays to make life easier. And that’s where the dream and the profit people in biotech are chasing after.
https://www.investopedia.com/terms/p/pick-and-shovel-play.as...
What disturbs me sometimes is that many people in software engineering approach debugging like empirical science. Run experiments, gather data, fit some curve, essentially. In the end, a program is a product of the mind. It's pure logic, and fundamentally it's understandable, unlike the real world which can at every turn have new surprises for you. We should think more and measure less.
https://www.vox.com/recode/2019/8/14/20804029/wework-ipo-tec...
https://news.ycombinator.com/item?id=25926188
The problem appears to be that our present text conversation climate is such that you must either pick Team Science or Other Team, and should you mention that there are faked results hiding in plain sight, people get upset because they think perhaps what you're saying is "Maybe vaccines do cause autism" and "Maybe HCQ does cure COVID-19".
But that's not the truth. There is no monolithic Science entity. It is a property of every organization that reward-alignment is a stronger long term mechanism than fact-alignment. And it is impossible to direct the two in the same direction. This means that among the science establishment, there will be those who have ruthlessly optimized for the reward.
This is part of why the general Internet arguer's approach of "There is some evidence that X is Y look at this study url://url.url/url.pdf" is often the blind leading the blind.
Non omnis scientia est scientia.
And some people treat it as a quasi belief system, which makes sense, with precepts like 'the world is deterministic, causes precede effects' and so on. You could do worse. But they aren't scientists and they don't generally grok that science is an activity whose output is sometimes useful models and not 'Truth'.
My mental model here is that you thus get people who trust in science as a paradigm, who then see doubt or skepticism as aimed against the whole edifice that is part of their fundamental understanding of the world, resulting in discomfort and leading to expected results.
In other words pretty normal stuff that resembles when you come across an argument whose implications are opposed to your principles, that you suspect of trickery, but lack the rhetoric to dismantle yourself.
Assume that you consider yourself unable to evaluate the truthfulness of scientific claims.
Being told that claims you'd thought were properly vetted are actually wrong, then means that you have no way to know if other claims are accurate or not.
If your process doesn't work, then any knowledge based on that process is called into question.
Part of the exasperation of "Team Science" is that so many from "Team Other" are not acting in good faith. Most of "Team Other" doesn't actually care about getting to truth--they only care about winning (for some value of that--the leaders grifting and the followers "sticking it" to the other team).
While biotech companies carry a lot of all-or-nothing R&D risk, their “product market fit” is all but guaranteed. Their addressable market size is well known based on patient numbers and how much they can charge nationalized healthcare agencies and insurance companies based on quality of life improvements is well understood. Once they have premarketing approval, they’re all but guaranteed an exit to one of the pharmaceutical giants. The most promising ones get acquired in phase II before efficacy is even 100% demonstrated.
Look no further than IPOs: on HN we lament all the unprofitable public tech companies but the average biotech IPOs pre-revenue! Before it’s even legal for them to sell a product! That’s increasingly been the case since the 80s due to the exhaustion of small molecules and other low hanging fruit.
In practice, scientific due diligence is harder to fake than aspirational user growth numbers which are far easier to game (case in point: Reddits ongoing bullshit)
That's why all those "I tell you about my secret to making 5 grand a day if you click the link below, for free" youtube ads are obviously all scams.
The way computer (and especially internet) companies work is that once you have a working product, you can sell it a practically infinite number of times for practically zero marginal cost.
This means that the successful software/internet companies are INCREDIBLY profitable, since they can sell the same product millions of times.
Are startup investors good at predicting which company is going to be successful? Oh god no, they are awful at it. 90% of the companies they invest in fail completely.
It doesn't matter, though, because they only need one success to make all those failures worth it. The cost of investing in a failure is WAY lower than the cost for NOT investing in a success.
This means they don't really care that they can't evaluate which companies will be successful, they just cast a wide enough net to not miss the successes.
In other words, if this single product the company is working on succeeds, then the unit cost can be brought down so much that it's still a gold mine.
Biotech companies are developing new technology, in a highly regulated environment where, like other comments point out, there's a predictable economic niche. They only share a denomination, the dynamics are wholly different.
You can experiment quickly, and you aren't super susceptible to lock in on any particular vendor that would eat your profits
Making copies is also basically free
Computers mostly are
* well understood at every step (well, at least it was all designed by humans and the bits and bobs do what they say on the tin).
* mostly involve producing some kind of output for everyday normal people to use, in the end.
So they are easy to understand and evaluate. Even some investors can manage it.
Biotech stuff seems to mostly involve making tools for other technicians and scientists. And the process by which is works is some ridiculous, quirk-laden byzantine madness that evolution happened to throw together. At least from an outsider point of view it seems like an absolute nightmare.