DL Software co-founder Martin Shkreli and founder Brian Armstrong argue tech systematically underestimates biological and medical complexity

Agree - everything in hard tech is more difficult than software (especially in bio where we barely understand how it works)

Arguably the best way to do a biotech company is: 1. make a billion dollars in software 2. self-fund the biotech (at least in part, to de-risk it)

If you don't have a software billionaire patron, the uncanny valley (series C crunch) can be brutal in bio.

It's still worth attempting hard things that can move society forward even if difficult tho. Arguably you have a responsibility to do so once you've made money in software.

I'd also say most biotechs are not as well run (from what I can tell). And they could learn a lot from the software industry.

i.e. techbio > biotech

And the 1-2 orders of magnitude complexity will get tackled with AI, not human brains.

i.e. bio-naive > bio-insight driven

@rheum_ai ... and clinical/healthcare people underestimate the impact that technology will have on the field(s) in the coming years. It's a two way street.

@rheum_ai You need a mix of physics, software/electrical engineering, chemistry, and medical knowledge to make good estimations. Tech people often overestimate their grasp of problem spaces.

@BurnZeZ Maybe the nuanced take is that estimation is hard for anyone and that we're all just bluffing until we get more data.

I will clarify that I agree with the people saying this is a good thing. Crippling pragmatism is my greatest weakness and I wish I had Stanford-dropout levels of unbridled optimism and audacity.

@rheum_ai Underestimating is good. You don't get scared off early by the complexity and by the time you figure out it is much more difficult, you are most likely already committed and will try to see it through.

@rheum_ai "Complexity" is a very underspecified term! I don't think it's clear that medical/biological problems are complex in the sense that the class of functions that fit the data are esoteric, it's just that each individual piece of data has very limited predictive value

@WillowChem The amount of data, time, and/or money required to solve.

@rheum_ai Complexity as measured by what?

@brian_armstrong 3) open r&d HQ in RTP 4) join nc biotech center 5) profit

@brian_armstrong Hard tech != biotech. Notably it is not something you discover not engineer your way out of, which may be uncomfortable for people from an engineering background

@brian_armstrong My plan was always make a billion in software then self-fund biotech. I am currently stuck on step one, going on forty years now. 🤠

@DoeSparkle So how can we save these mythical millions of lives if you can’t even articulate the problem(s)? This is fucking stupid.

@rheum_ai @MartinShkreli You are underestimating that computation can reduce the problem complexity in biology by 1 to 2 orders of magnitude. Stop thinking computation in 2026, start thinking about the computation scale and sophistication in 2040 to see. Intelligence is compression. Compute’s intelligent

@kemmishTree You had me at the 6th exclamation mark.

The problem is time-series scanning is denied over cost issues despite it being the best way to stop the so-called "over diagnosis" problem by getting baseline scans and showing changes over time. Its not really that complicated.

Also if you want a 21st century example: The 20-year lag on algorithmic drug-interaction cross-checks because doctors are dumb boomers.

@rheum_ai I would say 3-4 OOMs. And also, WE MUST BUILD UTILITY-SCALE MOLECULAR SENSING!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!OK?

@rheum_ai 19th century doctors called germ theory bullshit because they didn't want to be responsible for all the deaths they had caused. They got just as defensive when bloodletting was found to be worthless. And then there is the fact that they denied scurvy was nutritional.