That said, great research projects do not necessarily translate into great startups. People doing fundamental research with no clear path to market likely aren't a great fit for the program at this time.
There are some aspects of biology that likely translate better to the tech model. But for most aspects, especially anything to do with therapeutics, it is essentially impossible to avoid the hard slog. No software is going to change how you study efficacy and toxicity in animal models and human testing. With more targeted therapies the complexity only goes up.
I think YC should invest in the life sciences, but it should be a different model, perhaps an innovation on the existing tech transfer model, perhaps building an incubator that ends up being an IP funnel for pharma, or to provide some competition in an underserved area, but the software mindset is generally going to fail.
In biotech, once the technology has been developed, the innovative component has largely been completed. And it's the top scientists that have the institutional knowledge and non-dilutive funding to make that happen. The reason there are so many unmet needs in healthcare isn't for lack of trying or a lack of original thought in the top academic research labs - you can be sure that Bob Langer isn't just trying to up his h-index.
That said, you are right that a recent graduate won't be able to found a company in that model. That's mostly irrelevant though, since recent graduates are rarely capable of creating scientific breakthroughs on their own.
Except for the minor problem that any IP generated through their work is owned by the university. And if the IP's valuable, it typically flows straight into the hands of Flagship, Polaris, Third Rock, & co. The professor might have a say in what happens to the IP, but the students and postdocs have zero leverage with the TTO. And how could they? They are replaceable cogs in the machine (and often in the US on a student/work visa) - Bob Langer is not.
And my original point was that Paul Buchheit's request that "If you know great scientists solving important problems, please encourage them to apply" was unrealistic. YC isn't going to get the crème de la crème of scientists or scientific innovations - they're getting what's discarded by the TTOs.
To be more precise, nobody's going to get the top scientists (they're going to go right back to their labs) and the biotech VCs are going to get the top innovations (for which they often have an informal licensing option with the inventing professor before the academic research even begins). There's a record amount of funding in biotech right now, but why spread it evenly? Might as well just run dozens of trials in parallel for therapies derived from a single technology.
Some Universities don't automatically claim ownership of any IP you create there.