Ask HN: What is the reason YC switched to 4 batches?

I can't help myself but think that some fundamentals changed over the past 2-3 years in the software space.

To me it seems increasing throughput has long been a goal. For a batch processes, the options are larger batches or more frequent batches. YC has tried increasing batch size and later scaled it back (I have no idea where the current batch size sits in relation to the largest batch, they might be the same or different).

In other words, I suspect the change probably reflects changes at YC more than anything else.

This actually makes totally sense and most plausible.

https://www.ycombinator.com/launches

Also they can process applications faster now!

Scale? Given the demand for batches and the rate of successes they have PMF sufficient to justify more investments.

Wouldn't the investments then also spread thinner?

YC has substantial capital to deploy, including capital from interested alumni.

https://www.axios.com/2024/04/06/ycombinator-funds-alumi-inv...

They invest relatively little compared to the successes, to the point that they're even using SAFEs to double down. The thing is YC can't invest more; too much money is bad too for early stage. So it makes sense to invest into company with lower odds. They've also said that some of their best outcomes were just slightly above the bar, so there's probably a lot of potential unicorns slightly below the bar as well.

Agreed, the TLDR is they want as many cheap dice rolls as possible, and can afford it.

> I can't help myself but think that some fundamentals changed over the past 2-3 years in the software space.

LLMs enabled many new companies & created demand for more products/startups.

I'm not sure how LLM makes it easier to find PM/F.

Finding a real PM/F is like winning the lottery (I'm not talking about low TAM B2B products or services here).

In order to find one good idea, one need try lots of bad ideas. So, do you mean that LLMs enable to test and throw away bad ideas faster?