[Areibman]

In my experience, hackathons don't scale. The value is lost when the resources are spread too thinly across hundreds or thousands of attendees. Mentors can't teach, and hackers can't get feedback. It becomes impossible for meaningful discussions to arise, and winners start optimizing for all the wrong things (i.e. fancy slideshows, needlessly complex designs, worthless ideas) because nobody ends up solving real problems. Worse yet, the hosts celebrate all the wrong metrics such as attendee count and lines of code written, as if those mean anything.

As a prospective YC applicant, I see every batch size increase as a dilution to the value of the incubator. Most of YC's content is already free online; the key value is the personalized insights from partners, access to a professional network, and co-experiencing starting a company with other early stage founders. These sorts of things scale linearly, at best.

[tlb]

In almost everyone's experience, almost everything doesn't scale. Figuring out how to make something scale is a rare event.

I say this having heard, while debating the potential of various startups, probably 10000 arguments for or against something being able to scale. In the end, no amount of wisdom makes you very good at predicting it. Everything doesn't scale until someone figures out how to make it scale. So you bet on smart, energetic founders to figure it out.

Often the scalable version looks different than the initial version. Maybe the million-person hackathon looks more like Repl.it than a room with laptops and teetering stacks of half-empty pizza boxes.