I've been following various ETH and Sol hackathons for years, and to be honest—very few projects actually survive.

Most of the demos at hackathons are just flashes in the pan, and as for those that really take off? Almost never seen it happen. Projects coming out of incubators, on the other hand, occasionally manage to make a splash, but that's not because their demos are flashy—it all comes down to whether the team can quickly test and iterate like crazy.

Another key factor—the quality of the mentors. There's a huge difference between mentors who truly get involved, go the extra mile on details, and those who are just there for show. So when we evaluate projects, on-site performance actually carries the most weight, since it directly reveals the team's execution and adaptability.