Conferences to show off new startups are a good thing, but it seems like a bad plan to make the startups launch at the conference. Unless the date happens to be exactly when the startup would have launched anyway, this constraint is going to force them to launch either too early, or too late.
Launching late is very dangerous. Someone else is probably doing what you're doing, and if you launch even a couple weeks later your company could for the rest of its life be described as an X-like site, where X is your competitor's name.
Launching too early is even more dangerous. The right time to launch is as soon as you have something (a) useful and (b) not totally broken. Which means that launching early = launching something useless or broken.
Also, I read somewhere that launching at a conference means that you're one of dozens of launches over a short period of time, and you have to compete even harder for coverage from press/bloggers. Whereas, if you launch on a non-descript day, you may be the only interesting thing happening all day, and get a copious amount of coveage.
The alternative view is that having a hard launch date acts as a forcing factor to get your code up to shape. From personal experience, I found that having committed to giving a demo has made me work that much harder to finish it. Otherwise, you can keep on pushing back your launch date.
For what it's worth, last year, a handful of companies didn't end up launching on the day(s) of the conference (probably because they weren't ready vs strategy)
Launching late is very dangerous. Someone else is probably doing what you're doing, and if you launch even a couple weeks later your company could for the rest of its life be described as an X-like site, where X is your competitor's name.
Launching too early is even more dangerous. The right time to launch is as soon as you have something (a) useful and (b) not totally broken. Which means that launching early = launching something useless or broken.