Hacker News Reader

Comment by jasfi | Hacker News Reader

This stood out for me: "29% of the batch were accepted with only an idea". Kind of amazing!

I'm seriously thinking of submitting my project: https://lxagi.com. I've written a lot of code, but I'm not quite sure of when/if to apply.

View on HN

Well, I guarantee you those idea-only founders either have some pedigree (Stanford, MIT, Facebook, etc.) or have started successful startups before.

I bet you are right. Would love to see some kind of metric on this.

Read something recently that many serial entrepreneurs seek VC funding rather than fund the effort themselves (even when they have the cash to do so). Often the connections provided by a VC are just as important.

Geoff from YC and Patrick from Stripe did a chat with an Irish audience late last week. One thing that resonated with me was the fact YC provided encouragement along the way when the Stripe founders were unsure the venture would work.

My startup (courier.com • YC S19) was idea-stage and I did not have meaningful pedigree nor had I previously started a successful startup.

the deadline is gone, but I also applied after the deadline and got in.

you should apply now!

Thanks, I'll see if I can make it before the final deadline. I might be more organized by the next batch though.

Just do it - it's a great exercise and as the stats show people usually take more than one shot to get in.

I probably will. You're right, it might be worth it even if only to better my chances next time.

it'd be interesting to know what kind of companies those are. Don't want to give people a false sense of hope with "Uber for x"

could you explain what your project is? I didn't understand it from the splash page.

looks like a named entity recognition tool. Send some language and it parses it into actionable tokens

I'm working on a better description. The end goal is human-level conversation and related use cases, but that is too far off. The 1st real use case is that you can give it sentences and it parses them to JSON that can be used to call APIs. Those API calls provide the answers.

Hmm ok, so am I right in thinking that it would be used for natural language API integrations? Like a search-based interface or an Alexa Skills type thing? Interesting nonetheless - I'm unusually out of touch with the industry development of ML, despite it being extremely adjacent to my interests, so sometimes these ideas don't make immediate sense to me. Good luck in your endeavours, regardless!

Yes, that's the idea.

Reply on news.ycombinator.com