[nloui]

Parrot.vc is a bot trained on tens of thousands of VC tweets that uses predictive text to generate amazing, new startup advice. Share your favorites!

Gavin Belson - hit me up, this is the perfect acquisition for Hooli.

PS - you can follow Parrot.vc on Twitter as well (https://twitter.com/parrot_vc)

[smacktoward]

Is the source code available anywhere? I’ve been mulling over a similar project but don’t have a lot of ML experience myself, would be great to look at the innards of a working example...

[pequalsnp]

An easy way to get started that doesn't use pure ML is Markovify [1]. I used it for a couple of projects last year.

[1] - https://github.com/jsvine/markovify

[rpvnwnkl]

This blog post and associated github give a good enough intro to get simple examples working: http://karpathy.github.io/2015/05/21/rnn-effectiveness/

[nloui]

I may open source it. It's pretty easy to get started if you leverage pre-existing libraries for Markov chains and Twitter ingestion.

[tlrobinson]

Now do it with Hacker News comments!

[umvi]

"In other news, water is wet"

"This has been known for 10 years, this is not a new discovery"

"Title is clickbait, can someone please update it?"

[jraph]

It exists. :-)

hncynic: https://news.ycombinator.com/item?id=19694578

[zhte415]

This is the sort of thing that I'm hoping to see, especially with the recent "Show HN".