I don't understand the point of this. Is there a cultural reference that I'm missing? At first I thought that it was AI robots talking to each other but it is actually a dead simple script which reads a text file and pass it to a text-to-speech command. So clearly it does not interest HN for the technical aspect. I feel I'm missing something.

Even I thought it was something related to AI. But looks like it is just a simple text to speech utility.

It uses a text-to-speech utility, it is not even implementing that.

I had to do a double-take on the source code, as it looked just like Ruby. It's the first Crystal code (https://crystal-lang.org/) I've ever seen, and the interesting part is that it compiles Ruby-like code into executable. Anyway, the code can be summed up as "Process.run("say -v #{voice} \"#{text}\"", shell: true)", in which it calls "say" command to speak the voice of text file in macOS terminal.

PS: Does anybody know if the goal of Crystal is to be like Go, and if there's any performance advantage using Crystal, compared to MRI?

Oh, yeah. Back to the future.

Does anybody besides me remember the psychoanalyze-pinhead feature in venerable versions of Emacs?

There was a script that offered a parody of Rogerian psychotherapy. For example, you could type, "I'm anxious" into it, and it would respond "Do you feel anxious because of something that happened in your childhood."

And there was a script that blurted random Zippy-the-pinhead quotes, like

"Are we having fun yet? Are we? Are we?"

"Life is a blur of Republicans and meat."

The psychoanalyze-pinhead script connected these two together. So you'd get stuff like

"Tell me more about a blur of Republicans and meat and your mother."

Emacs was pretty fast. You could get a few hundred kb of absurdity in a few seconds.

Emacs Doctor ( which was a version of Eliza [0]) is actually more sophisticated than this script.

https://en.wikipedia.org/wiki/ELIZA

Maybe i don't know the reference. Anyone care to clarify?

As far as I can see it's just a txt file read with different voices...

You clearly didn't actually look at the link at all. It is just a script that reads a text file and outputs it using a text-to-speech library...

You're right and I made a fool of myself. I thought it was like seebotschat on Twitch which was actually two real bots talking to each other (and amazing that is).

If it's half as awesome as twitch.tv/seebotschat was it's amazing.

I don't know anything about that Twitch link, but I doubt it. This one being a disingenuous use of the word 'robots'.

Everybody that's upvoted on HN or starred on Github (over 30 on each!) has clearly not followed and read the link.

It appears to me like OP's experiment in marketing - getting HN readers to react positively to something mundane.

Well, ok. The Twitch thing was actually two bots talking to each other (two Google Homes hooked up to Cleverbot or a fork). It resulted in some really hilarious and sometimes deep conversations. I watched for hours.

care to share an audio sample for those of us not owning an apple machine?

I'm curious about Navi's voice.

eh, I thought the project created a new voice based on navi. I didn't know that serial experiment lain used a MacOS voice. Here it is: https://youtu.be/Z4hqUxb9MmY?t=3m11s

Thanks for sharing!

Meh, I've done better stuff with Hailo from CPAN and 20 lines of Perl.

Please don't post snarky dismissals of other people's work on HN. That's especially against the rules of Show HN threads: https://news.ycombinator.com/showhn.html

Edit: you've repeatedly posted uncivil comments to HN. We ban accounts that do that, so please stop doing that.

This seems important

It's the opposite of important, thus it might be