"Does it actually work? Kind of."

Iff "actually works" means "lorem ipsum generator." It generates something that looks superficially like programming answers, but makes no sense whatsoever; it seems to be on par with a simple Markov chain.

Example from "top 1000":

> What is console.log and how do I use it?

> In my configuration model I set up my InFile active output file: ./DevFine/mysite/ .py

In other words, AI cargo cult. "It sorta kinda works <small>(except for the one problem that it's supposed to solve. That part doesn't work, but all the bells and whistles do.)</small> Perhaps if we can make it look even more like SO, it will magically start working. Any minute now."

Hey everyone! Happy to announce the culmination of about a month of digging into ML and NLP.

Ask Roboflow is trained on millions of Stack Overflow question/answer pairs to mimic human responses.

It's learned lots of interesting things including how to insert HTML links and images into its replies, the syntax of several programming languages, and how to link to "relevant" documentation.

One thing it hasn't yet learned is the concept of "correctness" so most of the answers you'll see won't actually be helpful yet... I plan to continue to improve the model as time goes on. Hopefully one day it will actually be able to help new programmers get instant answers to their programming questions.

I just looked at the first most recent https://ask.roboflow.ai/question/30954728 and the answer is total nonsense. Left unimpressed.

I get a http://localhost:5000/question/4743730 when I click on a top question

Aw shoot. Fixing! I "cached" the top questions before deploying and forgot to update the URLs.

Edit: fixed!

Looks cool, but after the 5th refresh, and whitelisting even more botnet domains, I had to give up testing it.

Not sure I understand. The only external dependencies are cdnjs, Firebase, and google fonts.