[michjedi]

Current top 10:

1. "Apple introduces 16-inch MacBook Pro, the world’s best pro notebook" Bad: 0.9964 - Good: 0.0038

2. "Developing open-source FPGA tools" Bad: 0.3381 - Good: 0.6652

3. "Show HN: Can a neural network predict if your HN post title will get up votes?" Bad: 0.0598 - Good: 0.9307

4. "How internet ads work" Bad: 1.0000 - Good: 0.0000

5. "More Intel speculative execution vulnerabilities" Bad: 0.7413 - Good: 0.2306

6. "OpenSwiftUI – An Open Source Re-Implementation of SwiftUI" Bad: 0.9994 - Good: 0.0005

7. "How VCs Make Money" Bad: 0.9997 - Good: 0.0003

8. "OpenBSD: Why and How (2016)" Bad: 0.9988 - Good: 0.0013

9. "The Perl Master Plan: How to Put Perl Back on Top" Bad: 0.9997 - Good: 0.0003

10. "Jerry (YC S17) Is Hiring Senior Software Developers (Toronto)" Bad: 0.3142 - Good: 0.6800

So all in all, only 3 of today's top 10 had good titles... Either the titles could have been better but the content was too interesting, or this tool has very low recall.

[Reason077]

"Show HN" Bad: 0.0002 - Good: 0.9998

"Warning: bad economist" Bad: 0.0001 - Good: 0.9999

"Warning: bad artificial intelligence" Bad: 1.0000 - Good: 0.0000

[sdenton4]

Seems like the judge has a small conflict of interest.

[jansan]

So the answer "Can a neural network predict if your HN post title will get up votes?" is a clear "no", at least for this tool.

[skykooler]

The only one it predicts well...is itself.

[SamBam]

Perhaps because the author had access to the tool before writing the title? In which case, less of a "prediction."

[faizshah]

"My YC app: Dropbox - Throw away your USB drive" (Bad: 0.9970 - Good: 0.0029)

[Avamander]

I think this is a prime example of where AI could go wrong. When people just talk about social media AI curation they don't really understand it. But I personally really wish social media would do less AI curation, who knows what gems we've missed, just because they're maximising for our instant satisfication.

Kinda spooky even, who knows, social media totally might have already killed companies that sounded too different or even just political ideas that differ from mainstream (or sponsored) views?

[segfaultbuserr]

I think the problem is that the title is not a good indicator for current-event related submissions. "More Intel speculative execution vulnerabilities" may be a bad blogpost, but it's an important current event, so it still gets to the top regardless of the title selection.

Categorizing submissions to different types, and repeat the experiment, you'll find the program may predict blog/article and "Show HN" submissions with higher accuracy.

[mrburton]

> This project is far from credible. All the things I did were to satisfy my own curiosity. With that being said, the bigger limitation I can see is that I only had access to a few stories. I also cannot validated the neural network prediction, cause in order for me to do that, I would have to write a content, come up with a title and then post it choosing words that triggers a good value on the neural network and post that history on a Friday noon, to see if my story succeed.

This is from the Github project.

[brudgers]

  new mac

seven characters, 0.0060 bad, 0.9940 good

[DoctorOetker]

"I know"

6 characters Bad: 0.0085 - Good: 0.9915

EDIT: find a higher score than yours at one character less

"I went"

6 characters Bad: 0.0025 - Good: 0.9975

for 5 characters:

"I won" Bad: 0.0031 - Good: 0.9976

===

at 4 characters:

"J ML" Bad: 0.0002 - Good: 0.9998

[ekianjo]

"ruby rails nodejs bad" -> good at 0.9975 or something

[kordlessagain]

"Show HN: Cow Robots Neural Network Smells"

Bad: 0.0267 - Good: 0.9720