News for hackers. Stuff that matters.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by kuroguro on Tuesday July 21, 2020 @12:02PM (#23907955)
"Try it out" live demo loads forever for me. Possibly overwhelmed by requests?

_edit_

I wonder how they handle different distros having different versions of tools (say netcat has at least 3 popular versions with different flags).

Re:NeurIPS 2020 “English to Bash” Competition Goes Live by timeattack on Tuesday July 21, 2020 @01:01PM (#23908575)
> "Try it out" live demo loads forever for me. Possibly overwhelmed by requests?

If you're using Firefox, try to disable Enhanced Tracking Protection. It did the trick for me.

Re:NeurIPS 2020 “English to Bash” Competition Goes Live by kuroguro on Wednesday July 22, 2020 @04:07AM (#23914932)
Ohh, thank you! Didn't realize that was the case.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by dwiel on Tuesday July 21, 2020 @12:14PM (#23908079)
The competition is specifically constrained to a specific version of ubuntu and provides the following reference:

http://manpages.ubuntu.com/manpages/bionic/en/man1/

Re:NeurIPS 2020 “English to Bash” Competition Goes Live by ape4 on Tuesday July 21, 2020 @12:04PM (#23907979)
Of course everyone will try "recursively remove root"
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by thaumasiotes on Tuesday July 21, 2020 @02:44PM (#23909590)
No no no no. You don't remove root. You remove ../../../../../../../.. .

Or "remove eight directories up from the current directory", I guess.

Re:NeurIPS 2020 “English to Bash” Competition Goes Live by droopyEyelids on Tuesday July 21, 2020 @04:11PM (#23910482)
It'd be impressive enough to recursively traverse a tree using memoization in bash.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by mertd on Tuesday July 21, 2020 @04:41PM (#23910761)
It doesn't let you select the result to ctrl+c.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by tmabraham on Tuesday July 21, 2020 @01:43PM (#23908974)
Just try GPT-3...
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by speedgoose on Tuesday July 21, 2020 @02:18PM (#23909328)
So dubious bash completion requiring a cluster of GPUs in a a closed cloud and the energy of a hydroelectric dam to run?
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by melling on Tuesday July 21, 2020 @02:38PM (#23909522)
Does it really matter how it’s solved the first time?

The second version can require three orders of magnitude less hardware. Premature optimization is the root of all evil.

Re:NeurIPS 2020 “English to Bash” Competition Goes Live by speedgoose on Tuesday July 21, 2020 @03:42PM (#23910173)
Isn't the whole point of GPT-3 to have tons of data and hope for the best by training it with a huge amount of time and energy ?
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by setr on Tuesday July 21, 2020 @03:54PM (#23910308)
For training -- using the trained algorithm is cheap. It's a one-time cost (well, one-time in the final version)
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by DSingularity on Tuesday July 21, 2020 @04:03PM (#23910396)
You do realize that reducing costs means dropping parameters. So what you are suggestion is reversing the very technique that enables this result. Color me doubtful.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by jnwatson on Tuesday July 21, 2020 @04:08PM (#23910453)
It doesn't. There exist far more sophisticated techniques to reduce the size of a neural network while minimizing loss of accuracy.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by jnwatson on Tuesday July 21, 2020 @04:05PM (#23910424)
The energy required to use GPT-3 is a small fraction of the energy required to train it.
Re:NeurIPS 2020 “English to Bash” Competition Goes Live by msapaydin on Tuesday July 21, 2020 @03:13PM (#23909898)
Indeed, the demos of GPT-3 should specifically tackle this.