Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Building a neural network in pure Lisp using only atoms and lists (woodrush.github.io)
107 points by signa11 on Jan 17, 2022 | hide | past | favorite | 7 comments


Bit misleading, parameters here are already obtained - the only lisp part is applying pre-obtained weights to an input and computing the output (i.e. no training here).

That aside, this is very cool nevertheless.


This is super cool, next step - full autodiff support to train it in lisp too.


This was actually a Shown HN just a day ago: https://news.ycombinator.com/item?id=29954994


Interesting. Of course some basic built-in IO or interface with external world is important.


In case of SectorLisp, that's good ol' BIOS I/O for interacting with the keyboard and screen. Having any other interface would require implementing it first - remember that SectorLisp is meant to run on bare hardware.


All hail (((GPT-5)))!!!


I for one welcome (GPTγ GPTγ) :-)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: