tynet-lichat/shirakumo

7:48:10 |3b| ok, so regular grid would be either 1 group per item, or 1 group containing all. i think i read original question differently from what you intended, which is why i was confused

7:48:49 shinmera Yeah, sorry for not being clearer

7:49:04 |3b| (not that i have any useful input either way :/ )

7:49:18 shinmera That makes two of us :)

10:38:49 selwyn vr nearly up and running

10:48:42 shinmera Nice

11:49:48 selwyn how would you feel about trial source code being used as a prompt to an llm

11:53:53 shinmera I don't know what that means

11:54:08 Colleen "As a large language model, I have no feelings about using the Trial source code. However, it is important to consider ..."

11:54:08 hayley "As a large language model, I have no feelings about using the Trial source code. However, it is important to consider ..."

12:27:26 shinmera gingerale: Okey, things are integrated now and there's a very primitive test suite as well.

12:27:34 shinmera also wrote an extremely simple grid3

12:27:43 shinmera (dense array, single index)

12:28:02 shinmera I'm thinking about how things are handled for UPDATE, though.

12:28:25 shinmera Right now pretty much every container is required to also hold a hash table where they keep data about the object at last update/insert

12:28:54 shinmera I wonder if it would be better to instead force users to supply a previous location to UPDATE so that this is no longer necessary.

12:29:08 shinmera the advantage of course being not needing an extra table

12:29:22 shinmera the disadvantage being that the user has more work and that changes in size aren't caught by this.

12:29:31 shinmera So I'm a bit torn.

12:30:05 selwyn i plan to use a llm to translate voice commands to lisp snippets

12:30:26 selwyn in order to make the trial-vr voice coding concept more practical

12:30:48 shinmera so you mean using it as *training data*, not as a prompt

12:31:01 selwyn not training data

12:31:05 shinmera uh, ok?

12:31:24 selwyn gpt4 supports large prompts

12:31:54 selwyn theres the prospect of giving it the hyperspec and/or some libraries as a big prompt

12:31:57 shinmera oh, so you mean: what do you think of the legality of using source code spit out by gpt.

12:32:08 selwyn not so much that

12:32:13 shinmera then I'm still confused.

12:32:35 selwyn as the fact that it is your code, and there are some concerns about how these companies behave

12:32:41 selwyn e.g. with github copilot

12:32:45 shinmera Yes.

12:32:54 selwyn so i wanted to ask if you were fine with it as a courtesy

12:33:06 selwyn i understand if you don't want your code anywhere near this stuff

12:34:08 selwyn as i understand it there would be no training stage (yet) though they surely store the conversations

12:35:19 shinmera My thoughts on this are: 1. microsoft and other "machine learning" capitalist companies belong burned to the ground 2. AI bros should be guillotined 3. what I want has very little bearing on what will happen, my code is probably already in copilot and other garbage 4. I question the legality of using the output of GPT for code that you publish, so I would rather not include any such code in Shirakumo repositories.

12:35:59 shinmera That all said, if you're going to do it, I probably won't notice, and don't particularly care.

12:36:14 selwyn ok

12:37:12 shinmera I personally would be much more interested in a system that used "standard" voice recognition to do the code input, rather than hoping GPT will somehow magically make this method convenient.

12:37:31 shinmera Perhaps combined with special commands or gestures to aid the input.

12:37:54 Colleen "Can we have program synthesis" "We already have program synthesis at home"

12:37:54 hayley "Can we have program synthesis" "We already have program synthesis at home"

12:38:04 shinmera Heh

12:38:24 Colleen Oracle v Google implies you can't copyright types, so go nuts.

12:38:24 hayley Oracle v Google implies you can't copyright types, so go nuts.

12:39:09 shinmera I personally just feel uncomfortable with the idea of a program basically doing the equivalent of studenst plagiarising wikipedia for you.

12:39:33 shinmera So imo any code output by such things should be assumed to be plagiarised.

12:39:48 shinmera Because that's what it does

12:39:57 selwyn for the voice recognition itself i am using off the shelf stuff

12:40:04 Colleen The university sent us a fair bit of advice on using large language models. We are told to cite e.g. ChatGPT and Bing as if it were private communication.

12:40:04 hayley The university sent us a fair bit of advice on using large language models. We are told to cite e.g. ChatGPT and Bing as if it were private communication.

12:40:06 selwyn what do you mean by 'standard'

12:40:39 shinmera selwyn: anything that doesn't try to be free form and instead is tightly governed by user input.

12:40:51 Colleen That reads in Lisp code rather than a natural language prompt?

12:40:51 hayley That reads in Lisp code rather than a natural language prompt?

12:40:57 selwyn ok

12:40:59 shinmera voice recognition will also use "machine learning" methods, but it won't try to infer more than what you say.

12:41:09 selwyn as it ought to

12:41:32 shinmera I think using the controller buttons and gestures you could do stuff like handle paren management to ease the input

12:41:42 shinmera so you don't have to keep saying "open parenthesis" "close parenthesis"

12:41:55 selwyn this is close to my original idea

12:42:33 selwyn i wanted to create a command language of about 100 syllables to represent emacs commands, mostly paredit

12:42:56 selwyn but it would be tricky

12:42:57 Colleen I suggested to stylewarning to add a "Wadler, take the wheel" thing to Coalton, which just tries to come up with an expression of the right type. Not easy to implement, nor would the humour justify it though.

12:42:57 hayley I suggested to stylewarning to add a "Wadler, take the wheel" thing to Coalton, which just tries to come up with an expression of the right type. Not easy to implement, nor would the humour justify it though.

12:43:33 shinmera hayley: u should just let gpt fix ur gc

12:43:50 selwyn llms allow for the possibility to say out loud 'computer put a cube in the corner' and it spits out '(trial:enter ...)'

12:44:19 selwyn i then inspect the output before evaluating it

12:44:33 selwyn if i get lazy with that then bad things can happen

12:45:14 Colleen Did ask GPT to write a commit message for me. Didn't use it, but it wasn't offensively wrong.

12:45:14 hayley Did ask GPT to write a commit message for me. Didn't use it, but it wasn't offensively wrong.

12:46:12 Colleen It could also sketch out a shitty mark sweep, and then reinvented prefetch-on-grey after I asked for prefetching, and told it to prefetch earlier than just before tracing.

12:46:12 hayley It could also sketch out a shitty mark sweep, and then reinvented prefetch-on-grey after I asked for prefetching, and told it to prefetch earlier than just before tracing.

12:49:00 selwyn which fork/branch of sdf/bmfont am i supposed to use

12:49:25 shinmera selwyn: just use the dist, it'll have the right versions.

12:49:35 shinmera http://dist.shirakumo.org/shirakumo/

12:49:44 Colleen Confessed my benchmarking sins to the MMTk people today. Really need to defrag stuff; I couldn't get GPT to write a working copying GC.

12:49:44 hayley Confessed my benchmarking sins to the MMTk people today. Really need to defrag stuff; I couldn't get GPT to write a working copying GC.

12:50:13 shinmera What did the folks say?

12:52:00 Colleen There was some confusion as I forgot to say that I fixed the RNG and load balancing of my fuzz tester to make it deterministic. Not too keen on me running benchmarks in a virtual machine, but it wasn't easy to even get a VM, so.

12:52:00 hayley There was some confusion as I forgot to say that I fixed the RNG and load balancing of my fuzz tester to make it deterministic. Not too keen on me running benchmarks in a virtual machine, but it wasn't easy to even get a VM, so.

12:52:55 shinmera Hmm, yeah running in a VM isn't ideal. Lots of noise.

12:55:07 Colleen I think it averaged out fine. It was a VM on a dedicated machine too.

12:55:07 hayley I think it averaged out fine. It was a VM on a dedicated machine too.

13:02:23 Colleen Maybe I should ask if it knows that incremental compaction algorithm.

13:02:23 hayley Maybe I should ask if it knows that incremental compaction algorithm.

13:04:01 Colleen There is also the classic incremental incrementally compacting garbage collection, if one incremental isn't incremental enough.

13:04:01 hayley There is also the classic incremental incrementally compacting garbage collection, if one incremental isn't incremental enough.

13:09:05 Colleen ACTION gets ads from Azul while trying to look up papers on her phone.

13:09:05 hayley ACTION gets ads from Azul while trying to look up papers on her phone.

13:20:47 shinmera Just shout "dog food" a couple times and you'll get that instead.

13:21:58 selwyn i created trial-assets?

13:22:11 Colleen I only use the Gmail app out of all the Google ones, which is plain laziness on my behalf. Guess they can't really sell sbcl-devel.

13:22:11 hayley I only use the Gmail app out of all the Google ones, which is plain laziness on my behalf. Guess they can't really sell sbcl-devel.

13:25:26 shinmera selwyn: yes?

13:26:44 selwyn had completely forgotten

13:26:51 selwyn no wonder it won't load :p

13:27:40 shinmera Eh? It should load.

13:28:00 shinmera I messed with it only a couple weeks ago

13:29:48 selwyn actually trial-vr is the real problem

13:30:07 selwyn i just have to reconcile the changes to trial and it should work

13:30:42 selwyn no applicable method for trial:assets when called with (nil)

13:30:49 selwyn though i don't call it directly

13:56:19 shinmera Error reporting is one of those things that could be improved quite drastically

13:56:49 shinmera And it's not hard work, either

13:57:09 selwyn trial error reporting or my error reporting? :p

13:57:57 shinmera The former

13:58:07 selwyn nearly there

13:58:24 shinmera I think doing that would also significantly lower barrier to entry

13:58:29 selwyn can i still define texture cube map assets?

13:58:37 shinmera Don't remember

13:58:52 shinmera Maybe not :)

14:03:43 selwyn no skybox then

14:07:07 shinmera Another thing to fix for you ;)

15:55:41 shinmera gingerale: I pushed docs to 3d-spaces for everything but the quadtree. If you could fill in the missing bits in documentation.lisp, that would be great!

15:59:29 gingerale Sure!

19:09:59 shinmera :fart: https://shirakumo.github.io/3d-spaces/