lilbigdoot

5 時間前議論

I had this thought yesterday where I wondered if there was a way to run an Android app in the browser (definitely non trivial, but something I would expect to exist in some form) and nothing. If LLMs are as good as people say they are, I should be seeing plenty of useful projects like that popping up but.... I just see harnesses and other random things

lilbigdoot

一昨日議論

I've come to consider memory management the third

lilbigdoot

一昨日議論

I think this is a great abstraction of the process I've never considered in those terms and I like it. It explains a bit of how software development differs from some other building processes where the desired outcome is being discovered as you figure out how to make something that produces that outcome

lilbigdoot

一昨日議論

I was about to say a lot of algorithms or invariant bugs for me are slight mistakes at an edge of something

lilbigdoot

5 日前議論

That actually made me cackle ty

lilbigdoot

5 日前議論

I stopped posting there around then. Multiple times in a row I asked a question, got it closed as a duplicate/pointed to another thread/met with "why are you even trying to do that" despite being clear about my problem. It was obvious the people answering/closing the question skimmed it and I already gave a ton of context. It just stopped being useful

lilbigdoot

5 日前議論

I've tried GLM a few times but each time it ends up getting stuck in a loop and burning tons of tokens before I eventually kill it and restart. Has anyone had the same issue / know how to avoid it? I've been using Qwen 3.6 instead

lilbigdoot

5 日前議論

They also have huge influence over our military and laws. I immediately thought of the banana wars

lilbigdoot

6 日前議論

I've written multiple transpilers and their comment is how I do it. Safely mapping identifiers is one of those many edge cases you discover as you go along

lilbigdoot

7 日前議論

A sine is only a part of most kicks

lilbigdoot

18 日前議論

Agreed. Blending in CBD and other cannabinoids is way more effective than relying on strains. Basically everything out there is extremely high in THC and nothing else. Mixing in high CBD flower reduces anxiety a lot for me

lilbigdoot

18 日前議論

Yeah that's me. I quit for short periods but usually continue because my other coping mechanisms that replace it aren't any better. I naturally smoke little when life is busy and going well

lilbigdoot

20 日前議論

Seriously I've been on a sabbatical and have no pressure to use AI. I keep trying on my compiler and for coding they flail so hard every time. I love how good they are at helping me diagnose bugs, it's so much better than scouring google like I used to hoping for some result that matches what I'm seeing. Even then, they hallucinate all the time and just say odd things that don't make sense. As soon as you ask them something obscure they just.... merp

lilbigdoot

20 日前議論

I asked Qwen 3.7 pro to create a C# project that takes a string and reverses it, with a single file WASM target. It spun wheels for over 30 minutes and got nothing.

I use LLMs all the time to help me diagnose bugs and work through my designs, but again and again, I am super unimpressed by their coding abilities. I can see how in some cases with a proper harness they probably do a decent job at certain tasks, but almost everything I try to do, they flail.

lilbigdoot

20 日前議論

I tried to have multiple models convert a simple textmate grammar to a vim one, and none of them could do it. They couldn't even use the right names between the regex matches and the color definitions. I tried for about 30 minutes. It took me about 5

I tried having them work on a LSP. The fact I got a one shot half working autocomplete based on my existing work was cool, but again, they flailed on incredibly simple things like file path normalization / converting from a URI and I had to rewrite a decent amount of code. I don't think I saved any time

People keep throwing this out there but I keep wondering where are the receipts? I am seeing less interesting software released, anecdotally I know, since AI has taken hold, than before.

lilbigdoot

先月議論

I wonder if it's expertise gives you ability to see flaws and push the LLM past its acceptable point.

I haven't really used LLMs much for coding (sabbatical before LLMs got good at coding, now looking for work) but I found with chats that they are great at exploring well trodden territory but as soon as you go a little bit off the beaten path they flail horribly

lilbigdoot

先月議論

If someone was interested in trying it, the thinkythoughts folder contains an ASP.NET server side blazor project with monaco and basic (but broken ;)) syntax highlighting and debug output from inference. I don't expect anyone to be interested but if they are and want samples please don't hesitate to ask :)

lilbigdoot

先月議論

I've been working on a inference engine for a System F Omega inspired language and getting to a point where it is starting to feel good in small programs. I'm not sure why I started this, but I'm hoping to make something like an experience reminiscent of Lisp/Smalltalk by working in an interactive environment.

My goal with this language was to pick a set of primitives to compose and express as much as possible.

I don't have a demo up yet but if anyone was curious https://codeberg.org/lilbigdoot/gloo/src/branch/thinkythough...

The two main features I am missing right now are recursive types (I want to do proper mutual recursion and have been procrastinating) and some form of type classes or implicit modules. Structural typing has been useful and I'm finding a lot of features are falling out for free from that.

Long term goal is to create something with performance within a reasonable range of C# / Java etc generally, with tools for opting out of GC. I don't plan on chasing zero cost memory safety, since I want to spend my "budget" on tooling and expressiveness.

Until the language semantics stabilize I plan on generating some pretty naive JS/TS to play around with real programs, and eventually target .NET and native (likely via C++ transpilation)

コメント