Threadwell
Search
Search
Dark mode
Light mode
Reader mode
Explorer
Recent Notes
DemetriSpanos
Oct 06, 2026
archive
twitter
demetrispanos
NOTimothyLottes
Oct 06, 2026
archive
twitter
notimothylottes
bonzajplc
Oct 06, 2026
archive
twitter
bonzajplc
thegingerbill
41 items with this tag.
Aug 28, 2026
TheGingerBill
archive
twitter
thegingerbill
Aug 25, 2026
Examples of Odin's new inline `asm` templates!
archive
twitter
thegingerbill
Aug 25, 2026
This comment was asking if I used LLMs/agents for anything, so I wanted to explain why I have been recently using them for.
archive
twitter
thegingerbill
Dec 29, 2019
@FlohOfWoe @martin_cohen Sequential languages are still the bee's knees.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I can kinda see myself starting using Odom and learning assembly with it before I ever do that with C 😂
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill neat!
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Liked the syntax
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Isn't it sometimes useful to use nop instructions to align relevant code in memory (like in your example after jmp)?
archive
twitter
thegingerbill
nop
Aug 25, 2026
@TheGingerBill @TheGingerBill would it be possible that asm templates prevent the bug in https://youtu.be/PNWABi6Dcl8? Cc @LowLevelTweets
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill This is very much a personal preference, but all the big sites let you give the AI some specific "personality" details that apply to all convos.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Hear hear.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill This is exactly where I sit with using LLMs for programming
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Sometimes LLMs bad ideas are great at sparking good ones
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Please look into the physics behind it instead, and then you will see that "LLMs" are unscalable by nature due to the noise-denoise aspect they introduce.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Broadly agree, with one caveat: giving agents code to review for mistakes/find bugs is in my opinion a good use case.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Sir what would u say about coding is becoming obsolete as dhh, uncle Bob etc are saying, will the ones who are technically competent have advantage over others in the future
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill You mentioned them being passive aggressive.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Don't use LLMs as a beginner to substitute producing real work, sacrificing the learning that comes with it.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I wrote a basic deep learning framework but with a generalized version of backpropagation.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill They're pretty nice for generating bindings too, but you gotta make them triple check
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Slippery slope, this is how all accomplished programmers start ;).
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Training on the mean programmer's code is the most honest explanation I've seen for why agent demos always pop and then fall apart on a real codebase.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I use LLMs the same way, except I use them as a chatbot through opencode a and give them 0 permissions to do anything ever.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill What if you dgaf about learning to code and just use LLMs for product development?
archive
twitter
thegingerbill
Aug 25, 2026
People said much the same thing about high level languages, interpreters, etc.
archive
twitter
thegingerbill
Aug 25, 2026
LLM "agents" are a tool tho, it just means there will be a software passing messages down to the LLMs so they can execute things automatically, there's no magic on that, but it's convenient.
archive
twitter
thegingerbill
Aug 25, 2026
A lot of people will give you nonsensical shit about this.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I even have a hard time blindly trusting generating tables and charts etc from data.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Although I agree with some of that, I think you should at least try to use agents for such tasks (in good faith) so that you can accurately assess whether they could be useful or not.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I think they can be used to make learning more effective, especially for concepts where reading a book doesn't give you the full context you need.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I also find LLMs great at generating comments and documentation for an existing code base.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill That's how I was using AI until about February of this year.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Would you advice against reading a technical book using llms?
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill I ask the LLM to generate VERY small programs using an API I need to use.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill mine called my parser design naive, lol
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill i use llms the exact same way.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill When you work with LLMs are you working with them out of the web chat UI or are the apps downloaded onto your machine, or are you using third party AI tools?
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill @Wassimulator I mostly use them to narrow down/distill documentation and to elucidate errors that are very poorly explained.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill You will not deny "beginners" this: they will be all over it esp.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill mmm, they can and should be used as a search tool on your existing code base or when using an existing library for the curious.
archive
twitter
thegingerbill
Aug 25, 2026
@TheGingerBill Missing from this list - your recent blog post seemed, at best, very heavily LLM assisted.
archive
twitter
thegingerbill