7 ms·

Claude is perfectly capable of writing assembly. Here's a working (basic) Prolog interpreter that Claude Fable 5 wrote in WebAssembly in 61 minutes for $16.75 i

by ekidd 3mo ago

Claude is perfectly capable of writing assembly. Here's a working (basic) Prolog interpreter that Claude Fable 5 wrote in WebAssembly in 61 minutes for $16.75 in token costs: https://github.com/emk/fable-wasm-prolog/blob/main/prolog.wat https://github.com/emk/fable-wasm-prolog/blob/main/prolog.wa... WebAssembly is slightly easier than real assembly, but here Fable used WASM GC extensions, which are poorly documented and not yet super common. Fable didn't even need to debug it; I believe essentially all the assembly worked correctly on the first try. I have feelings about this, but I'm not pretending it isn't real.

dude250711 3mo ago
The guys with unlimited Fable/Mythos access are for some reason incapable of producing a flawless Claude Code app built entirely in native assemblies.
- bflesch 3mo ago
  They're so busy winning that they don't care about improving their own web apps ;)
- byzantinegene 3mo ago
  they use React to render their TUI :)
- cpursley 3mo ago
  You’re getting down voted, but it would really be nice if they rebuilt the thing in Rust or something like that. I’m just tired of these enormous JavaScript bundles pretending to be desktop apps. Especially when we need to squeeze as much juice as possible out of our machines.
viccis 3mo ago
Yeah I had it remake my favorite TI-89 graphing calculator game in Python and it one shotted it perfectly.
avador 3mo ago
There’s a very real dichotomy at play here. Position 0: humans depend upon AI. Position 1: AIs depend upon humans. At first blush, it seems achingly obvious that position 1 is true, whilst position 0 is a false play by con artists. Well, at first blush I agree! But first blushes are notorious for being famous last words (blushes). You see, we always knew that the tool shapes the hand. That is, as we use computational (discrete) devices, we ourselves become more computational, discrete. But what we did not anticipate is that the tool would fool others as the ACTUAL HAND. I am so fooled. Daily. My friends. Does technique encompass being? Answer me this!
furyofantares 3mo ago
I have a personal game framework that I have LLMs write games in, which is in AssemblyScript. AssemblyScript is certainly closer to TypeScript than it is to WebAssembly, but it's still this thing where the host shares some big chunk of memory with the script and you pick some memory locations to read and write as your means of exposing APIs, and there's not a lot of training data on games written in AssemblyScript and even less in my game framework specifically (none) - and the LLM does an excellent job.
- overtomanu 3mo ago
  I guess since fable has learnt deeply about 10+(maybe 100+) popular programming languages, it can just easily reason about the new 101st programming language.
- newaccountman2 3mo ago
  This is for games that run purely in the browser?
  - furyofantares 3mo ago
    No, I typically want to target either mobile or PC first - but I strongly want stuff to run well in the browser too, for early playtesting and for demos at the very least.
jimbokun 3mo ago
Just interesting to see Claude’s hourly rate is around $17. Much cheaper than a human software engineer. But comparable to the wages for some human workers.
- marcus_holmes 3mo ago
  And ofc the Chinese workers are a lot cheaper.
- wartywhoa23 3mo ago
  As soon as human software engineers are Ubered away from the field, the price will ramp up to that of a human software engineer or more, have no doubt. ROI won't make itself up.
  - voakbasda 3mo ago
    It is hard to believe that no one sees this coming. I can’t imagine it will happen any other way.
  - jimbokun 3mo ago
    Then everyone just starts running open weight models locally.
    - wartywhoa23 3mo ago
      Running it on RAM and GPUs which cost like a private jet plane, or have to be smuggled in like cocaine due to import/export tariffs or bans.
    - lenkite 3mo ago
      U.S. is actively considering restricting highly advanced "open-weights" models due to cybersecurity and national security concern. Chinese models will be enlisted in NATS - the National Terrorism Advisory System. Can't have the fat, rich big boys who give the campaign contributions fail.
  - mycall 3mo ago
dilyevsky 3mo ago
Which effort setting did you run this with?
quantumHazer 3mo ago
> Fable didn't even need to debug it; I believe essentially all the assembly worked correctly on the first try. so you don't know if it works properly or not
- soco 3mo ago
  It's also conveniently forgetting that AI was trained on the whole internet, a corpus of code including every garbage solution in SO wrong answers or random junior blogs. I didn't see any claim AI could select "best practices" from their sources, so putting AI to correct another AI has high chances of statistically selecting the same wrong solution, because that was just a common thread in those old posts they used as reference.
  - mattmanser 3mo ago
    I think you're pretty wrong here, but I'm no expert so anyone who wants to correct me please do so. As I understand it, there's post-training step that does exactly that, they get real developers to select good and bad code and help the AI figure that out. AFAIK it's the reason why Claude was able to leap-frog OpenAI in coding last year, they did this refinement step a lot better. And now they've taught the older models this, they can use existing models to bootstrap the newer models without humans involved, and bring humans in to just assess the trickier stuff.
    - soco 3mo ago
      Claude's coding efficiency is largely due to their control plane governing the actual execution of the model - rules, evaluators and basically looping over results. Not related to the model training or fine tuning.
      - mattmanser 3mo ago
        The models were already much better before Claude Code came out, let's not rewrite history here.
- 3mo ago
weinzierl 3mo ago
WebAssembly has as much to do with assembly as JavaScript has with Java. I highly doubt the "slightly easier" and my attempts with LLM's and assembly so far were largely disappointing, but I also think there is no compelling reason this has to be the case.
- ekidd 3mo ago
  I have written assembly for about 5 different processors, including 65C02s, 680x0s, cute little DSP3210s that managed the CPU cache manually, utterly cursed TI320C40s, and (of course) a bit of Intel. WebAssembly is simpler than some of these architectures in some ways. But it's not that much simpler. And once you add the WASM GC stuff, WebAssembly gets weird. It's a Harvard architecture with separate value memory, linear memory, GC memory and "tables", all accessible in completely different ways, with a weird mandatory type system (especially for the GC stuff). And the docs are often terrible. And yes, I've also written WebAssembly by hand. So yes, I would, overall, classify WebAssembly as "slightly easier". But not dramatically so. And the training data for actually writing non-trivial things by hand isn't that great, not compared to something like Intel assembly. (Don't talk to me about TI320C40 assembly. If Fable can one-shot a Prolog interpreter written using that without finding a reference manual, it's time to hang up my hat and learn to make goat cheese.)

by ekidd 3mo ago

dude250711 3mo ago

The guys with unlimited Fable/Mythos access are for some reason incapable of producing a flawless Claude Code app built entirely in native assemblies.

bflesch 3mo ago
They're so busy winning that they don't care about improving their own web apps ;)
byzantinegene 3mo ago
they use React to render their TUI :)
cpursley 3mo ago
You’re getting down voted, but it would really be nice if they rebuilt the thing in Rust or something like that. I’m just tired of these enormous JavaScript bundles pretending to be desktop apps. Especially when we need to squeeze as much juice as possible out of our machines.

viccis 3mo ago

Yeah I had it remake my favorite TI-89 graphing calculator game in Python and it one shotted it perfectly.

avador 3mo ago

There’s a very real dichotomy at play here. Position 0: humans depend upon AI. Position 1: AIs depend upon humans. At first blush, it seems achingly obvious that position 1 is true, whilst position 0 is a false play by con artists. Well, at first blush I agree! But first blushes are notorious for being famous last words (blushes). You see, we always knew that the tool shapes the hand. That is, as we use computational (discrete) devices, we ourselves become more computational, discrete. But what we did not anticipate is that the tool would fool others as the ACTUAL HAND. I am so fooled. Daily. My friends. Does technique encompass being? Answer me this!

furyofantares 3mo ago

I have a personal game framework that I have LLMs write games in, which is in AssemblyScript. AssemblyScript is certainly closer to TypeScript than it is to WebAssembly, but it's still this thing where the host shares some big chunk of memory with the script and you pick some memory locations to read and write as your means of exposing APIs, and there's not a lot of training data on games written in AssemblyScript and even less in my game framework specifically (none) - and the LLM does an excellent job.

overtomanu 3mo ago
I guess since fable has learnt deeply about 10+(maybe 100+) popular programming languages, it can just easily reason about the new 101st programming language.
newaccountman2 3mo ago
This is for games that run purely in the browser?
- furyofantares 3mo ago
  No, I typically want to target either mobile or PC first - but I strongly want stuff to run well in the browser too, for early playtesting and for demos at the very least.

jimbokun 3mo ago

Just interesting to see Claude’s hourly rate is around $17. Much cheaper than a human software engineer. But comparable to the wages for some human workers.

marcus_holmes 3mo ago
And ofc the Chinese workers are a lot cheaper.
wartywhoa23 3mo ago
As soon as human software engineers are Ubered away from the field, the price will ramp up to that of a human software engineer or more, have no doubt. ROI won't make itself up.
- voakbasda 3mo ago
  It is hard to believe that no one sees this coming. I can’t imagine it will happen any other way.
- jimbokun 3mo ago
  Then everyone just starts running open weight models locally.
  - wartywhoa23 3mo ago
    Running it on RAM and GPUs which cost like a private jet plane, or have to be smuggled in like cocaine due to import/export tariffs or bans.
  - lenkite 3mo ago
    U.S. is actively considering restricting highly advanced "open-weights" models due to cybersecurity and national security concern. Chinese models will be enlisted in NATS - the National Terrorism Advisory System. Can't have the fat, rich big boys who give the campaign contributions fail.
- mycall 3mo ago

dilyevsky 3mo ago

Which effort setting did you run this with?

quantumHazer 3mo ago

> Fable didn't even need to debug it; I believe essentially all the assembly worked correctly on the first try. so you don't know if it works properly or not

soco 3mo ago
It's also conveniently forgetting that AI was trained on the whole internet, a corpus of code including every garbage solution in SO wrong answers or random junior blogs. I didn't see any claim AI could select "best practices" from their sources, so putting AI to correct another AI has high chances of statistically selecting the same wrong solution, because that was just a common thread in those old posts they used as reference.
- mattmanser 3mo ago
  I think you're pretty wrong here, but I'm no expert so anyone who wants to correct me please do so. As I understand it, there's post-training step that does exactly that, they get real developers to select good and bad code and help the AI figure that out. AFAIK it's the reason why Claude was able to leap-frog OpenAI in coding last year, they did this refinement step a lot better. And now they've taught the older models this, they can use existing models to bootstrap the newer models without humans involved, and bring humans in to just assess the trickier stuff.
  - soco 3mo ago
    Claude's coding efficiency is largely due to their control plane governing the actual execution of the model - rules, evaluators and basically looping over results. Not related to the model training or fine tuning.
    - mattmanser 3mo ago
      The models were already much better before Claude Code came out, let's not rewrite history here.
3mo ago

weinzierl 3mo ago

WebAssembly has as much to do with assembly as JavaScript has with Java. I highly doubt the "slightly easier" and my attempts with LLM's and assembly so far were largely disappointing, but I also think there is no compelling reason this has to be the case.

ekidd 3mo ago
I have written assembly for about 5 different processors, including 65C02s, 680x0s, cute little DSP3210s that managed the CPU cache manually, utterly cursed TI320C40s, and (of course) a bit of Intel. WebAssembly is simpler than some of these architectures in some ways. But it's not that much simpler. And once you add the WASM GC stuff, WebAssembly gets weird. It's a Harvard architecture with separate value memory, linear memory, GC memory and "tables", all accessible in completely different ways, with a weird mandatory type system (especially for the GC stuff). And the docs are often terrible. And yes, I've also written WebAssembly by hand. So yes, I would, overall, classify WebAssembly as "slightly easier". But not dramatically so. And the training data for actually writing non-trivial things by hand isn't that great, not compared to something like Intel assembly. (Don't talk to me about TI320C40 assembly. If Fable can one-shot a Prolog interpreter written using that without finding a reference manual, it's time to hang up my hat and learn to make goat cheese.)