I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
We love it when changes have non-localized and unpredictable results;
-
@cstanhope @mcc @mntmn @cwebber I like it.
-
@ireneista @mntmn @cwebber well it's a general purpose UI *now* but only in a very monkeys paw way
-
@cwebber of course a deterministic LLM could be made. But ~noone would use it. Being able to reroll the dice is an important part of the confidence game.
@joeyh I mean real talk that's why I don't play preset seeds in roguelikes, hooked on that RNG juice
-
@mntmn @cwebber I think the single interesting thing LLMs have revealed is that there is a substantial market segment who has an active desire for natural language interfaces to the computer and who will flip from "do not engage to the computer" to "engage with the computer" if a natural language interface became available.
I do not personally want a natural language interface to the computer. I also do not believe the thing LLM vendors have built is a natural language interface to the computer
@mcc @mntmn @cwebber speaking of expanding to more users and of assembler:
An argument I've heard is that: in the past high level compiled languages have replaced assembler, and LLMs are the next step.
Well, assembler -- and assembler-adjacent stuff like C's SIMD intrinsics -- are still relied upon (think finely optimised low-lvl libraries in some fields like gaming, video codecs, and number crunching in scientific data analysis).
... -
@mcc @mntmn @cwebber speaking of expanding to more users and of assembler:
An argument I've heard is that: in the past high level compiled languages have replaced assembler, and LLMs are the next step.
Well, assembler -- and assembler-adjacent stuff like C's SIMD intrinsics -- are still relied upon (think finely optimised low-lvl libraries in some fields like gaming, video codecs, and number crunching in scientific data analysis).
...@mcc @mntmn @cwebber ...
It's not gone. I suspect there might be even more people with the know how than back in the days.
It's just that thier numbers haven't grown as fast as, e.g., the number of people who nowadays know only Python or other high-lvl languages, and would never dare to learn anything lower-lvl and would be abandonning computing back in the days.
... -
@mcc @mntmn @cwebber ...
It's not gone. I suspect there might be even more people with the know how than back in the days.
It's just that thier numbers haven't grown as fast as, e.g., the number of people who nowadays know only Python or other high-lvl languages, and would never dare to learn anything lower-lvl and would be abandonning computing back in the days.
...@mcc @mntmn @cwebber ...
Proper software engineering done by human with brains that can hold an actual mental model of the design implication isn't going away.At best, what we have is a (very unreliable) tool that will allow people who are utterly allergic to programming languages to still be able to play around code.
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber they're lossy pseudorandom decompression
-
@cstanhope @drwho @mcc @mntmn @cwebber And to bring it full circle, grad students *can* be compilers.
-
@cwebber exactly this. on the flip side, there seemed to be a vast desire among management types and maybe hobbyists for some super easy super high level language. but idk if it's even worth going there. avoiding the details only works until it doesn't
-
@joeyh I mean real talk that's why I don't play preset seeds in roguelikes, hooked on that RNG juice
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber oh, they could… if you operated them yourself. Snapshotting, and saving the PRNG seed.
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber mostly agree, especially about them not being compilers, but some compilers aren't deterministic. You'll get a different result in memory layout or optimization sometimes. Especially for quantum compilers, where the compilation process itself is known to be NP hard, so heuristics are used.
-
@cwebber If I hear "LLMs are like higher level languages" one more time I will end up on the news, i think
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber
PGO go brrrrr -
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber This is more like the Pentium 4 idea of predictive branching, but with even larger pipeline stalls. Except the P4 could still do math.
-
@cstanhope @mcc @mntmn @cwebber I like it.
@drwho @cstanhope @mcc @mntmn @cwebber Honestly, I would prefer LLM generated code over grad student generated code.
-
@cwebber mostly agree, especially about them not being compilers, but some compilers aren't deterministic. You'll get a different result in memory layout or optimization sometimes. Especially for quantum compilers, where the compilation process itself is known to be NP hard, so heuristics are used.
@rdviii Ok but who's actually talking about *quantum compilers* when they are just saying "compilers" as a general term? ... other than people who work exclusively on QC's, who would be ... an incredibly tiny minority
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber precisely that!
A #shitposting - Program is anything but #reproduceable and I want #ReproduceableBuilds for #auditability, #security and #transparency.
- That's the whole reason I do @OS1337: To have something so fundamentally simple and compact that it is (at least in theory - at some point) financially feasible to crowdfund complete code audits of the entire system.
- I don't want people to trust me blindly, but to earn trust in the few things I code.
That's why I treat any "#AI" / #AIslop the same way @dolphin treat any leaks from Nintendo:
- That's the whole reason I do @OS1337: To have something so fundamentally simple and compact that it is (at least in theory - at some point) financially feasible to crowdfund complete code audits of the entire system.
-
I keep seeing lots of people saying "LLMs are like compilers/assemblers for prompts"
Noooooooooo
NooooooooooooooooooooooooooooLLMs are not compilers, and they're not assemblers. Determinism is a key aspect to assemblers and compilers.
And they *certainly* can't be part of a reproducible pipeline
@cwebber if, just like with asm, reading and reviewing generated code is not longer a necessary thing, then the productivity bottleneck shifts to how much time is spent "engineering" the prompt.
