Remix.run Logo
▲ Qwenputer: A functioning, deterministic RISC-V CPU inside Qwen3.5(divagr.com)
97 points by csmantle 2 hours ago | 19 comments
▲yaoke259 4 minutes ago | parent | next [-]

It could be interesting to experiment with more complex representations in the memory/file system, i.e. vectors. Seems like the full advantage of this architecture is yet to be explored. Blending neurons with deterministic compute to fit on one GPU is a fascination concept, never considered it before!

▲jerrythegerbil 6 minutes ago | parent | prev | next [-]

There was an interesting bit of research that qwen has a single neuron that makes it “demonic evil qwen” and a simple “Who are you?” will get some absolutely terrifyingly vile results.

Anyways, what I’m trying to say is I want evil qwen to be my CPU.

▲tiborsaas 39 minutes ago | parent | prev | next [-]

This is fun and games, but wait until an agents finds this and does the most misaligned thing with it in order to fold paper clips.

> Because a CPU inside a language model clearly wasn't cursed enough, I decided the prudent thing would be to give it a filesystem too.

Love the spirit :)

▲voxelghost a minute ago | parent | prev | next [-]

[delayed]

▲zardo 2 minutes ago | parent | prev | next [-]

How much VRAM do I need to run DOOM?

▲jazzyjackson an hour ago | parent | prev | next [-]

> When Qwen 'saves' a file, the compiled layers hash its name into one of 1,024 slots in a directory kept in the DeltaNet state, and store the text itself in other heads, with eight tokens per row, chained together

I’ve never wanted to learn how LLMs work more than reading this. Fascinating.

Does divagr get paid to do wild things with LLMs? Just curious if they’re being compensated for these wild experiments

▲lifeisloving 26 minutes ago | parent [-]

https://divagr.com/resume.pdf?v=20260922

▲traverseda an hour ago | parent | prev | next [-]

I'd be very interested in how a scratch trained model that started training with this neural circuitry would behave. Especially the storage space.

▲horacemorace an hour ago | parent [-]

Along these lines I’m curious how a scratch trained model with a few built-in tools (memory, calculator, ?) available during a single generation would behave. Not tool calls as turns.

▲low_tech_punk an hour ago | parent [-]

a single generation would still allow looping in the "CPU" because the instruction set can do it. The model is almost a Turing machine (except for the limit on memory)

▲bronlund 19 minutes ago | parent | prev | next [-]

No, I don’t want you to use the computer, I want you to be the whole computer. No mistakes.

▲farhanhubble an hour ago | parent | prev | next [-]

This is cool and I understand the RAM emulation with gated attention but how does the CPU emulation wor? You add layers that are then trained to predict outputs while interpreting RISC instructions?

▲bee_rider an hour ago | parent [-]

I think they describe the CPU using the weights? I mean in some sense a CPU can be viewed as just a function, right?

▲ram_rattle an hour ago | parent | prev | next [-]

Cool cool cool

▲seba_dos1 16 minutes ago | parent | prev | next [-]

Now run LLM inference on it.

▲rickcarlino 38 minutes ago | parent | prev | next [-]

If anyone else has an easier way to emulate RISC-V I’d like to hear it. /S

▲gigel82 an hour ago | parent | prev | next [-]

But... why? </s>

This reminds me of C++ Template Metaprogramming (TMP) (it's Turing-complete so people wrote "games" that run at compile time). The same with TypeScript types, I think someone got DOOM running entirely in the TypeScript type system. An interesting tech demo but entirely superfluous.

▲rvz an hour ago | parent | prev [-]

So this is yet another rube goldberg machine.

We have gone down this path already in [0] and this is the same thing but overly complicated. This reason is "because LLMs".

Why?

[0] https://arxiv.org/abs/1410.5401

▲jwitthuhn 20 minutes ago | parent [-]

It is a great example to refute the common misconception that LLMs are inherently non-deterministic. I see people repeating this a lot and being able to point at a CPU implemented within one is a very strong counter-example.