HN

Comments (60)

• Menu

  tiborsaas
  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 :)
• jedbrooke
  I wonder if this approach would work for AdderBoard[1][2] (a competition to make hand written transformers that can add two 10 digit numbers in the least amount of parameters possible), or if it uses way too many parameters to be competitive[1]: https://github.com/anadim/AdderBoard [2]: hn discussion https://news.ycombinator.com/item?id=47170030&utm_source=per...
• Menu

  jerrythegerbil
  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.
• jacquesm
  Heh, this is very timely, it is more or less exactly what I'm busy with right now but I'm taking the idea in a completely different direction. Fascinating.
• giveexamples
  https://www.percepta.ai/blog/can-llms-be-computersThey implemented wasm in transformer with cpu decoding for the wasm part.
• Menu

  jazzyjackson
  > 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 togetherI’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
• Menu

  hislaziness
  This is amazing, now waiting for someone to run an llm on the cpu. Eventually it will be LLMs all the way down...
• Menu

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

  andai
  > Now, here's where our fun experiment starts. I found out that by setting the gate to 1,1 the write strength (beta) to 1,2 and using one-hot key encoding,3 every write we perform will overwrite exactly one row, and every one-hot query can retrieve an exact value. This means that the GDN layers themselves can work exactly as RAM, at a far lower cost than using normal attention as RAM, since a GDN head's state stays the same size however long the conversation gets, unlike attention's KV cache.So like... can't you use a computer as a computer without doing all that roundabout stuff? Like if you want to integrate a LLM with hard computation there's probably a more efficient way to do that.Or is this building towards "but can it run DOOM?" :)
• mkturkcan
  Nice project and write up! I’ve done similar work with a custom tiny network I call Loom at https://arxiv.org/abs/2604.08816 ; you need far fewer layers, and you can run Doom, very slowly: https://mehmetkeremturkcan.com/Loom/demos/
• Founderarcstone
  That’s great new!
• Ericson2314
  The real question is, do current training methods allow one to learn this when training on enough, say, assembly, or can the weights only be achieved via hard-coding?A fun and important thought experiment about the the space of nerual networks vs the image of current training methods.Most discourse conflates neural networks and trainable neural networks.
• bronlund
  No, I don’t want you to use the computer, I want you to be the whole computer. No mistakes.
• throwaway81523
  Wow! I wonder if it can start programming itself soon.
• Menu

  blurbleblurble
  This is cool but it'd be way more useful if they had made a GPU, seeing as how expensive those have gotten
• yaoke259
  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!
• lacoolj
  I 100% see the potential here and am dying to see it in 30 daysOr maybe ill join your discord and hopefully we can riff it out to get there faster
• jonrouach
  i wonder if you could implement "authenticated token authorship" and self-verified running with this. the model verifies an input as signed by an authorized requester, and signs its own outputs. eventually it also authenticates its own running via e.g. snarks? :) there's lots of meta available with these toys..
• Menu

  farhanhubble
  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?
• zardo
  How much VRAM do I need to run DOOM?
• voxelghost
  Makes me want to run agents on it
• dools
  Have you any idea how it feels to be a Qwenputer living in a Qwenbot's Qwenputer's world?
• rumwerfer
  But can it run Doom?
• 0x20cowboy
  This is very clever. An inversion of tools. Awesome.
• anon
  undefined
• Menu

  seba_dos1
  Now run LLM inference on it.
• fr2029r
  [dead]
• favori995749721
  [flagged]
• hujun236
  [flagged]
• cinderwick
  [flagged]
• rickcarlino
  If anyone else has an easier way to emulate RISC-V I’d like to hear it. /S
• Menu

  gigel82
  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.
• Menu

  ansk
  [flagged]
• Menu

  rvz
  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
• Menu

  ram_rattle
  Cool cool cool