> 25K parameters is about 70 million times smaller than GPT-4. It will produce broken sentences. That's the point - the architecture works at this scale.
Since it seems to just produce broken and nonsensical sentences (at least based on the one example given) I'm not sure if it does work at this scale.
Anyway, as written this passage doesn't really make a whole lot of sense (the point is that it produces broken sentences?), and given that it was almost certainly written by an AI, it demonstrates that the architecture doesn't work especially well at any scale (I kid, I kid).
forinti 9 hours ago [-]
How does it compare to a Markov chain generator I wonder.
jll29 7 hours ago [-]
The Transformer is the more powerful model than Markov chain, but on such a weak machine as the C64, a MC could output text faster - but it surely would sound "psychedelic", as the memory limits a MC to a first-order or second-order model, so to predict one word, only the two words before would be taken into account as context (and no attention).
On a plain vanilla C64, the Transformer cannot really show what it's capable of doing. An implementation using 2 bit per weight (vectorized) could be slightly better, perhaps.
pizza234 8 hours ago [-]
[dead]
mixmastamyk 8 hours ago [-]
Just reminded me of the random sentence generator program on my Vic-20. I had changed most of the words to all the bad words a preteen could think up. So many laughs with the neighborhood kids.
arketyp 34 minutes ago [-]
I love these counterfactual creations on old hardware. It highlights the magical freedom of creativity of software.
It (v3) mostly only says hello and bye, but I guess for 25k parameters you can't complain. (I think the rather exuberant copy is probably the product of Claude et al.)
anyfoo 8 hours ago [-]
This would have blown me away back in the late 80s/early 90s.
(Or maybe not, if it doesn't perform better than random, I haven't actually tried it out yet. Some more examples would have been nice!)
I wonder how far you could push this while still staying period correct, e.g. by adding a REU (RAM Expansion Unit), or even a GeoRAM (basically a REU on steroids).
SuperCPU would also be an option, but for me it's always blurring the line of "what is a C64" a bit too much, and it likely just makes it faster anyway.
LeFantome 6 hours ago [-]
How fast is the “new” Commodore 64?
Have not heard much about it since launch. Although, now that I look, it seems they are just shipping now.
RAM can be increased to 16 MB and CPU speed to 48 GHz.
wk_end 2 hours ago [-]
I’m sorry how many Hz???
borsch_not_soup 6 hours ago [-]
Interesting, I’ve always thought neural network progress was primarily bottlenecked by compute.
If it turns out that LLM-like models can produce genuinely useful outputs on something as constrained as a Commodore 64—or even more convincingly, if someone manages to train a capable model within the limits of hardware from that era—it would suggest we may have left a lot of progress on the table. Not just in terms of efficiency, but in how we framed the problem space for decades.
dpe82 6 hours ago [-]
YOU> hey
C64> HELLO! RE SOUNDS ME. MEFUL!
60s per token for that doesn't strike me as genuinely useful.
Very, very cool project though!
chillingeffect 5 hours ago [-]
not useful in a disaster scenario:
YOU> HELP I'M DROWNING
C64> YOU' HERE!
YOU> OH NO I'M ON FIRE
C64> IGLAY!
YOU> IM BEING SWALLOWED BE A SNAKE
C64>
YOU> BIRDS ARE NIPPING ON ME
C64> YOU
Razengan 4 hours ago [-]
Reminds me of Terry Davis' random word generator :')
Maybe there is deeper wisdom in there that we have yet to unearth
numpad0 2 hours ago [-]
Next-word prediction features always existed for flip phones...
classichasclass 8 hours ago [-]
If you're running this in VICE, run it under the SuperCPU with warp mode on.
bartread 8 hours ago [-]
That's a good idea because, although I love this, 1 minute per token is absolutely savage. Whereas if you can juice the performance you're into semi-credible Jar Jar Binks simulator territory.
It does also make me wonder what you could do with somewhat more powerful retro hardware. I'd love to see what a transformer running on a PSX or an N64 could do.
7 hours ago [-]
djmips 4 hours ago [-]
Dissapointed - there was no 6502 code in the GitHub repo.
brcmthrowaway 8 hours ago [-]
How does this compare to ELIZA?
Geee 7 hours ago [-]
ELIZA is better, because this doesn't seem to generate anything coherent. You can try the original ELIZA with DOCTOR script here: https://anthay.github.io/eliza.html
jll29 7 hours ago [-]
Jopsph Weizenbaum's ELIZA was rule-based and ran on even slower (1960s) hardware, but because it relied on simple pattern matching instead of neural nets, it would easily have been more responsive (the Emacs editor/operating system has an implementation included, start it with: M-x doctor RETURN).
ELIZA was not written in assembler, but (different versions) in COMIT, FORTRAN and LISP.
Eliza called, and asked if we saw her grand kids...
tclancy 9 hours ago [-]
What makes you say that? This is about you, not me.
(Came here to say an update to Eliza could really mess with the last person still talking to her.)
Vaslo 6 hours ago [-]
Load”*”,8,1
Brings back memories
Lerc 9 hours ago [-]
Ok now we need 1541 flash attention.
I'm not sure what the venn diagram of knowledge to understand what that sentence is suggesting looks like, it's probably more crowded in the intersection than one might think.
dnnddidiej 1 hours ago [-]
How many 40+ AI pillers? Assume 10M devs in the world. 10% heard of flash attention, 1% heard of 1541 then 10,000
Lerc 17 minutes ago [-]
Ahh but you also have to know the significance of the 1541 that makes the Flash attention reference work
bighead1 9 hours ago [-]
i hate ai, and i love the c64, but i'll allow it.
ghstinda 8 hours ago [-]
but can you make mac keyboards feel like a c64c?
Rendered at 07:31:43 GMT+0000 (Coordinated Universal Time) with Vercel.
Since it seems to just produce broken and nonsensical sentences (at least based on the one example given) I'm not sure if it does work at this scale.
Anyway, as written this passage doesn't really make a whole lot of sense (the point is that it produces broken sentences?), and given that it was almost certainly written by an AI, it demonstrates that the architecture doesn't work especially well at any scale (I kid, I kid).
On a plain vanilla C64, the Transformer cannot really show what it's capable of doing. An implementation using 2 bit per weight (vectorized) could be slightly better, perhaps.
It (v3) mostly only says hello and bye, but I guess for 25k parameters you can't complain. (I think the rather exuberant copy is probably the product of Claude et al.)
(Or maybe not, if it doesn't perform better than random, I haven't actually tried it out yet. Some more examples would have been nice!)
I wonder how far you could push this while still staying period correct, e.g. by adding a REU (RAM Expansion Unit), or even a GeoRAM (basically a REU on steroids).
SuperCPU would also be an option, but for me it's always blurring the line of "what is a C64" a bit too much, and it likely just makes it faster anyway.
Have not heard much about it since launch. Although, now that I look, it seems they are just shipping now.
https://www.commodore.net/product-page/commodore-64-ultimate...
If it turns out that LLM-like models can produce genuinely useful outputs on something as constrained as a Commodore 64—or even more convincingly, if someone manages to train a capable model within the limits of hardware from that era—it would suggest we may have left a lot of progress on the table. Not just in terms of efficiency, but in how we framed the problem space for decades.
Very, very cool project though!
YOU> HELP I'M DROWNING
C64> YOU' HERE!
YOU> OH NO I'M ON FIRE
C64> IGLAY!
YOU> IM BEING SWALLOWED BE A SNAKE
C64>
YOU> BIRDS ARE NIPPING ON ME
C64> YOU
Maybe there is deeper wisdom in there that we have yet to unearth
It does also make me wonder what you could do with somewhat more powerful retro hardware. I'd love to see what a transformer running on a PSX or an N64 could do.
ELIZA was not written in assembler, but (different versions) in COMIT, FORTRAN and LISP.
https://dl.acm.org/doi/pdf/10.1145/365153.365168
(Came here to say an update to Eliza could really mess with the last person still talking to her.)
Brings back memories
I'm not sure what the venn diagram of knowledge to understand what that sentence is suggesting looks like, it's probably more crowded in the intersection than one might think.