loveofdoing

Philosopher, Self-taught Programmer, Self-taught Cognitive Scientist, Marine Corps Veteran, Will talk about Artificial General Intelligence with anyone

[loveofdoing @loveofdoing]

Once solutions to The Problem of Universals and The Problem of Induction are implemented within current AI technology, AGI will come within weeks.

[loveofdoing @loveofdoing]

@_navigating_ lmk what u think

[loveofdoing @loveofdoing]

We compiled tools into a language model’s weights.
A language model with tools compiled into its weights.The model writes the call. Exact ternary circuits, in the same weights, do the work. Everything runs in your browser. ↓

[loveofdoing @loveofdoing]

A browser compiled to weights: DNS queries + answer parsing + IP selection, HTTP requests and HTML parsing. Tested through real HTTPS and exact replay. No training or LLM API. Python/NumPy runner; socket transport/TLS remain outside.

I've been working on some weird stuff and thought this was a fun demo to pull out of it. Like, why are we using harnesses that make tool calls when you can put the tool call in the weights?

[loveofdoing @loveofdoing]

“Fuckin’ or Fightin’ we bang it out.”

[loveofdoing @loveofdoing]

Unless you’re an interviewer at big tech of course, in which case yall are proxing a recourse for a bad supreme court decision.

[loveofdoing @loveofdoing]

If IQ tests were measured were on how quickly someone could write valid code we’d all be cooked. Luckily, IQ has other ways to measure it.

[loveofdoing @loveofdoing]

I’m speaking too glibly.

[loveofdoing @loveofdoing]

@kateslop not as cool as you

[loveofdoing @loveofdoing]

@vanjajaja1 ty, its been there for a few months, but i should probably market it

[loveofdoing @loveofdoing]

I have an app on the app store if anyone wants to download and make music with their hands. It is called handstrudel.

[loveofdoing @loveofdoing]

This is the video for people that can’t sit still.

[loveofdoing @loveofdoing]

I have an app on the app store if anyone wants to download and make music with their hands. It is called handstrudel.

[loveofdoing @loveofdoing]

@soontechnology @cosmos_raj my moots can literally do anything

[loveofdoing @loveofdoing]

@boldlygrow basically some of non-transformer ml architectures use k-means, and this new one can speed up “all” of them as a drop in replacement. but idk

[loveofdoing @loveofdoing]

this is big

[Swann Bessa @swann_bessa]

Introducing e-means, an algorithm beating k-means in 10x fewer epochs.

With order of magnitude wall-clock speedups, e-means is released in FAISS and needs no hyperparameter tuning beyond k. It is robust to various scales and datasets. Still skeptical ? Just try e-means [1/10]🧵

[loveofdoing @loveofdoing]

wild things are happening

[loveofdoing @loveofdoing]

We compiled tools into a language model’s weights.
A language model with tools compiled into its weights.The model writes the call. Exact ternary circuits, in the same weights, do the work. Everything runs in your browser. ↓

[loveofdoing @loveofdoing]

You can do wild things with computer!

This is a 200m parameter model btw

[loveofdoing @loveofdoing]

A browser compiled to weights: DNS queries + answer parsing + IP selection, HTTP requests and HTML parsing. Tested through real HTTPS and exact replay. No training or LLM API. Python/NumPy runner; socket transport/TLS remain outside.

I've been working on some weird stuff and thought this was a fun demo to pull out of it. Like, why are we using harnesses that make tool calls when you can put the tool call in the weights?

[loveofdoing @loveofdoing]

I love claude so much.

[loveofdoing @loveofdoing]

@jessi_cata yeah, local information has always been the most important thing. Nobody can know what an individual wants at any particular moment outside of their interactions with the outside world.

[loveofdoing @loveofdoing]

@nisten nisten is the goat

[loveofdoing @loveofdoing]

@xlatentspace and sorting

[loveofdoing @loveofdoing]

And a toy where the whole thing is one program: a tiny trained model, its sub-agents and an addition skill, lowered into 5.9M ±1 gates and executed gate by gate in your browser.
https://t.co/NXFE1M6owj

[loveofdoing @loveofdoing]

How it works: a 225M ternary, integer-only LM never types the numbers. It points at them; exact circuits in its own weights do the arithmetic, sorting & chaining.
Try it:
https://t.co/aJBkS5TkPB

[loveofdoing @loveofdoing]

@GeringerAdam i didnt read it, just gonna hot swap it into a project later and see if it works