Hand range test

Experimenting with mediapipe hand-tracking for value adjustments.

Experimenting with running a modified version of Masks on movies (Fight Club).

Me recording my face using masks

Masks - Collage faces using your webcam.

Dev notes

Working om masks

I want to think more about some connected ideas. Silicon Jungle's prototypes where he is exploring breaking ideas into cards, I think called lenses, that can then be further explored.

Linus Lee's talk about using LLMs as a microscope for thoughts.

Some work experiments where we've been using LLMs to break scenes down into parts, then applying a transformation to one of those parts.

New desk experiments - stashing a beheaded desktop below the desk with string loops.

Bottom view - laptop body tied to the underside of the desk

Top view - strings coming out of drilled holes

New desk setup as of this weekend

Process

Materials, with extra for later

Thinking a lot lately about making a comfortable environment for myself. Both in real space and digital space.

In real space I think about my desk setup. I think about nostalgic 90s photos of cozy computer setups.

In digital space I think about my neovim and terminal setup. After getting things pretty well setup I've kind of let them sit - and added a few things that never quite worked right. That cruft started to build up and now I'm starting to clean it out.

Back on the train and in the glasses deck. Sitting next to a big coffee spill on the flor and a man slowly, methodically eating a sandwich.

I don't know if I'll actually publish this as a post, but I wanted to collect my thoughts by writing.

This morning I was thinking of buying a new kitchen knife and working on cooking a couple of dinners a week again. I'd like to find some sort of system for all of that that feels good.

I'm doing some more experimentation with raspberry pi's around the house. Working on a base ubuntu setup, that is pretty close to what I would do with ubuntu on any other computer.

Flashing

Install ubuntu-server through rpi-imager. Annoying rpi-imager doesn't work with sway, something about QT apps and security. I've been flashing from my work macbook so far. I'd like to get a better setup. I could do it from the command line with dd but I'm not sure how to replicate the settings presets that the imager has, which is especially nice for automatically connecting to wifi. Still, I should dig in and get this figured out.

Organizing a new quotes repo: https://github.com/grantcuster/quotes

pokemonsay hooked up to my quotes text file

Inspired by dynamicland to try some real world manipulation linked to digital objects. I'd guessed coarse color matching would be enough here - it may not be.

Red construction paper in real life moves sloth on computer

Demo that has some promise where all data is in the image file. Sources on the left, canvas layout on the top right. Data encoded as image pixels on the bottom right.

But I'm not excited to jump back into it lately so I need to figure something out.

I think it gets more interesting when I have the three ways of working with the images.

Brainstorming on the train. Laying out the pieces.

I can use mediapipe hand landmark. I know the positions of the fingers. I'll use pointer finger as cursor, if they have two hands up I can use both of them...

I can use pointer and thumb pincer as click. I can normalize distance based on lower knuckle thumb distance probably. Click and drag works with that motion too, although maybe it's a shame you can't just point and drag - maybe also some ability to toggle based on the other hand.

The elements: mini PC, AR glasses, portable battery, bluetooth keyboard

Simulation of POV from the porch. A floating window over the world.

After trying out a DIY deck version, I've redone my cyberdeck experiment using AR glasses. It's not perfect but I've been using it pretty consistently on subway commutes and porch nights.

Some of the most interesting responses I saw to Ted Chiang's piece:

Mat Dryhurst tweet:

The easy unlock re the cyclical AI art culture war that will seem obvious in retrospect is that the people creating models are artists

Redid CSS zoom method again. Hopefully a base to build on for future projects at https://github.com/GrantCuster/zoom

Zooming in on the sloth

More response writing! Here we go. Same rules, don't summarize, pick out what's interesting to me even if it leaves a bunch of other stuff out.

First his definition:

art is something that results from making a lot of choices.

Source, output, data. More wyswig work.

I'd like to get into the habit of using this blog to write down responses and appreciations of things. So let's do that:

I've been listening to the "Moving Beyond Syntax" episode from Future of Coding over the past couple days. Overall I've just really been enjoying Future of Coding episodes as a break from AI hype. As a reminder of all the interesting things there are still to do in programming as programming.

I also think the addition of Lu has been great. Now (and especially in this episode) there are three distinct points of view from Lu, Ivan and Jimmy, that bounce off and contrast with each other. I think sometimes I get stuck wanting to know / cover everything, and listening and enjoying the back-and-forth is a good reminder that it's good to be coming from a perspective. That you're part of a larger conversation and you don't have to cover every angle yourself. A good thing for me to write about is what angle I am coming from these days.