A to-do list of projects. Not to be taken too seriously, and not all about programming.
- A one-button app that you can press for food to be delivered to your exact location
- Implement a neural networking library from scratch
- Implement the Doom engine from scratch
- A neural network that can expose provably automated accounts on social media
- Figure out the best way to confuse visual Neural Networks - how could one confuse facial recognition in ways that aren’t as obvious as wearing a mask? What about gait recognition?
Arduino air quality sensor - done!
- A python art generator - pixel/glitch art based on conditional logic, for loops and RNG. Secondary draw pass can include databending & filters.
- A programming language that is vague by definition and can be interpreted multiple ways.
(this is the worst idea and I couldn’t care less)
- A really angry programming language. All the method names are in caps, exclamation points mark line endings, the standard library uses language that is as angry as possible. Lowercase is only used for comments. It insults you when it fails to compile.
- Typable, monospace Tengwar (and a programming language written using its features - how can vowel diacritics be used in programming?)
- A tape delay pedal (honestly, £200 for a digital stompbox that serves a function we fixed 60 years ago? Music gear is daylight robbery)
- A personal DLP solution
- A musical language that’s based on an arbitrary set of cultural and physiological constraints - would a society with only four fingers really be forced to use a pentatonic scale? How could different vocal cord physiology affect singing - higher frequency of overtones or circular breathing? vocal cords working in a different way to humans altogether, like crickets rubbing their legs against their abdomens to create sound? How would linguistics be affected by this?
- Build a robot hand myself to see if it can be done, with a NN-control scheme based on skin electrode input, and to ideate for my character in Tom’s tabletop RPG game. Design a corresponding workshop for the handyman.
- How can FPGA’s be applied to machine learning? The idea of switchable hardware profiles for different NN tasks is really exciting.