Engineering Blog

We have reached another AI tool inflection point: guitar youtubers and other musicians are using Reverse Engineer Anything to clone plugins, and hacking firmware in their guitar pedals to let them to what they want. Without any prior knowledge of software development. This could mean less hardware sold, or following Jevons' paradox, much more.

3D printing of small objects can now be done by taking some photos and chatting with an LLM, no specialist CAD software needed. This brings this kind of work where software was two years ago. Specialist smaller models for geometry are also up and coming.

Communicating our intent clearly is helpful in getting the results we want. I find the pattern Intention, Location Details as described by Emily Bache quite handy, It took me a while to find sources, I am collecting them here, before expanding on my own musings.

In which the author uses a local text-to-speech model to draft blog posts, starting from talking while sketch noting.

Will it Blend was a series of videos promoting a blender brand, in which various unlikely objects were put in a blender to see if they would blend.

I made a prototype and let it sit for over two months. why? attention is all i need - focus can be frustrating, because i can't do the other things, but it is also what lets me enjoy holidays and get good results out of my work.

Synthetic Test Driven Development is when you have a coding agent generate the tests and the code, and maybe also do the refactor step with an agent.

Next annoying variation in my blog writing workflow is adding images to a post. I wanted to add drag and drop. I let a coding agent generate something that almost worked, and then did the finishing touches by hand, simplifying the UX and code as I went.

Checking out a repository on a fresh machine often involves some fiddling around. An init scripts costs only one prompt and gets you going now, and months later.

In which the writer reflects on believing AI is conscious

The last ten days have seen a phase shift in quality and speed of open weight models and inference. I am still adjusting.