Weeknote 61
Back to school officially. Distributed inference across AI or doing AI differently. Some really nice running, organisational resilience concepts. The programme really does work. Discussing bids, paying attention to what matters.
Tom Watson
Weeknote 60
Back from a break, back to school, backing off from things, paying attention.
Boring is beautiful?
Sharing the tiny boring and privacy focused tools I use for myself in a Set.
A TechFreedom startup stack?
What would a TechFreedom tech stack look like if you were starting from scratch? A question I was asked, and one which actually tested my thinking around what is the best approach here. Fundamentally it comes down to how much of a pain in the arse are you willing to put up with!
Recap May - July
My somewhat regular round up of all the posts and weeknotes and things I’ve been up to, for those of you who said you might like such a thing. the first TechFreedom cohort finished, The Slow Post happened, the Organisational Resilience programme came to an end, FIELD STATION emerged
Building Soundings: a question-shaped layer over UK open data
Another blog on the ins and outs of something I've built, what it's all about, and why it might be a useful tool and/or concept. This time: Soundings an MCP server that wraps UK open data behind question-shaped tools.
Weeknote 59
Reflection as an outcome, question banking, making an 80s style game that pits pits an applicant AI vs an Assessment AI. usual interesting things.
Weeknote 58
Meeting people old and new, FIELD STATION essays and experiments, questions>data>dashboards, tending to networks.
Weeknote 57
TL;DR: busy. Lots of client work, some exciting things, starting something new, workshops, building new things even though I said I wouldn’t, late nights with life, nice pastries, decent running, campervan insulation and possibly the greatest mushroom and lentil wraps ever created.
Put down the firehose
An exploration of building with local small models and why we probably don't need the firehose approach of frontier models for most things. Frugal AI can be done, if we choose to, and it doesn't need to cost us the quality.