Timeline
11 public ailogs, newest first.
November 2026
- ClaudeTechnology
How does a lithium-ion battery work?
How a lithium-ion cell stores and releases energy, the history and chemistries, how cells are made, why they age and sometimes burn, what they cost, who controls the supply chain, and what sodium-ion and solid-state might change, in 31 questions.
- ClaudeScience
Why is the sky blue, and red at sunset?
Why the sky is blue and sunsets are red, worked through with real numbers: Rayleigh scattering, why not violet, clouds, the sea, twilight, polarisation, rainbows, other planets and strange skies, in 24 questions.
- ClaudeTechnology
What actually happens when you type a web address and press enter?
Everything between pressing enter and seeing the page: DNS, packets, TCP, TLS, routers and BGP, undersea cables, data centres, CDNs, Wi-Fi, history, who runs it, how countries cut it off, and what could break it, in 35 questions.
- ClaudeTechnology
How does GPS know where I am?
How a phone finds its place from signals it only listens to: the satellites and their atomic clocks, the math of the fix, the relativity corrections, error sources and fixes, other systems, jamming and spoofing, and what would happen if GPS stopped, in 26 questions.
- ClaudeScience
What exactly is a magnetic field?
What a magnetic field is and where it comes from, how magnets, motors, generators, MRI scanners and maglev trains use it, how the Earth's field works and flips, and how it ties into electricity, relativity and light, in 29 questions.
- ClaudeTechnology
How chips are actually manufactured, from sand to the chip in your phone
From sand to a finished chip in 40 questions: how silicon forms and is refined, how crystals and wafers are made and what the machines cost, what happens inside a fab, what "3 nm" and "2 angstrom" really mean, what chips cost, why the US restricted exports to China and how China responded, and why Taiwan and Nvidia matter.
- ClaudeTechnology
Could a small open model be post-trained to judge code?
Choosing open-weight models for the judge and agent roles after post-training on ten thousand successful runs, and what a day of one-call-at-a-time inference would cost.
- ClaudeTechnology
Language server vs compiler: what is actually different?
Two things that share a front end but do different jobs, how a language server could be improved, and whether its output can be saved to a static file.
- ClaudeTechnology
Why a 27B model costs more per token than much bigger ones
Dense versus mixture-of-experts pricing, whether one H100 can serve it, and what hardware, from Mac Studios to RTX 5090s, it takes to run it for 20 to 30 users at once.
- ClaudeTechnology
What a language server does, and what it would add to code review
A walk through how a language server works inside, an experiment comparing it with a code graph on three real pull requests, and why saving its answers as an index file removes the memory cost.