Blog of Ken W. Alger

Calendars, evidence, and the difference between done and stuck I expected my pear wine to spend two to three weeks in primary fermentation.

Context isn’t free.

I spent a weekend running two small language models against three jobs I would normally hand to a language model without thinking about it.

Every introduction to programming eventually reaches data types.

Nine years in, I finally worked out what else to count.

Attribution, allocation, and why “unknown” belongs in the schema.

AI can generate the implementation.

What a refractometer taught me about observability In July, I wrote about a farmer walking a vineyard row at dawn, crushing a grape onto a refractometer prism, and typing “13.5 Brix” into a chat window.

There is an API.

Part 9 of the Building the AI Memory Stack series This series began with a deceptively simple observation: a context window is not memory.

I stored passwords in plaintext on a floppy disk in the 1980s.

Part 8 of the Building the AI Memory Stack series Over the past seven articles we’ve built an architecture that treats memory as infrastructure rather than as an oversized prompt.

A career has edges.

AI can generate the code.

AI can generate code faster than ever.

Part 7 of the Building the AI Memory Stack series Back in Part 1, I made a promise: memory eventually becomes voice.

I went looking for a better runtime check and found a different way to think about the problem.

Part 6 of the Building the AI Memory Stack series At the end of the last article, I left one question unanswered.

Because “we’ll be in touch” is not an observable state.

Building a Write-Side Custody gate in Go AI memory systems spend most of their design budget on retrieval.

Part 5 of the Building the AI Memory Stack series The previous articles introduced the Reasoning Ledger and then worked through what a single ledger record should contain.

A companion to Part 4 of the Building the AI Memory Stack series.

The memory store may know the truth.