Field note 01Portland, Oregon

I moved from monkey knees to machine minds.

I’m Shane Neeley: a data and software engineer, former biology researcher, and author interested in the long, strange path from evolved intelligence to artificial intelligence.

15
years building software and data systems
11
scientific publications co-authored
01
book edited by Darwin’s great-great-grandson
02

Context is the model

Generic AI knows the internet. These models know my margins.

Each answer is assembled from a deliberately small, personal corpus—not a synthetic persona. Different model sizes receive different cuts of the same evidence.

03

Selected evidence

The through-line is translation.

  1. 2010–13

    Biology into systems

    Viral genetic engineering, animal surgery, and a Rice master’s in bioengineering and systems biology programming.

  2. 2013–24

    Research into useful software

    Clinical diagnostics, search, APIs, machine learning, cloud systems, consulting, and the revenue-driving product at a company later acquired by Roche.

  3. Now

    Data into better decisions

    Senior data engineering for consumer healthcare, with AI workflows that have to survive contact with real people and real constraints.

Live context experiment

Same Shane. Three sizes of mind.

Ask once and compare what survives compression. The two smaller voices run on your machine; the frontier voice uses a low-volume hosted endpoint.

  1. PocketSmolLM2 · 135M · in your browser
  2. BrowserQwen 3.5 · 0.8B · WebGPU
  3. FrontierNemotron 3 Nano · NVIDIA NIM

Local models download once and remain in your browser cache. On phones, the hosted voice answers alone. Compatible browser agents can search the same library through WebMCP.