I build web systems, run them, and write down what breaks.
I’m Lance Keay, a senior full-stack developer and DevOps engineer.
This is my notebook: hands-on reviews of AI tools, agents and self-hosted software, and a monthly look at what model tokens cost.
Latest posts
Each one ends with a verdict. Here it is up front.
-
Aider with local models through Ollama: two weeks on my own backlog
Worth a trial.A 32B local model handles small, well-described edits; for refactors across files I still reach for a hosted model.
-
AI token market, November 2026
Takeaway:Output tokens still cost several times input across the hosted tiers; hosted open-weight models stay the cheap floor.
-
Backups I actually test: restic, a second disk and a monthly restore drill
Keep using.Boring, cheap and it restored everything I asked of it. The drill matters more than the tool.
-
Letting a coding agent open its own pull requests for a week
Skip for now.Review time ate the time it saved. Useful for dependency bumps, not for features.
What a million tokens costs, by tier
The monthly market post compares list prices per million tokens across providers. This is the shape of it.
Price per million tokens, USD
Sample figures to show the format, not current list prices.
Show as a table
| Tier | Input | Output | Output ÷ input |
|---|---|---|---|
| Frontier, largest | $5.00 | $25.00 | 5.0× |
| Frontier, mid-size | $3.00 | $15.00 | 5.0× |
| Frontier, small and fast | $0.80 | $4.00 | 5.0× |
| Open-weight 70B class, hosted | $0.60 | $0.80 | 1.3× |
| Open-weight 8B class, hosted | $0.10 | $0.10 | 1.0× |