Apps
Things I have built, am building, or plan to build. More will land here as I finish them.
Local AI Index
What’s worth running at home
Every open model I can run on a 128 GB AMD Strix Halo machine, scored 0–100 from graded tests — plus the models on the radar.
Local AI intelligence lab
Models, harnesses, and real work
Measured local model and coding-agent results, with honest limits and the next assistant tests.
Will it run?
Does this model fit on my machine?
Pick an open model, a quant, a context length and your hardware. See if it fits in memory and how many tokens per second to expect.
Model cards
One page per tested model
Score, strengths, speed and coding-agent results for every open model measured on the test bench — each on its own page.
Frontier Watch
Quantum, photonics, humanoid robots
Dated, sourced milestones at the edge of computing — what happened, why it matters, and the catch nobody puts in the headline.
Market history
Thirty days, refreshed once a day
Daily open, high, low, close and volume for seven assets, kept current and free to scrape as JSON or CSV.
The Last 30 days in AI tech and news
Useful AI and developer news from the past month
Original summaries of material developments in AI tools, coding agents and open-source software, with source links and honest coverage limits.
Prompt Playground
Same prompt, several models, side by side
Real outputs from local models on the same prompts, saved from benchmark runs — compare answers, not just scores. Opens after the next benchmark round.
Dashboard
Your corner of this site
A customisable panel of widgets that remembers what you picked. Sign in and arrange it however you like.