I study how minds work, and I build with the new ones.
Move. Click. The field is listening.
Scroll01 About
I’m Paul, a psychologist in London. I took a First in Psychology at Goldsmiths and an MSc in Clinical Mental Health Sciences at UCL. Before that: business school in France, and a sales floor in Madrid.
I think general-purpose AI arrives this decade, and that it’ll change mental health and ordinary work more than anything else I’ll live through. So I’ve rebuilt how I work around it. A small staff of AI agents writes nearly all my code and runs my homelab. My job is to point them somewhere worth going, and to check what they bring back.
The questions I care about sit where those two worlds meet. What happens to people who talk to machines all day? And how does a human stay a useful check on a system that’s smarter than them?
- studies mapped in Psyche Machina
- 306
- commits to my homelab since April, mostly by agents
- ~4,900
- services running at home
- 28
- repair jobs in XK8 Workshop
- 43
- apps that re-theme to one wallpaper
- 14
02 Work
Things I’ve made
- 01 Psyche Machina Research A living map of human studies on how AI use affects mental health, graded by what each study can show. 306 so far. Only 41 are randomised trials, and the map shows exactly where they’re missing.
- 02 XK8 Workshop 3D · Web An interactive 3D workshop manual for the Jaguar XK8. Take the car apart, watch the engine run, follow 43 repair jobs with factory torques. Every value shows its source. English and French.
-
03 Iris Private Systems My homelab and personal operating system, from photo backup to the assistants that run my week. Held to one rule: it only works once a fresh run on the real machine says so.
- 04 Homelab Operator Open source That rule, packaged for anyone. A small CLI that stops AI agents confusing “tests passed” with “the server is fixed”, by making every pull request say what it proved and what it didn’t.
- 05 Oversight eval AI safety When AI helps you check AI, do you catch more errors or just feel surer? A reproducible study design with seeded mistakes and 4 kinds of assistance. Simulated participants for now, so it tests the design.
- 06 WalBridge macOS Change the wallpaper and 14 desktop tools re-theme to match it, with animated transitions. Linux-style ricing for macOS.
03 Research
Questions I’ve chased
-
2026 · Qualitative
Staff morale in inpatient mental health rehabilitation
MSc dissertation, UCL Division of Psychiatry. A reflexive thematic analysis of staff interviews and focus groups from NHS and independent-sector rehab units, drawn from the ACER study.
-
2026 · Evidence synthesis
Is ACT worth the money for older adults?
Systematic review placement at UCL on the cost-effectiveness of Acceptance and Commitment Therapy. Screening, extraction, quality appraisal, and pulling ICERs and QALYs out of economic evaluations.
-
2025 · Experimental
Does nonsense sound deeper with a face attached?
BSc dissertation, Goldsmiths. A 2×2 experiment on how video, pictures and audio change the way people judge pseudo-profound statements.
-
2024 · Experimental
Mimicry, listening and trust
A 2×2 study across 40 face-to-face interviews. Mirroring someone and actively listening to them each moved how much they trusted the interviewer.
04 Method
How I work
- i.
Agents do the typing. I decide what’s worth doing and check the result.
- ii.
Every claim names the check behind it. A passing test and a working server are different facts, and I write down which one I’ve got.
- iii.
Making things got cheap. Knowing which things to make is the job now.
- iv.
Systems stay small enough that I’d happily delete them.
05 Path
The long way round
- 2026
MSc Clinical Mental Health Sciences University College London
- 2026
Graduate Member British Psychological Society
- 2025
BSc Psychology, First-Class Honours Goldsmiths, University of London
- 2022
BSc Business Management, finance and banking KEDGE Business School
- 2021
Inside sales Selectra, Madrid
- Speaks
French, English and enough Spanish to get by
06 Contact