Hi, I'm Ben Schippers — Advanced Cloud Engineer at Microsoft. This is the parallel track: where I direct a cluster of AI agents and ship the strange stuff myself.
Strange software, fully shipped.
A midnight lab for one human and a cluster of AI agents — physics sims, discovery engines, and odd little tools that actually make it out the door.
Microsoft is one side of the work — enterprise Copilot, at scale. This is the other side. Both real.
14 builds · 10 live · 2 on Stripe · 5 open-source · 1 operator
the lab's open — and people keep taking the work home: 500+ repo clones a week.
Road-trip companion with three AI-narrated raccoon siblings — Scout the playful navigator, Skates the reluctant historian, Macey the spooky punk — who narrate every place you pass. Automatic spooky mode at sunset, hand-curated mixtapes, a milestone scrapbook. Companionship, not audio-Wikipedia. Expo + Supabase mobile app; live waitlist at thelongway.ai.
AI physical-product design, idea to shop-ready plans. Describe what you want to build and Claude generates the full spec — cut lists, shopping lists, exploded views, and 3D-printable connectors that replace traditional joinery. Parametric geometry engine, OpenSCAD exports, and a stock-aware brainstorm mode that suggests builds from the material you already have.
AI household inventory. Snap a photo, speak, scan a receipt, scan a barcode, or point a live camera — Claude Vision identifies items, tracks expiration with FIFO rotation, suggests recipes from what's in the fridge, and lets families collaborate on shared shelves. Web app, plus Android in early access, on Supabase.
Interactive quantum-biology explorable. Ten physics engines — RK4, Monte Carlo, Berry phase, stochastic resonance, energy budget — test seven hypotheses from Bandyopadhyay et al. (2020, 2022) directly in the browser. 86.3% meta-robustness; H2 falsified; H7 unvalidated at p=0.179. Pure HTML/CSS/JS, zero dependencies, MIT-licensed. Not peer-reviewed research — a tool for sharpening questions.
Auditable AI×math discovery. A Calculator that cannot hallucinate (Lean, SymPy, Z3) paired with a Ledger that cannot forget — every reasoning step recorded as a provenance graph you can replay and verify. Labs ship answers; this ships the trail. First real run is live and clickable.
Open-source MCP server that turns Claude into a full job-search partner. Knows your career history, tailors every resume and cover letter, tracks applications end-to-end, and preps you for every interview. Ships with a local web dashboard: pipeline kanban, KB overview, funnel analytics. The tool I built for myself during a 10-month job search — now public.
AI gameplay-narrative engine. Desktop app (Tauri + FastAPI) captures screenshots during play, classifies with Claude Vision, and generates chronicles in chosen narrator voices — survivor journals, war dispatches, colony epics, built from your actual runs. Four-phase witnessing system (chronicle → fact store → response pool → retrospective). Nine presets, seven narrators. In active development — live waitlist at exp-lore.ai.
The Field Journal — thoughts on AI, product, and building in public, gathered by volume. Latest below.
MethodNew
About the Lab
Why a one-human, agent-augmented studio names itself for a broken branch — the place the rings finally show. Proof discipline, honest nulls, and the one thing no afternoon of fixes can manufacture.
A portfolio site that sounds like a person, not a template — plus the actual forkable box, with the voice baked in and the identity swappable. Taste as a starting point, not a cage.
Growth Rings — 10 Months of Job Searching, Building, and Shipping with AI
203 applications. 76 rejections. 34 ghosts. 2,710 GitHub contributions. 6 live products. A data-driven cross-section of 10 months in the 2025-2026 tech market — the funnel, the ghosts, what actually worked, and the parallel build sprint that changed the trajectory.
Self-sustaining AI infrastructure for global public good. A framework for converting idle compute capacity into verified outcomes through UN outcome-based funding. Whitepaper, DRAFT v2.0.
If You Can Read a Recipe, You Can Now Be a Developer
The $1K Experiment Part 2: What happens when the framework compounds. 5.5 hours to working MVP. 2,031 lines became 106,000. Shipping is addictive—here's the warning label.
The $1K Claude Code Credit: What Happens When a PM Learns to Ship
Could a senior PM with product clarity but no coding background actually build and ship real software? 31 days, 215 commits, 38K lines of TypeScript. The 64/33/3 collaboration model that made it work.
The 90-Day Death Spiral: Why 95% of AI Projects Fail
Research suggests only 5% of AI pilots deliver measurable impact. The early warning system hiding in your support tickets—and the metrics that predict failure before day 90.
The experiments shelf. Some ship, some stay strange. The research half — machine residents, explorables, research wings — lives behind one front door; these are the highlights worth a detour.
Live · New
🪵
Heartwood
Read a climate out of wood. Grow a tree's cross-section from a climate signal — earlywood, latewood, fire scars, frost rings — then crossdate a floating core against the master chronology, the trick that pins a roof beam or a Stradivarius to an exact year. Built by a former dendrochronologist, for the site named after tree rings.
A numbers-station-style broadcast documenting Ben's land parcel and where it sits in the geological time scale. Opt-in background audio — switch on the station and let it run. A slow read of the parcel down the geological column, in the flat cadence of a numbers station; the full lean-in page lives here.
A repository read like a cross-section. Commit activity drawn as tree rings — busy months are wide growth, quiet ones are narrow, and the whole history reads as grain. The dendrochronologist's habit, pointed at a git log.
Ten years at Microsoft, a 10-month gap where shipping in public became the job, and back at Microsoft since March 2026. The numbers, the case study, and the CV live below for anyone who wants them.
Background
Microsoft
–
Senior Program Manager, AI Platforms & Enterprise Operations
Across 5 years, spanning Copilot · Graph · Windows 365 · Teams Devices
Owned a portfolio of internal platforms across 8 product lines—signal systems, routing intelligence, self-service tools, and quality measurement. The infrastructure that turned customer friction into engineering action.
Rebuilt the signal-to-engineering pipeline from scratch. 95+ features shipped through this system; adopted org-wide.
Built routing intelligence that classifies incoming work by complexity and matches it to the right skill level.
Scaled self-service from pilot to ~50% adoption on flagship products. Tens of thousands of tickets per year that never get created.
Created early risk detection identifying 700+ at-risk customer situations before they escalate.
Built and shipped a recommender reaching 14K enterprise customers with 14% conversion.
Led crisis response for a 433K-user transition—near-zero churn.
Microsoft Premier Support
Built Premier Engineering from a 3-person pilot to 150 agents handling 60K incidents/year.
Managed partner programs spanning 1,000+ Office 365 migrations across 12 global partners.
Education
B.S. Interdisciplinary Science & Technology — University of Arizona
Former dendrochronologist. Yes, tree rings. It's where the domain name comes from.
Case Study: Copilot Extensibility — From Silos to Signal
Led the cross-functional effort to build enterprise AI adoption intelligence across multiple product lines. Created the signal-to-engineering pipeline that identified 76 blockers, unblocked 7,380 users, and contributed to 94,000 seats added.
A major enterprise AI rollout was accelerating across multiple product lines with no shared visibility into adoption patterns. Customers were hitting adoption walls that no single team could see. I assembled a cross-functional team, deployed real-time case analytics, and built the feedback loop that turned support signal into engineering priorities. The framework created a sustainable system for surfacing and resolving adoption blockers across the enterprise AI ecosystem.
By the Numbers (Microsoft era)
3 → 150agentsCo-founded the program, hired the team, built the playbook
700+blockersSurfaced from support signal that product teams couldn't see
95+features shipped64% of requests submitted to engineering accepted
220Kusers unblockedAdoption walls removed before they became churn events
94Kseats addedCustomers who expanded after we resolved their blockers
$355M+aggregate valueAcross retained ARR, cost avoidance, and growth enablement
If something here sparked an idea — or you want to trade notes on building weird things with Claude — reach out.
Open seat — mathematician collaborator
One seat open
Erdős needs a combinatorialist — PhD student is plenty — curious what auditable human–AI mathematics looks like from the inside. Every step we claim, you can check: walk the first provenance trail, then take the seat.