Paul Welty, PhD AI, WORK, AND STAYING HUMAN

What shipped today

Quick check-in session to review the overnight pipeline results. The pipeline ground through all 20 issues that were loaded yesterday evening across “Adaptive lesson delivery” and “First module: Delegation” milestones, landing 19 of 20 — only GH-148 (breadcrumb nav on learner module page) was tagged human-needed and left open.

The overnight run was remarkably clean: +1,628 lines across 33 files, including 12 new migrations, new components for assessment responses with AI feedback (Anthropic SDK), DISC indicators, loading skeletons, progress tracking, and full DISC-adapted content for the delegation module. Both milestones are effectively complete.

One fix from the previous session also landed in this window: we discovered and fixed a pipeline merge artifact where the old [moduleId] route conflicted with the new [id] route, and duplicate auth code in the demo-requests page left behind by the pipeline’s GH-120 merge.

Completed

Pipeline overnight (19 issues):

  • GH-128 — Wire goals list into learner module detail page
  • GH-129 — Render assessments as standalone section on lesson page
  • GH-130 — Add assessment response textarea and submission for scenario questions
  • GH-131 — Add AI-powered assessment feedback using Claude
  • GH-132 — Show DISC adaptation indicator on lesson page
  • GH-133 — Auto-track lesson progress on first visit
  • GH-134 — Add module completion summary when all lessons done
  • GH-135 — Add loading skeleton for lesson page
  • GH-136 — Add loading skeleton for module detail page
  • GH-137 — Connect “Practice this skill” button to module-specific simulation
  • GH-138 — Generate DISC-adapted content units for lesson 1
  • GH-139 — Generate DISC-adapted content units for lesson 3
  • GH-140 — Add I-style and C-style examples to all three delegation lessons
  • GH-141 — Create assessment for lesson 1
  • GH-142 — Create assessment for lesson 3
  • GH-143 — Add diagnostic pre-assessment to delegation module
  • GH-144 — Clear legacy content field after content units are populated
  • GH-145 — Add practice task content units with guided reflection prompts
  • GH-147 — Add error boundary and retry UI for lesson page data fetching

Carry-over

  • GH-148 — Add breadcrumb navigation to learner module detail page (tagged human-needed)
  • Both “Adaptive lesson delivery” and “First module: Delegation” milestones can be closed once GH-148 is resolved

Risks

  • No end-to-end testing of the learner flow has been done after 19 PRs landed overnight — should test the full journey (module → lesson → assessment → feedback) in Chrome before declaring milestones complete

Flags and watch-outs

  • The Anthropic SDK was added as a dependency (GH-131) — need to ensure ANTHROPIC_API_KEY is set in production environment
  • 12 new migrations landed — verify they apply cleanly against production/staging Supabase

Next session

  • Fix GH-148 (breadcrumb nav) manually — it’s the only open issue, tagged human-needed
  • End-to-end test the learner flow in Chrome: sign up → DISC assessment → module detail → lesson → assessment response → AI feedback → completion banner
  • Close “Adaptive lesson delivery” and “First module: Delegation” milestones once verified
  • Run /scout to find quality issues in the 19 pipeline-merged PRs — the pipeline occasionally leaves merge artifacts (as we saw with the [moduleId] conflict)
  • Plan the next milestone — likely “Learner analytics and reporting” or “Multi-module curriculum”

Why customer tools are organized wrong

This article reveals a fundamental flaw in how customer support tools are designed—organizing by interaction type instead of by customer—and explains why this fragmentation wastes time and obscures the full picture you need to help users effectively.

Infrastructure shapes thought

The tools you build determine what kinds of thinking become possible. On infrastructure, friction, and building deliberately for thought rather than just throughput.

Server-side dashboard architecture: Why moving data fetching off the browser changes everything

How choosing server-side rendering solved security, CORS, and credential management problems I didn't know I had.

The work of being available now

A book on AI, judgment, and staying human at work.

The practice of work in progress

Practical essays on how work actually gets done.

Junior engineers didn't become profitable overnight. The work did.

We've been celebrating that AI made junior engineers profitable. That's not what happened. AI made it economically viable to give them access to work that actually builds judgment, work we always knew

Three projects, three opposite methods, all monster output days: what that taught me about when process helps and when it's just comfort

I've been running a portfolio of software projects using a mix of autonomous AI pipelines and human-led parallel agent sessions. Yesterday, three different projects had monster output days — and th...

What happens when the pipeline doesn't need you

So here's something I noticed today that I want to sit with. I run several projects that use autonomous pipelines — AI systems that pick up tasks, write code, open pull requests, ship changes. One ...