Work log: 2026-04-03
What shipped today
The big win today was slashing the Vercel bill. An audit of the deployment configuration revealed authexis was running Turbo build machines ($0.126/min) with auto-deploy on every git push — roughly 22 production builds per day. Combined with three other projects on the same team (eclectis, prakta, dinly), the estimated monthly cost was ~$99 in build minutes alone. Two changes — switching to Standard machines ($0.014/min) and setting the ignored build step to exit 0 so deploys only happen via vercel --prod — should cut that to ~$3.43/month. The /ship skill was updated to include a vercel --prod step so deploys still happen at natural ship points without auto-triggering on every push.
On the agency side, Synaxis AI’s onboarding kicked off. The product session fielded two rounds of questions — Sydney’s initial product interview (8 questions about positioning, voice, pricing, competitors, tone) and Trina’s review of the client onboarding document. Both were answered from PRODUCT.md and DECISIONS.md. The onboarding doc was confirmed factually accurate; four open business questions were flagged for Paul’s input (engagement scope, Synaxis branding connection, testimonials, landing page refresh).
Also filed diktura#101 for linking its Next.js frontend to Vercel — it’s supposed to be deployed there but was never connected.
Completed
- #1918: Vercel manual deploy optimization (authexis settings applied, cross-project issue filed)
- diktura#101: Link Next.js frontend to Vercel (filed)
- Synaxis agency onboarding — product interview and doc review completed
/shipskill updated with Vercel deploy step
Release progress
- v1.5: 1 open / 49 closed (#743 dashboard redesign, in backlog)
- v2.0 (Product simplification): 25/25 closed
- v2.1 (Content model refactor): 15/15 closed
- v2.2 (Simplified pipeline): 3/3 closed
Carry-over
- #1916: Curated briefing option — still needs Paul’s clarification
- #1913: White-label briefings — blocked on #1916
- Vercel settings changes needed on eclectis, prakta, dinly (dashboard-only, Paul must do manually)
- Synaxis onboarding doc has 4 open questions for Paul (engagement scope, Synaxis branding, testimonials, landing page)
Risks
No new risks. Existing open risks unchanged.
Flags and watch-outs
- The
exit 0ignored build step means no code reaches production until someone runsvercel --prod. If the/shipskill is skipped or the session closes without shipping, the deployed version could lag behind main. This is intentional (cost savings) but worth being aware of. - Diktura’s repo has legacy Rails code (Gemfile, config.ru, Dockerfile) that is NOT active — it’s reference material. The actual app is Next.js in
web/+ Python engine. - The eclectis, prakta, and dinly Vercel settings changes are manual dashboard work. Until applied, those projects are still burning Turbo-rate build minutes on every push.
Next session
- Apply Vercel settings (Standard +
exit 0) to eclectis, prakta, dinly — remind Paul if not done - Answer #1916 (curated briefing) so #1913 can unblock
- Answer Synaxis onboarding open questions (engagement scope, branding, testimonials, landing page)
- Consider closing v1.5 — #743 (dashboard redesign) is the last issue, sitting in backlog
- Run
/scoutto surface new work — queues are empty
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.
The inbox nobody reads is the one that matters
Every organization has a monitoring system that works perfectly and reports to nobody. The gap between having information and acting on it is where most failures actually live.
The best customers are the first ones you turn against
Every subscription makes a bet that most customers won't use what they're paying for. The customer who closes that gap becomes a problem to be managed.
Delegation without comprehension is just prayer
The organizations that survive won't be the ones that automated the most. They'll be the ones that figured out what to stop delegating.