Current

What has my attention right now

A snapshot of what I’m building, learning and puzzling over. Updated when things meaningfully change, not on a schedule.

Last updated September 2026 · Bengaluru

Notes

  1. This site now carries VeraKeep, GrowthOS, five specialist AI skills, and Selling What's Complicated — evidence of one operating philosophy, not four unrelated projects.

  2. Spine was renamed VeraKeep. Older articles on this site that mention Spine describe an earlier stage of the same project.

  3. Turning the Gulfood letters of intent into a smaller set of real, structured conversations. Fewer, deeper is the right instinct here.

Publishing

Selling What's Complicated

Production-final and gate-passed; KDP Previewer and a physical proof are still pending before an Amazon listing goes live. No ISBN or ASIN exists yet. The free Field Kit is live at /fieldkit — the book already prints that URL with a QR code.

Building

VeraKeep

In working use, including in the production of this website. Formerly Spine, renamed in August 2026. Released under Apache-2.0; not offered as a priced commercial product.

Developing

GrowthOS

Read-only R&D: it diagnoses and prepares actions, but every write to a production website, listing or campaign still requires explicit human review.

Codifying

Specialist AI skills

DecisionOS, GlobalDive, Humaniser, the Context-Aware Hook Writer and HyperFrames AI + Technology WOW — turning working methods into reusable, evidence-governed AI skills.

Continuing

NxSync ERP

Growing the working multi-tenant foundation into its first business modules — deliberately, one reviewable stage at a time. Not commercially launched.

T57

Developing the early relationships from Gulfood into structured commercial plans, and continuing the market, pricing and partnership work that has to be ready before the platform is. Pre-launch.

Learning

  • How AI changes the way a business should be run — as a daily operating question, not a slogan.
  • Enterprise-platform architecture for multi-tenant, AI-native systems.
  • Market-entry economics for emerging-market trade platforms.

Reading

  • Notes and papers on how far autonomous agents can run before they need a human checkpoint.
  • The economics of purchasing-power-adjusted pricing in cross-border trade.

Open questions

  • What does a genuinely AI-native ERP look like once the marketing language is stripped away?
  • How large can a piece of autonomous AI work safely be before it needs a human checkpoint?
  • What should a founder keep doing personally, and what should the system do?

If any of this overlaps with what you’re working on, I’d like to hear from you.