Skip to main content

tech radar · v2 · august 2026

What I'm betting on, and what I'm not

Years ago I launched my first company's Tech Radar to give engineering teams a shared formulary for what to adopt, trial, assess, and hold. This is the personal version. Seventy-one blog posts and one year of writing, distilled into the bets I am willing to defend out loud.

last reviewed August 1, 2026 · click any blip or any item below for the argument behind it

ADOPTTRIALASSESSHOLDEngineering PracticesAI Tools & ArchitectureLeadership PosturesIndustry Narratives1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374
Adopt — use today. default in new workTrial — in production. convinced of directionAssess — watching. not yet bet onHold — refuse, or warn against

quadrant 01

Engineering Practices

How software actually gets built, reviewed, and shipped in an agentic SDLC.

Adopt

Use today. Default in new work.

  • 01Spec-driven development
  • 02Standards in files, not heads
  • 03Architecture as agent guardrails
  • 04Pipeline-first delivery
  • 05Verification-first development
  • 06Fail to unknown, never to green
  • 07Conformance suite as executable oracle
  • 08Name an owner for every agent
  • 09Separation of duties for agents
  • 10Context Spine / Intent Stack
  • 11Enforced design systems, not voluntary ones

Trial

In production. Convinced of direction.

  • 12Spec review as the new code review
  • 13Multi-agent specialist workflows
  • 14Prompt review as quality gate
  • 15Workflow inversion over agent loops

Assess

Watching. Not yet bet on.

  • 16Dark factory model for software
  • 17Cross-team multi-agent workflow sharing

Hold

Refuse, or warn against.

  • 18Vibe coding without specs
  • 19Kubernetes for one service
  • 20Agent teams as a default
  • 21Line-by-line review as primary quality bar

quadrant 02

AI Tools & Architecture

The seam between local and frontier, the patterns that compound, the ones that do not.

Adopt

Use today. Default in new work.

  • 22Claude Code as build system
  • 23Hybrid local + frontier inference
  • 24MCP as integration substrate
  • 25AI providers like cloud providers
  • 26Design for sovereign model risk
  • 27Own the learning loop, not the model

Trial

In production. Convinced of direction.

  • 28Open-weight models for regulated workloads
  • 29Persistent personal AI agents (Roci)
  • 30Custom-trained frontier models (Nova Forge style)
  • 31LLM-consumable design systems

Assess

Watching. Not yet bet on.

  • 32Fully autonomous coding agents
  • 33MCP gateway / RBAC patterns
  • 34AI design tools beyond Figma

Hold

Refuse, or warn against.

  • 35Multi-cloud as a hedge
  • 36Prompt-whispering as a discipline

quadrant 03

Leadership Postures

How engineering leaders should show up, what they should reward, and what they should refuse.

Adopt

Use today. Default in new work.

  • 37Stay close to the craft
  • 38Principles over personalities
  • 39Accountability flows upward
  • 40Clarity beats consensus
  • 41Turn complexity into leverage
  • 42Servant leadership in tech
  • 43Invest in framers, not finishers
  • 44Judgment as the scarce resource
  • 45Bargain honestly with disruption
  • 46Value per verified outcome

Trial

In production. Convinced of direction.

  • 47FinOps as a core engineering competency
  • 48Build for succession
  • 49The AI Operating Ledger

Assess

Watching. Not yet bet on.

  • 50The apprenticeship problem
  • 51Pairing on specs and reviews

Hold

Refuse, or warn against.

  • 52Mandate-driven AI adoption
  • 53Building monuments to yourself
  • 54Headcount-first AI reorgs
  • 55Personal agents as AI strategy
  • 56PowerPoint-only leadership

quadrant 04

Industry Narratives

The stories the industry tells itself. Some are true. Most are someone's quarterly earnings call dressed up in a hoodie.

Adopt

Use today. Default in new work.

  • 57"Specs beat vibes"
  • 58"The bottleneck is never the stack"
  • 59"Agents advise; humans decide"
  • 60"From toil to judgment"
  • 61"Trust the gate, not the actor"

Trial

In production. Convinced of direction.

  • 62"Trust compounds faster than velocity"
  • 63The builder’s bargain
  • 64"The frame is the bottleneck"

Assess

Watching. Not yet bet on.

  • 65AI infrastructure capex bubble
  • 66Open-source funding under AI
  • 67Capability as a governed resource
  • 68The rewrite taboo was a price tag
  • 69Intelligence as infrastructure

Hold

Refuse, or warn against.

  • 70"Workforce optimization" language
  • 71"AI is coming for your job"
  • 72"AI replaces engineers"
  • 73The salary-for-subscription swap
  • 74Vendor-as-operating-model