Skip to main content

tech radar · v2 · october 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. 96 blog posts written since May 2025, distilled into the bets I am willing to defend out loud.

last reviewed October 4, 2026 · click any blip or any item below for the argument behind it

ADOPTTRIALASSESSHOLDEngineering PracticesAI Tools & ArchitectureLeadership PosturesIndustry Narratives123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990
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
  • 08Count the work for everyone
  • 09Name an owner for every agent
  • 10Separation of duties for agents
  • 11Context Spine / Intent Stack
  • 12Enforced design systems, not voluntary ones

Trial

In production. Convinced of direction.

  • 13Answer-key-first modernization
  • 14Operability in the spec
  • 15Spec review as the new code review
  • 16Multi-agent specialist workflows
  • 17Prompt review as quality gate
  • 18Workflow inversion over agent loops

Assess

Watching. Not yet bet on.

  • 19Dark factory model for software
  • 20Agent workflows as versioned releases

Hold

Refuse, or warn against.

  • 21Vibe coding without specs
  • 22Kubernetes for one service
  • 23Agent teams as a default
  • 24Instructions as access control
  • 25Line-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.

  • 26Claude Code as build system
  • 27Hybrid local + frontier inference
  • 28MCP as integration substrate
  • 29AI providers like cloud providers
  • 30Design for sovereign model risk
  • 31Own the learning loop, not the model
  • 32Opinionated defaults, explicit escape hatches

Trial

In production. Convinced of direction.

  • 33Open-weight models for regulated workloads
  • 34Personal agents for everyday follow-up
  • 35Custom-trained frontier models (Nova Forge style)
  • 36Design the platform for the agent persona
  • 37Context as a platform capability
  • 38Invisible AI over the portal chatbot
  • 39Keep the agent replaceable
  • 40LLM-consumable design systems

Assess

Watching. Not yet bet on.

  • 41Intent interface over the golden path
  • 42Fully autonomous coding agents
  • 43Delegated identity for agents
  • 44AI design tools beyond Figma
  • 45Decision models for typed judgments

Hold

Refuse, or warn against.

  • 46Multi-cloud as a hedge
  • 47Prompt-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.

  • 48Stay close to the craft
  • 49Principles over personalities
  • 50Accountability flows upward
  • 51Clarity beats consensus
  • 52Turn complexity into leverage
  • 53Servant leadership in tech
  • 54Invest in framers, not finishers
  • 55Judgment as the scarce resource
  • 56Bargain honestly with disruption
  • 57Value per verified outcome

Trial

In production. Convinced of direction.

  • 58FinOps as a core engineering competency
  • 59Build for succession
  • 60Designed apprenticeship over accidental
  • 61The AI Operating Ledger
  • 62Spend the saved hour on purpose
  • 63The developer experience ledger

Assess

Watching. Not yet bet on.

  • 64Pairing on specs and reviews

Hold

Refuse, or warn against.

  • 65Mandate-driven AI adoption
  • 66Building monuments to yourself
  • 67Headcount-first AI reorgs
  • 68Agent time as a success metric
  • 69Personal agents as AI strategy
  • 70PowerPoint-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.

  • 71"Specs beat vibes"
  • 72"The bottleneck is never the stack"
  • 73"Agents advise; humans decide"
  • 74"From toil to judgment"
  • 75"Trust the gate, not the actor"

Trial

In production. Convinced of direction.

  • 76"Trust compounds faster than velocity"
  • 77The builder’s bargain
  • 78"The frame is the bottleneck"
  • 79Open weights as commoditization strategy

Assess

Watching. Not yet bet on.

  • 80AI infrastructure capex bubble
  • 81Open-source funding under AI
  • 82Capability as a governed resource
  • 83The rewrite taboo was a price tag
  • 84Intelligence as infrastructure
  • 85Safety rules as incumbent moats

Hold

Refuse, or warn against.

  • 86"Workforce optimization" language
  • 87"AI is coming for your job"
  • 88"AI replaces engineers"
  • 89The salary-for-subscription swap
  • 90Vendor-as-operating-model