I asked Claude about my AI opinions.

Based on my transcripts and blog posts, find the three claims I make most consistently, the three I’ve quietly reversed, and the one assumption I’ve never questioned but everything depends on.

Here are things I’ve changed my opinion on:

  1. THEN: One frontier model will win - not specialization. NOW: Gemini for media, Claude for strategy/style, GPT for rigor. SLMs as tools.
  2. THEN: Carefully curate my course content. NOW: Give students prompts directly.
  3. THEN: Web apps are differentiated artifacts. NOW: HTML is easier to generate than PPT - a signal of slop, not craft.
  4. THEN: Human in the loop. NOW: Human NOT in the loop, bottlenecking it. On-the-loop, etc. is fine.
  5. THEN: Minimal single-agent loop, avoid sub-agents" NOW: Multi-agent, sub-agent, and agent teams.
  6. THEN: Avoid MCP, prefer SKILLS.md. NOW: Use MCP because integrating with Claude / ChatGPT / … is easy.

There are the top contradictions in my opinions.

  1. “Vibe code everything, end-to-end” vs “Don’t commit based on vibe-coding, it’s not fully reliable”.
  2. “Experience is a liability” vs “Domain expertise is an edge”
  3. “Paid plans are fine for private data” vs “Don’t assume paid = private”
  4. “Get out of the way / delegate everything” vs “Verification is the whole job”
  5. “Build, don’t plan” vs “Plan → Correct → Execute”

Some things, I have never questioned.

  1. For numbers, math, or correctness, make AI write and run code – never trust prose arithmetic.
  2. Judge AI against human accuracy, never against perfection – because experts disagree among themselves.
  3. Use AI heavily – reach for it first, high volume.
  4. Don’t build a foundation model from scratch – steer existing general models instead.

Things I usually say, but there are exceptions.

  1. “Always verify”. But blindly trust AI in non-core areas where you are unskilled (such as personal finance).
  2. “AI is improving fast”. But there is a jagged edge, to verify before upgrading.
  3. “Don’t build models”. But go ahead if it’s easy and beats LLMs clearly.
  4. “Models keep getting cheaper”. But not in May 2026.