Skip to content
wojciech.io
Insights

Field notes on AI systems, GTM, and revenue architecture.

Practical breakdowns from systems I build and run. Not theory: architecture, stack choices, and operator lessons from live deployments.

New23 Sept 2026 · 8 min read

Claude Opus 5.5 won the index. GPT-6 Sol won the invoice.

Opus 5.5 scores 58 to Sol's 48 and costs $5.98 per task against $1.06. Both read cache at $0.20. Where the gap is real and where it disappears.

Read the comparison →

52 articles

Every comparison I have run

Price, independent score, and which one I would put in the loop.

25 comparisons
Claude CodevsCodexcoding agent benchmark and cost16 Sept 20267 min readClaude Fable 5.1vsFable 5what changed, what breaks16 Sept 20267 min readGPT-5.6 LunavsClaude Fable 5.150x the input price for 15 points16 Sept 20266 min readGPT-5.6 TerravsFable 5.1five times the price for 11 points16 Sept 20266 min readClaude Haiku 4.5vsGPT-5.6 Lunafive times dearer, 20 points behind16 Sept 20266 min readClaude Haiku 4.5vsSonnet 5double the price, 20 points, 5x the window16 Sept 20266 min readClaude Mythos 5.1vsFable 5.1the real difference16 Sept 20265 min readClaude Opus 5vsFable 5.1double the price for under three points16 Sept 20267 min read

Playbooks

Longer pieces about systems rather than models.

23 pieces
AI Systems · 6 min read

Fable 5 scored 60 in July and 49.7 today. The model did not change. The index did.

What v4.3 measures, what changed on 7 September, the current ranking to two decimals, and why AI model scores from July cannot be compared with today's.

AI Systems · 6 min read

Add 50,000 tokens to a GPT-6 Astra prompt and the bill more than doubles. The 272K line, explained.

Above 272K input tokens, OpenAI bills the whole request at 2x input and 1.5x output. Claude keeps one rate to 1M. Worked costs for every current model.

AI Marketing · 11 min read

An AI CMO without a plan is just a very fast intern

What an AI marketing stack really takes off a CMO, and what it cannot. The open source Claude Code plugin I built, and the two files it needs first.

Operator Playbooks · 5 min read

Replatforming a large B2B website without losing the pipeline

The website rebuild I ran at Symfonia, treated as a revenue system: protect the SEO equity, keep the lead paths, tie the new site to pipeline.

Products · 5 min read

The browser bookmark is the wrong primitive

Why I built Kade as a local-first link vault: not another bookmark list, but a private queue for links that need context before they are opened again.

AI Systems · 6 min read

How I use Claude Code in client GTM work

Beyond the outbound agent stack: how Claude Code runs content operations, CRM hygiene, and research inside real client engagements: what the delivery model looks like from the outside.

Newsletter

Want to know more first?

New articles on AI systems, GTM architecture, and growth operating models, delivered when they publish.

Subscribe