Stillpeak · Field NotesThinking on AI and analytics

Field
Notes.

Thinking on AI, analytics, and what it actually takes to redesign how teams work. Written from inside the work, not above it.

01Written by Sean↓·02Worth reading↓

— all entries

No. 01·FoundationsMay 2026

How AI handles the unfamiliar.

Every AI model is built from a fixed snapshot of the past. Understanding where that boundary sits — and what happens when you push past it — is one of the most practical things you can learn about these tools.

48 pp. · Dot Grid · 7 min readRead →

— worth reading

Field Link·AI Productivity Research

2026

What is the impact of AI on productivity?

Individual task speedups are real. Firm-level productivity gains are not — because the workflow, the review cycles, the coordination overhead haven't changed. Giving a team faster tools doesn't make the team faster. That's the distinction Stillpeak is built around.

Via Alex ImasRead ↗

Field Link·AI Productivity Research

2026

AI coding productivity paradox: 93% adoption, 10% gains

Developers using AI coding tools were 19% slower than those working without them — but believed they were 20% faster. That gap between felt speed and actual output is where most AI enablement efforts quietly fail, and it's not a tool problem.

Via Philipp DubachRead ↗

Field Link·AI Productivity Research

Feb 2026

Thousands of CEOs admit AI had no impact on employment or productivity

An NBER study of 6,000 executives across four countries found the vast majority see little measurable impact from AI on their operations — and economists are calling it the new Solow Paradox. The productivity is theoretically there. The organizations haven't caught up.

Via FortuneRead ↗

Field Link·AE Ecosystem

May 2026

AI-assisted analytics engineering: Docusign's framework for scaling dbt unit testing

Docusign cut dbt unit test authoring from 5 hours to 30 minutes — not by finding better prompts, but by building a framework around how AI gets used. The headline is the time savings. The story is the workflow design.

Via dbt BlogRead ↗

Field Link·Perspective

May 2026

The bonfire of our vanities

Benn asks whether the current frenzy of AI-powered building is actually producing anything meaningful, or whether it's an addictive loop of novelty that will look emptier in hindsight. A useful mirror for anyone claiming transformation — including us.

Via Benn StancilRead ↗

Field Link·AE Ecosystem

Jun 2026

How Anthropic enables self-service data analytics with Claude

Anthropic automated 95% of their internal analytics queries — not by pointing Claude at a data warehouse, but by encoding workflow into skills. Without them, accuracy sat at 21%. With them, it jumped to 95%+. The infrastructure matters less than the procedural knowledge built around it.

Via AnthropicRead ↗

Field Link·AI Productivity Research

Feb 2026

55% of CEOs who fired people 'because of AI' already regret it

HBR found 60% of companies cut headcount anticipating AI would replace workers — only 2% based on actual deployment. Forrester found 55% of those who did already regret it. AI amplifies output by 14–126% depending on the task; the mistake is treating it as a headcount equation before the enablement work is done.

Via Avi GoldfingerRead ↗

Field Link·AI Costs & ROI

Aug 2025

The GenAI divide: why 95% of AI investments fail

Despite $30–40B in corporate AI spend, 95% of enterprise pilots never make it from testing to production. The 5% that do share one trait: they embed AI into core workflows rather than layering it on top of existing operations. The investment isn't the bottleneck. The redesign work is.

Via Sundeep TekiRead ↗

FieldNotes.

How AI handles the unfamiliar.

What is the impact of AI on productivity?

AI coding productivity paradox: 93% adoption, 10% gains

Thousands of CEOs admit AI had no impact on employment or productivity

AI-assisted analytics engineering: Docusign's framework for scaling dbt unit testing

The bonfire of our vanities

How Anthropic enables self-service data analytics with Claude

55% of CEOs who fired people 'because of AI' already regret it

The GenAI divide: why 95% of AI investments fail

Field
Notes.