Claude CampResearch

Who actually learns Claude Code

Before every cohort we ask people what they can already do. Here is what 35 of them said, and the short version is that almost none of them are the person you would picture.

n = 35 · self-reported · August 2026

// the finding

The people paying to learn Claude Code mostly do not code

60%
had never opened Claude Code, or tried it once
69%
rarely or never use a terminal
9%
are professional engineers

Claude Code is a command-line tool, so the intuitive assumption is that the people learning it are software engineers. In our rooms that is not close to true. Three in five have barely touched it, seven in ten are not comfortable in a terminal, and fewer than one in ten writes software for a living.

They are not confused about what they signed up for. Claude Code was the second most requested thing on the list. They want the tool precisely because they cannot already use it.

// what they could already do

Comfortable in the chat app, beginners everywhere else

Self-rated from 1 (never touched it) to 5 (use it daily). The gap between claude.ai and everything else is the whole story: people arrive fluent in the chat box and near-zero on the surfaces where the leverage actually is.

SurfaceMeanDistribution 1 to 5Never or once
claude.ai (the chat app)3.94
17%
Claude Cowork2.69
51%
Claude Code2.37
60%
The terminal / command line2.26
69%
Coding in general2.11
69%

Left to right in each bar: rated 1, 2, 3, 4, 5. Darker is more fluent.

// who they are

Mostly founders and operators, and a third who fit nothing

Other (did not fit our categories)
11 · 31%
Founder / Entrepreneur
10 · 29%
Product / Marketing / Ops operator
7 · 20%
Solopreneur / Indie builder
3 · 9%
Professional engineer
3 · 9%
Career-switcher / self-taught
1 · 3%

The largest single answer is “Other”, which is a finding about our question rather than about them. Nearly a third of people did not see themselves in a list that included founder, engineer, operator, solopreneur and career-switcher. We are keeping the free-text answers to rebuild the categories rather than quietly rounding people into the nearest box.

// what they want

They want work that runs without them, more than they want a tool

Automations / Routines
25 · 71%
Claude Code
21 · 60%
Cowork
18 · 51%
The operator workflow
18 · 51%
Claude Design
14 · 40%
MCP
13 · 37%
Agent SDK
10 · 29%

Automations and routines beat every named product, including Claude Code itself. Read alongside the familiarity numbers, the request is consistent: an audience that does not write software is asking for work that happens on a schedule without them, and reaching for a developer tool to get it.

Agent SDK and MCP sit at the bottom. Those are the two most discussed things in the developer conversation about Claude and the two least requested by the people actually buying training.

// method, and what this does not show

35 responses to the intake form every enrollee is asked to complete before their cohort, collected through August 2026. Familiarity is self-rated on a 1 to 5 scale. The most-wanted question allows multiple answers, so those shares sum to more than 100%.

The limits, stated plainly. This is a sample of people who chose to pay for live Claude training, so it describes that population and not Claude Code users generally. It is self-reported, and people are not reliable judges of their own skill. 35is enough to see a 60/9 split clearly and nowhere near enough to split the data by role. It is one program's rooms, and our marketing reaches operators more than engineers, which certainly shapes who shows up.

Aggregate only. No names, no email addresses, and no free-text answers are published here or shared.

Numbers are regenerated from the source data rather than typed by hand. If you want to cite or challenge any of this, the fastest route is to ask directly.

This data comes from the intake form for the Claude Camp 2-Day Intensive, a live cohort across the whole Claude stack. If you are weighing it against the alternatives, we keep an honest comparison that includes when to buy something else.