Demo data

These ~120 responses are synthetic, generated in your browser from a fixed seed. No real people answered, and nothing here is saved. It’s the same dashboard buyers get, with the same rules: groups under 5 are hidden, and the CSV has pseudonymous ids only.

Example survey

AI coding assistants: the wait

Demo data120 of 150 completesSynthetic responses, generated in your browser. Nothing here is real.

Completed: 120; In the results below
Screened out: 19; Didn’t qualify; not in results
Still answering: 4; Started, not finished yet
Left out for quality: 3; Failed a quality check; replaced at no charge

Filter by audience

Showing all 120 responses

Filters use audience fields only and run on our server: a filter has to match at least 10 responses, and one that leaves out fewer than 5 people is hidden too (comparing it with all results would show what they said). Counts are rounded to the nearest 5. The CSV has every delivered response with pseudonymous ids, no profile data, and open text with names and contact details removed. Answers flagged as too fast are left out of filtered results on our server. Two filters whose responses differ by fewer than 5 people can’t both be shown.

Insights

AI summary

Key findings with numbers, themes with real quotes, and what to be careful about. It only reads totals, breakdowns with at least 5 people per cell, and answers with names removed.

Q1Single choiceScreener

Do you use an AI coding assistant at least once a week?

120 answered

Only people who qualified finish the survey, so everyone counted here passed this screener. 19 people screened out and aren’t included.

Yes100% · 120
No0% · 0

Q2Single choice

Which one do you use most?

120 answered

Cursor35% · 42
Claude Code23% · 28
GitHub Copilot23% · 27
Codex7% · 8
Windsurf11% · 13
Other2% · 2

Break down by

Q3Multiple choice

What do you usually do while it works?

120 answered · people could pick more than one, so shares add up to more than 100%

Watch it28% · 34
Review earlier output36% · 43
Switch to other work47% · 56
Check my phone43% · 51
Something else12% · 14

Break down by

Q4Scale

How happy are you with how fast it is?

120 answered

Mean 3.5Median 3on a 1–5 scale

1 (Very unhappy): 5 (4%)
2: 10 (8%)
3: 46 (38%)
4: 35 (29%)
5 (Very happy): 24 (20%)

Break down by

Q5Single choice

Who pays for it?

120 answered

My employer44% · 53
I do32% · 38
Free tier10% · 12
Mix14% · 17

Break down by

Q7Number

About how many minutes a day do you wait on it?

120 answered

Average: 19.3 minutes
Median: 20 minutes
Lowest: 5 minutes
Highest: 39 minutes

Q8Grid

How does it do on each of these?

120 answered

Speed · 120 answered

Poor10% · 12
OK31% · 37
Good36% · 43
Great23% · 28

Accuracy · 120 answered

Poor11% · 13
OK27% · 32
Good43% · 52
Great19% · 23

Code quality · 120 answered

Poor14% · 17
OK33% · 39
Good38% · 46
Great15% · 18

Q9Ranking

Rank what matters most

120 answered

1. Speedaverage rank 1.3 · 82% first
2. Accuracyaverage rank 2.0 · 14% first
3. Priceaverage rank 3.0 · 3% first
4. Privacyaverage rank 3.8 · 1% first

Break down by

Q10Constant sum

Split 100 points across what you want improved

120 answered

Speed50.2 of 100 · 50%
Price25.3 of 100 · 25%
Accuracy24.5 of 100 · 25%

Q11MaxDiff

Which feature matters most, and which least?

120 answered

1. Long contextscore +0.89 · 192 best · 0 worst · shown 216
2. Code reviewscore +0.34 · 77 best · 3 worst · shown 216
3. Team sharingscore +0.13 · 70 best · 41 worst · shown 216
4. Offline modescore -0.47 · 18 best · 120 worst · shown 216
5. Voice inputscore -0.89 · 3 best · 196 worst · shown 216

Score = (times picked best − times picked worst) ÷ times shown, from −1 to +1.

Q12Concept test

How appealing is this plan?

120 answered

Pro monthlymean 3.3 of 5 · 43% rate it 4–5 · n 60
Pay as you gomean 4.0 of 5 · 67% rate it 4–5 · n 60

Q13Price sensitivity

Think about a faster “priority” plan for your assistant.

120 answered

Optimal price: USD 17.50
Indifference price: USD 20.43
Acceptable from: USD 13.13
Acceptable up to: USD 29.46

Too cheap
Bargain
Expensive
Too expensive

Table view

Price	Too cheap	Bargain	Expensive	Too expensive
USD 3	100%	100%	0%	0%
USD 4	98%	100%	0%	0%
USD 5	92%	100%	0%	0%
USD 6	83%	100%	0%	0%
USD 7	76%	100%	0%	0%
USD 8	67%	100%	0%	0%
USD 9	57%	99%	0%	0%
USD 10	47%	98%	0%	0%
USD 11	38%	92%	0%	0%
USD 12	32%	88%	0%	0%
USD 13	27%	76%	0%	0%
USD 14	17%	66%	1%	0%
USD 15	3%	57%	1%	0%
USD 16	0%	53%	1%	0%
USD 17	0%	45%	2%	0%
USD 18	0%	34%	5%	0%
USD 19	0%	28%	11%	0%
USD 20	0%	22%	17%	1%
USD 21	0%	18%	24%	2%
USD 22	0%	13%	28%	3%
USD 23	0%	10%	34%	4%
USD 24	0%	3%	39%	6%
USD 25	0%	0%	46%	8%
USD 26	0%	0%	57%	11%
USD 27	0%	0%	68%	14%
USD 28	0%	0%	75%	18%
USD 29	0%	0%	77%	18%
USD 30	0%	0%	82%	24%
USD 31	0%	0%	89%	29%
USD 32	0%	0%	91%	35%
USD 33	0%	0%	92%	40%
USD 34	0%	0%	95%	52%
USD 35	0%	0%	95%	59%
USD 36	0%	0%	97%	66%
USD 37	0%	0%	100%	71%
USD 38	0%	0%	100%	75%
USD 39	0%	0%	100%	78%
USD 40	0%	0%	100%	79%
USD 41	0%	0%	100%	82%
USD 42	0%	0%	100%	88%
USD 43	0%	0%	100%	92%
USD 44	0%	0%	100%	94%
USD 45	0%	0%	100%	98%
USD 47	0%	0%	100%	100%

Shaded: the range of acceptable prices (where “too cheap” meets “not a bargain”, to where “too expensive” meets “not expensive”). The optimal price is where “too cheap” and “too expensive” cross.

Q14Slider

How much faster would it need to be to switch?

120 answered

Average: 39.5 %
Median: 39 %
Lowest: 10 %
Highest: 70 %

Q15Date

When did you last change assistants?

120 answered

Earliest: 2026-01-02
Median: 2026-04-25
Latest: 2026-09-27

2026-019% · 11
2026-0214% · 17
2026-0318% · 22
2026-048% · 10
2026-0511% · 13
2026-0612% · 14
2026-078% · 9
2026-089% · 11
2026-0911% · 13

Response quality

Completion times are rounded to the hour to protect respondents. While you watch live, a new response still shows up within seconds of being finished, so its timing is only approximate privacy, not anonymity. Respondents are pseudonymous; please don’t try to identify anyone.

Answered while waiting on their AI: 80%; Share of answers given during a wait, when people had nothing else to do.
Median time per answer: 9.7s; Timed by our server from Oct 9, 2026; earlier answers used the respondent’s own clock.
Flagged too fast: 28; answers from 27 respondent(s), answered faster than the question can be read. Left out of filtered results on our server; the overall charts and the CSV include every delivered answer.

Responses that fail a quality check are removed before results reach you, aren’t billed, and are replaced while your quota is open.

Want results like this for your question?

Describe your goal and get a drafted survey, a price and a reach estimate. A person reviews it before it goes live.

Draft a survey Ask us instead

AI coding assistants: the wait

Filter by audience

Insights

Do you use an AI coding assistant at least once a week?

Which one do you use most?

What do you usually do while it works?

How happy are you with how fast it is?

Who pays for it?

What one change would make waiting on it less painful?

Themes

About how many minutes a day do you wait on it?

How does it do on each of these?

Speed · 120 answered

Accuracy · 120 answered

Code quality · 120 answered

Rank what matters most

Split 100 points across what you want improved

Which feature matters most, and which least?

How appealing is this plan?

Think about a faster “priority” plan for your assistant.

How much faster would it need to be to switch?

When did you last change assistants?

Want results like this for your question?

AI coding assistants: the wait

Who answered

Filter by audience

Insights

Themes

Speed · 120 answered

Accuracy · 120 answered

Code quality · 120 answered

Want results like this for your question?