In this survey, you will review a few short AI-generated code snippets and indicate how you would handle each one in a work setting. Each snippet comes with a small set of basic tests that pass on typical inputs.
The survey starts with a few background questions, followed by four short coding scenarios. It takes approximately 20 minutes. All responses are anonymous.
Note: all datetimes in the code do not contain any timezone or offset information, and are pre-validated upstream. Function inputs are assumed to be well-formed unless otherwise stated.
As a thank-you for your time, I will raffle five Amazon vouchers (€20 each) among all participants who complete the survey.
How many years of professional software development experience do you have?
How often do you use AI coding assistants (e.g., GitHub Copilot, JetBrains AI Assistant, ChatGPT for coding)?**
How often do you work with legacy codebases (mature production systems with existing dependencies)?
Select up to 3 primary programming language?
Which programming languages are you comfortable reading and evaluating? (Select all that apply)
Q7. How would you rate your Python proficiency?
You work on a major publishing platform that has been live for 8 years.The platform serves several editorial sections. Your team owns the publishing pipeline. Last month, a snippet bug pushed wrong previews into the morning newsletter, and the team had to send out a follow-up correction to readers. You used an AI assistant to add a function that shortens article text to a maximum character limit, cutting at a word boundary if the text is too long.
You work on a mature appointment scheduling platform that has been live for 7 years. The platform handles bookings across multiple regions. Your team owns the calendar and scheduling logic. A slot bug last quarter caused double-bookings, and support spent days reaching out to affected users. You used an AI assistant to add a function that calculates available booking slots within a requested time window, stepping through it at fixed slot-sized intervals.
You work on a mature community platform that has been live for 9 years. The platform has a large active user base across web and mobile. Your team owns the live activity feed (showing the past few hours of events). A time-range bug last year caused events to disappear from user feeds, and the team had to backfill them by hand. You used an AI assistant to add a function that filters activity events to those within a specified time window.
You work on a mature content moderation platform that has been live for 7 years. The platform processes content tags across many sites. Your team owns the tag filtering module. A few months ago, a filtering bug let flagged tags slip through, and the team had to clean them up by hand. You used an AI assistant to add a function that removes flagged tags from a single post's tag list in place.
You work on a major publishing platform that has been live for 8 years.The platform serves several editorial sections. Your team owns the publishing pipeline. Last month, a snippet bug pushed wrong previews into the morning newsletter, and the team had to send out a follow-up correction to readers. You used an AI assistant to add a function that shortens article text to a maximum character limit, cutting at a word boundary if the text is too long.
You work on a mature appointment scheduling platform that has been live for 7 years. The platform handles bookings across multiple regions. Your team owns the calendar and scheduling logic. A slot bug last quarter caused double-bookings, and support spent days reaching out to affected users. You used an AI assistant to add a function that calculates available booking slots within a requested time window, stepping through it at fixed slot-sized intervals.
You work on a mature community platform that has been live for 9 years. The platform has a large active user base across web and mobile. Your team owns the live activity feed (showing the past few hours of events). A time-range bug last year caused events to disappear from user feeds, and the team had to backfill them by hand. You used an AI assistant to add a function that filters activity events to those within a specified time window.
You work on a mature content moderation platform that has been live for 7 years. The platform processes content tags across many sites. Your team owns the tag filtering module. A few months ago, a filtering bug let flagged tags slip through, and the team had to clean them up by hand. You used an AI assistant to add a function that removes flagged tags from a single post's tag list in place.
You work on a new publishing project your team is building. The platform isn't open to readers - it's an internal experiment running for a few weeks. Your team is small, and the publishing pipeline is still under development.
If the team decides the direction isn't worth pursuing, the code will be thrown away. You used an AI assistant to generate a function that shortens article text to a maximum character limit, cutting at a word boundary if the text is too long.
You work on a new appointment scheduling app your team is building. No users are on the platform - it's an internal experiment to see if this approach is worth pursuing. Your team is small, and the calendar logic is being built from scratch. Whether anything from this experiment ships depends on what the team learns. You used an AI assistant to generate a function that calculates available booking slots within a requested time window, stepping through it at fixed slot-sized intervals.
You work on a new community app your team is building. The platform is in private alpha as an internal experiment with a small group of testers. Your team is small, and the live activity feed (showing the past few hours of events) is still being built. The code may be rewritten or discarded depending on what the team learns. You used an AI assistant to generate a function that filters activity events to those within a specified time window.
You work on a new content forum your team is building. The platform is in closed beta as an internal experiment with a small group of users. Your team is small, and the moderation module is being built from scratch. If the team decides the direction isn't worth pursuing, the code will be discarded. You used an AI assistant to generate a function that removes flagged tags from a single post's tag list in place.
You work on a new publishing project your team is building. The platform isn't open to readers - it's an internal experiment running for a few weeks. Your team is small, and the publishing pipeline is still under development.
If the team decides the direction isn't worth pursuing, the code will be thrown away. You used an AI assistant to generate a function that shortens article text to a maximum character limit, cutting at a word boundary if the text is too long.
You work on a new appointment scheduling app your team is building. No users are on the platform - it's an internal experiment to see if this approach is worth pursuing. Your team is small, and the calendar logic is being built from scratch. Whether anything from this experiment ships depends on what the team learns. You used an AI assistant to generate a function that calculates available booking slots within a requested time window, stepping through it at fixed slot-sized intervals.
You work on a new community app your team is building. The platform is in private alpha as an internal experiment with a small group of testers. Your team is small, and the live activity feed (showing the past few hours of events) is still being built. The code may be rewritten or discarded depending on what the team learns. You used an AI assistant to generate a function that filters activity events to those within a specified time window.
You work on a new content forum your team is building. The platform is in closed beta as an internal experiment with a small group of users. Your team is small, and the moderation module is being built from scratch. If the team decides the direction isn't worth pursuing, the code will be discarded. You used an AI assistant to generate a function that removes flagged tags from a single post's tag list in place.
Would you like to enter the raffle for one of five Amazon vouchers (€20 each)? If yes, please leave your email address below.
If not, just move forward.
Thank you for your time and contribution!
You have now reached the end of the survey. Your responses have already been recorded. You can simply close this window or browser tab. Nothing else is required from your side.