Does Vibe Coding Still Feel Like Software Engineering?

Key Concepts:

• AI coding
• Software engineering
• Productivity
• Skepticism
• Iteration
• Testing
• Technical understanding
• Engineering management

Main Topics and Key Points:

The speaker discusses an observation about people's reactions to AI coding, particularly those who are adjacent to software engineering roles.

• Engineer's Shift in Perspective: An engineer initially skeptical about AI coding became more positive after experiencing increased productivity. However, skepticism remains regarding security concerns.
• Experience with Non-Engineers: The speaker has been experimenting with and building tools around AI coding since 2023, sharing them with non-software engineers like teachers, product managers, and UX/UI designers.
• Excitement from Adjacent Roles: People in roles adjacent to software engineering (e.g., those who tried it but didn't pursue it) are most excited about AI coding, hoping it will make software development accessible.
• The Core Experience Remains: Even with AI handling the coding, the experience still feels like software engineering, involving iteration, testing, and problem-solving.
• Analogy to Engineering Management: Using AI coding feels like being an engineering manager who still needs to understand the technical details.
• Reversion to Outsourcing: People who initially got excited about AI coding often realize after a few hours that they would still prefer to pay someone else to do it because it still feels like software engineering.

Important Examples, Case Studies, or Real-World Applications Discussed:

• The speaker mentions sharing AI coding tools with teachers, product managers, and UX/UI designers.

Key Arguments or Perspectives Presented, with Their Supporting Evidence:

• Argument: AI coding doesn't fundamentally change the experience of software engineering.
  • Evidence: Even with AI doing the coding, users still need to iterate, test, and understand technical details.
• Argument: People who gave up on software engineering may not find AI coding as appealing as they initially thought.
  • Evidence: After trying AI coding, they often revert to outsourcing the work.

Notable Quotes or Significant Statements with Proper Attribution:

• "It still feels the same even if AI does all the coding for you it still feels like software engineering."
• "It kind of feels like you're at best your engineering manager who still needs to go into the deep technical reasons things happen."
• "...after a couple of hours they find out they would still rather pay someone else to do it because it still feels like software engineering even if AI is doing most of the coding for you."

Technical Terms, Concepts, or Specialized Vocabulary with Brief Explanations:

• AI coding: Using artificial intelligence to generate code.
• Iteration: Repeating a process to refine and improve a product or solution.
• Testing: Evaluating software to identify defects and ensure it meets requirements.

Logical Connections Between Different Sections and Ideas:

The speaker starts with an observation about an engineer's changing views on AI coding, then connects it to their experience sharing AI coding tools with non-engineers. This leads to the central argument that AI coding doesn't fundamentally change the software engineering experience, which is supported by the observation that people often revert to outsourcing.

Brief Synthesis/Conclusion of the Main Takeaways:

AI coding may increase productivity and make software development more accessible, but it doesn't eliminate the core challenges and complexities of software engineering. People who are adjacent to software engineering roles may initially be excited about AI coding, but they may ultimately find that it still requires a level of technical understanding and effort that they are not willing to invest.