Automatic Programming - How Far Can Machines Go

Explore the potential and limitations of automatic programming in this 52-minute conference talk from YOW! 2018. Delve into program synthesis, its current capabilities, and future possibilities as a tool for developers. Learn about self-programming systems, human-computer collaboration in coding, and realistic expectations for synthesis engines. Examine different strategies, interaction models, and programming by example techniques. Discover how to navigate challenges when AI models outthink programmers and the importance of maintaining understanding throughout the process. Gain insights into read-eval-synth loops and their role in advancing automatic programming technologies.

Intro
Program synthesis to the rescue
What if we have automatic programming?
Query knowledge about some kind of code
Commit Strip said it best...
Adjusting our expectations
What does this all mean?
Realistic expectations for realistic synthesis
The synthesis engine
Trying a different strategy altogether
The Interaction Model
Specifying and re-specifying intent Task: find the median of a list
Programming Not only by Example
When models "out-think" the programmer
Keep understanding what's going on
Read-Eval-Synth Loops