Post #1540641

@lritter I won't make predictions "AI will never... " anymore. The last times I though the rather simple architectures won't be able to do certain things I was wrong. Currently I more expect coding to be solved "good enough" for 90% of the use cases, possibly together with a babysitter that tells the AI what choices were not as expected when prompting for the feature. And much progress doesn't even require better models, but will come from better tooling. Even though more context would be nice.