3 ms·

So, the point here I think is: - Flow-based programming enables a more composable way of developing software which avoids a lot of errors coming from incorrect

So, the point here I think is: - Flow-based programming enables a more composable way of developing software which avoids a lot of errors coming from incorrectly connecting parts of code together, making it more safe for allowing AI to the loop, and designing the graph of connected components. - The graph is also more easily understood by humans, making the verification of the produced program easier to do for a human. So in summary: Having an easier to understand and more composable way of developing software helps both for collaboration between humans, but perhaps extra much so when letting "still kind of unproven" AI take part.

bergie 4y ago
Yup! Similar approach would certainly work with more traditional programming methods as well. But FBP enforcing that separation of concerns makes it easier to split the workload between GPT and human.
jacobr1 4y ago
In some ways, this has always been good programming design. Have clean interfaces. Encapsulate state, and have clear propagation of state between components. Identify appropriate abstractions that represent the components of your system such that they can be composed, potentially reused, and can be clearly reasoned about and tested. Pulling this into the explicit programming framework rather than just being a set of good design patterns does seem like it help when collaborating with semi-trusted participants. This is how most "low-code" systems work.