@Jesse_Livermore@ballsandgutters BS jobs may provide value but implicit or ironic rather than the literal function
They might offload social dynamics, provide supporting evidence for a narrative of effort, add inertia for status quo beneficiaries, etc.
Value, but externally illegible
@antirez Automated tests, design documents, declarative vs procedural code, high and low level code, static vs dynamic, LLM prompts; all these seem to be complementary statements of intent, with different specificity and rigor, and strongest when combined well
@Grady_Booch Code is both asset and liability
An asset we value for what it can do; for crystallizing hard-won procedural insight
A liability requiring upkeep, with inertia owing to its commitment to a point of view and endless accidental detail
@benlandautaylor For experienced teams, in many cases the goal is to net-reduce the amount of software, and improve quality rather than quantity. Lines of code is a liability as much as an asset
@nicbarkeragain One of the ways I’ve helped kids learning algebra is pointing out these notational pitfalls in a way they feel motivated to demonstrate being clever by dodging them. t vs + is another
@AgileJebrim@nicbarkeragain Also interesting the role non-semantic formatting and layout plays in both text and graph representations
You can use a fmt tool for code and auto-layout graphs but in both cases it may project away a dimension of guidance
@AgileJebrim@nicbarkeragain My experience has been that big enough systems inevitably incorporate both text and visual flow representations. If they favor one as the primary human interface the other shows up as a diagnostic alternative.
@twiecki But yeah the way we yearn to humanize our interactions with AI is fascinating. To make it a dialogue; to imagine it is “thinking” and “smart” rather than just a trained numeric algorithm; to give it a voice and speak to it. Even since the days of ELIZA and HAL9000
@twiecki Intuit has this for TurboTax, where they make the progress bar move slowly to suggest it is working hard on behalf of the user, and to avoid the reality that the computation requires << 1 millisecond
@justinskycak There’s a lot of implicit conditionality on the right side, where “it depends” and it’s rarely worthwhile to enumerate cases in throwaway discourse. Many of the best discussions are predicated on shared context or “skin in the game” for longer term iterated games
@justinskycak@zdrks The smartest people I know seem to carry a larger and more organized toolkit of mental models
Right tool for the job makes all the difference, both when working on physical things or mental things
They weren’t born with a toolkit — they assembled theirs, day by day
@justinskycak@zdrks This is similar to how people may disagree over what is intuitive
Intuition is a product of our unique past experience, so we share intuition to the degree that we have shared experiences
This is also why it feels great to share the paths we have taken in life, to tell stories