While I don't disagree that there's no craft most of the time, it's interesting to note that the anti-analogy with art and artists doesn't really work (or works in an unexpected way).
I'm visiting a lot of museums and exhibitions for a number of years now, without having any kind of art education, etc. After some time you start seeing patterns and understand more even if you still haven't read your art history. What strikes me time and time again, is that there's a huge amount of repetition even in great artists' work. It's definitely not an assembly line, but they do the same thing over and over again for years and years. That was true centuries ago and that's true today.
That is a very thoughtful piece. Thank you for posting it. I especially like the idea of that new mode of programming (or problem solving, whatever you call it) when people very much enjoy getting an LLM to do what they want.
This is both new and old, because it's the same joy (or dopamine hit) of making a machine do your bidding. Honing your prompts is not that different to honing your shell scripts. I think many people overlook this aspect.
I do a fair bit of "prompt honing" myself, but there remains a fundamental difference between that and traditional hacking of yore: the latter is predictable. The system may be very complex, but you can map it out step by step, and the more you do so the more advanced things you can cajole it into doing - and, importantly, those things don't suddenly stop working "just because".
But with LLMs, your well-honed prompt might be an abject failure on a very similar task that just happens to have enough differences in context that the "magic" in the prompt no longer works. And debugging those kinds of issues basically amounts to throwing things at the wall and seeing what sticks. And then of course the providers switch models from under you and retire the old ones. So it's not quite the same joy of making a machine do your bidding. In the former case, the joy was from knowing that you can do this now and in the future - that you grok the thing enough to make it do this bidding at any point. With LLMs, it's more like the joy of successfully casting a spell, but you still have no idea why that particular spell worked better than a dozen others, nor can you rely on it working in the future.
Many people mentioned the Farnam Street newsletter, but Shane Parrish has also published a book in 3 volumes: The Great Mental Models, see https://fs.blog/books/
There is a way around that. First of all, if you and your wife have two free accounts and share workspace subtrees with each other, you can double the free content limit.
And if you pay for just one personal account, you can still share content with a free account. This is very close to having two people sharing one paid account, feature-wise.
Main difference to Notes is that Bear uses Markdown as the data format. It has a rare capability of converting webpages to Markdown when you save them as a note using a sharing extension on iOS. The conversion is decent.
But as mentioned above it also lacks important organisation features: no folders, no dated notes, etc.
There’s also another competitor to those mentioned, called Coda (https://coda.io) The biggest downside they have compared to Notion is that docs are separate from each other. There’s no structured workspace. Dropbox Paper also has this problem.
From my recent experience on a real project DDD inspires very rigid thinking sometimes on the boundary of being religious (e.g. the quest for pure Values and Aggregates eclipsing the problem domain understanding).
And for a regular "aggregate" entity I would never prefer a JSON storage to a data mapper with SQL (as long as you have an option to use SQL). You should have a reason for JSON and such, otherwise it's just NoSQL hype.
I'm visiting a lot of museums and exhibitions for a number of years now, without having any kind of art education, etc. After some time you start seeing patterns and understand more even if you still haven't read your art history. What strikes me time and time again, is that there's a huge amount of repetition even in great artists' work. It's definitely not an assembly line, but they do the same thing over and over again for years and years. That was true centuries ago and that's true today.