I think @ibdknox made that version

Pygmalion is fascinating, and very similar to what I was thinking. I realize having seen the video, @Mariano Guerra, that the nature of the conversation between the user and the environment is significantly different depending on the paradigm of the language. Imperative languages need to repeatedly ask "what next?" But declarative languages would need to ask something more like "defined how?"

Definitely helps me understand what I'm aimed at, better, thanks.

@Corey, I wasn't familiar with Eve, and my mind is blown. That it could not find a place to live is so disappointing. The use of colours instead of variable names, the use of containers to represent conjunctions... All very cool.

I notice that in none of these cases, though, does the interface explicitly request information. That, to me, is the crux of the idea... are there interfaced where the environment actively solicits information from the user?

or proof assistants

Not necessarily in a textual format, like that, but that sort of conversation. Where the environment actively prompts the user based on an understanding of the syntax and semantics involved in the underlying language...

Good point on theorem provers. If you have premises and a conclusion, they give you a set of intermediate steps to aim toward, and ask for a step in that direction. But the "asking" is basically by showing the new targets, I think? The actual communication of "what strategy should we explore for deriving X from Y?" is sort of implicit?

It is more iterative than conversational, maybe? Is that fair?

Also, @Corey the eve example has given me some interface ideas for a different problem I have been thinking about, which is greatly appreciated. Specifically, how to bridge the gap between wanting to use typing for speed of entry and visual representations for clarity and beginner ease of use. Very interesting approach.

Is there a postmortem on Eve somewhere? Something that talks about what it has to teach about building businesses around innovative programming interfaces? Last night I heard about a friend's similar dead-end in a very innovative approach to something, and I wonder whether there is some lesson I can glean... Is it a cautionary tale about solving problems people don't yet know they have? Or what? Because I see my own work headed in a frustratingly similar direction.

That was a pretty good post mortem

anyone have a link to that?

oh wait here it is

For reference on what Eve did

Thanks for the links! 👀

Inspiring as shit. Thanks, guys.

I still don't get it, which makes me nervous. I don't know what went wrong, there, so I have no way to avoid it. 😬

True progress is extremely time intensive because it requires a crazy amount of experimentation. This means you basically have two options: find a way to give yourself a long time to find the kernel of something that is geniunely better by a large degree (and then raise to try and productize it) or you do something incremental.

There were a number of incremental off ramps for us (we basically built notion, roam, airtable, bubble, and a number of others along the way...), but they weren't going to move the bar much for what we cared about; making programming more accessible for new comers and more productive for those of us who already did it.

You could argue that it was a bad choice to give up on businesses that could move the needle 10%, but we had the opportunity to try to swing for the fences and so that's what we did.

It's a low probability bet though.

retool too