Put it in plan mode.

Then, repeatedly ask Claude to criticize the plan and use the "AskUserQuestion" tool to ask for your input.

Keep criticizing and updating the plan until your gut says Claude is just trying to come up with things that aren't actually issues anymore.

Then unleash it (allow edits) and see where you get. From there you may ask for one off small edits. Or go back into plan mode again

Haven't resized a window with a mouse since using aerospace

It's not as one-dimensional as good vs bad. Transformers generally are extremely useful. Do I want to read your transformer generated writing? Fuck no. Is code generation/understanding/natural language interfaces to a computer good? I'd have to argue yes, certainly.

I cry every time somebody tries to frame it one dimensionally.

I cry every time someone uses the f-word in their writing for no apparent reason.

I'd rather read "fuck" than "f-word". The latter is like eating dumplings for lent because surely $deity won't see the meat in there.

Cultured people have no need for such words in public discourse so I'd rather not see either of them or the need. People are judged by the words they use.

Cultured people? I'd certainly argue that words which people feel are "uncultured" strictly DO have a use case.

It's as if you're saying "smart people only color inside the lines." Take a step back.

The usage of those words is to accurately convey the speaker's emotion. If you don't see any use in that then that's your problem.

There are far better words availble than to use those often referred to as "gutter talk".

Also, unfortunately I have in my global instructions to never use em dashes...

Maybe I'll get over it eventually.

nice

Yeah, I can feel it too, it _mostly_ works but.. feels like it needs a rewrite.

for example Amp "feels" much better. Also like in Amp how I can just send the message whenever and it doesn't get queued

* I know, lots of "feels" in there..

Mostly speculative

I'm a fan human pruning (which you refer to as "write gates").

At this point, I'd argue any "automated memory" is a failure over time. I don't know if I'll be convinced otherwise.

If one were to go full automation, then I'd want generations of what to prune ("I haven't used this memory in a while... "). Actually, that might be useful in your tool as well. "Remove-gates"?

Agree on human pruning piece for the most part. The system is built on the assumption that the model shouldn't necessarily be trusted to decide what's permanently important.

Remove-gates are a good extension of the same philosophy. Right now /recall-maintain lets you manually review and clean up, but it doesn't proactively surface "hey, you haven't referenced this in two weeks, still relevant?" That would be the natural counterpart to the write gate: system surfaces candidates, human decides.

Am going to think on this, but will likely add. Appreciate you!

Opened a design issue for this if anyone wants to follow along or weigh in: https://github.com/davegoldblatt/total-recall/issues/2

Nice work!

Please do

Wait what..? please elaborate or provide any references for further reading!

Sure!

The first is an attempt to provide a semantics for activity diagrams as constraints on a state machine and thereby allow folks to specify correctness properties for the state machine using a visual language. Existing work on semantics for activity diagrams already exists but doesn’t come with tooling in the way that temporal logic does (https://arxiv.org/pdf/1409.2366)

The second is an attempt to fix a long standing problem with state machine specification languages. While many support composition operators (parallel and/or nesting) none of them come with strong theorems about when temporal properties proven about constituent elements will remain valid in the composite.