iiinova — where your digital self keeps evolving

Replies

dan @danabra.mov · 22/09/2025

the thing i'm worried about is whether AI will decimate the "flow states" of knowledge work. i realize this is very selfish. i enjoy being in a flow state. i don't want to constantly be babysitting an unlearning intern who repeatedly violates instructions. even if they eventually produce good work

1019510

Phillip Broth @thebrochures.bsky.social · 22/09/2025

This has been my experience. I’m very powerful now but enjoy it less. Flow is the reason I’ve lasted in this career.

020

Josef Grahn @jgrahn.bsky.social · 22/09/2025

I think it's a matter of how the tools are designed. I could see AI facilitate flow by tackling obstacles that would break it, large or small, and that might turn out to be the most effective way to utilise the human part of a symbiotic relationship.

120

a14y @a14y.eu · 22/09/2025

it’s the same for me. and i even enjoy writing specs. but that’s because it forces me to think about the problem. i even did this for me before the agentic stuff. currently for me it seems i use agentic stuff for things i know how to do but are time consuming and not interesting.

000

Andreas Møller @andreasmoller.dk · 22/09/2025

I have the same thought. It feels so disruptive to my productivity when I have to wait for an AI all the time. Event if it did generate good code I don’t think I would like it

130

Tom Hicks @hicksyfern.bsky.social · 22/09/2025

I feel this with the light of 1000 suns. As useful as they are sometimes, they definitely prevent flow. Usually my flow sessions follow a load of failing LLMs then fuck it I’ll do it myself which tbh I should probably just do as I’m an experienced adult

010

dan @danabra.mov · 22/09/2025

even if the end result arrives faster, *this type of work* is not enjoyable to me. it tires me. it's like i used to build an invisible canvas of concepts and now i'm begging a slot machine. i'm worried that the only enjoyable version of that is the one that's also smart enough to not need me at all

6921

James Snell @jasnell.me · 22/09/2025

I've been experimenting with a few different ways of using agents like Claude in my workflow. Rather than relying on it to do the initial work, I've been having it instead serve as a review step and a task organizer, both of which it appears to excel at. No interruption to flow.

030

sawyer @sawyer-p.me · 22/09/2025

After having used my corporate Cursor account for a while, I think you just get used to knowing precisely when to invoke the model and when not to. a) Text manipulation more complex than a vim macro can handle b) Related to configuring/setting up libraries c) Dead simple util functions

100

AntmanCodes @antman.bsky.social · 22/09/2025

This is why I don't use it — I enjoy writing code, and unless you process information, you are not actually learning. Recognition provides the illusion of learning while recalling and processing provides deep learning. You've got to write code, understand, build pattern muscle memory, then level up

040

TQ White II (he/him) @tq.tqwhite.com · 22/09/2025

I wired Claude to a graph database and it’s learning. It’s the strangest thing. I still share your concern about not getting flow but Claude having persistence is very cool.

000