Профиль

I feel like I get "actualization burnout" Used to be I had so much time to mull over a new inspiration while completing the current thing The other thing is, it used to be easy daydreaming about doing something I knew I'd never get done. Now I have so many ideas I *know* I can do

the agents iterate so fast. Switching to not-doing-the-dumb thing has a really low activation cost

I remember my manager coming to me and saying "I'm using this new browser, you should try it."

theory of mind is hard and annoying

Lawyers not making something complicated challenge: failed

One time I chatted with Opus to spin a series of my bad short stories into the plot for a novel, and that hour was glorious

Also have Fable make some new factions

er 11 hours

the rest I understand?

Taking at it as a Left view of, well, anything, her critique kind of lines up. But way more dismissive than the movie deserves--it was oblique to what she wants from the story Nolan himself kind of did a structure-preserving collapse of the story along the lines of Willpower and Trauma 🤷‍♂️

One time I pulled a bunch of people into a meeting and my first question was, "Why is this a crisis?"

Her criticisms were along an entirely different dimension, and, along that dimension, actually correct and insightful

Chart Nazi: NO CHART FOR YOU!

crank stuff, 5.6 humoring me 😂 No idea what this means

You can't imagine how much this dynamic means to me 🤣 I love to have the models talk about my bad short stories, same reason

maybe its time🤷‍♂️

I like your personhood discussions Because getting to a checkmark in that box is a long journey, and after we check it there's a huge question of what's next

...which, to your point, is the opposite of delving into machine instructions

So my other ihmo is that we're going to move away from frameworks per se and move toward gigantic toolsets, where the models can piece together functionality from a giant span of tooling We may end up with an "Everything Library" which bundles functionality cognitively, just extremely broadly +

My imho is that the beasties have a much, much higher limit of cognitive capacity than humans do (eg I can't fit all of Moby Dick in my working memory). But there's still a limit So the measures we take to reduce cognitive load in humans *will* also benefit the models, just not as obviously +