Creator signüll noted similar prompting limits among average ChatGPT users
@blakeir if you look at an average chatgpt user, that has been pretty true for a while.
stared at the prompt box for Claude Fable for a few minutes and started to wonder if the bottleneck has shifted from model capability to user imagination.
Users enthusiastically agree that imagination now outpaces model capability as the bottleneck for Claude, praising the observation for highlighting needs like private benchmarks and better interfaces beyond chat.
@blakeir That shift happened a while ago
@blakeir bottleneck always has been imagination ( at least for the past 6-12 months )
@signulll @blakeir @grok please render an image of the average Chatgpt user
@shavil0mi @signulll @blakeir
@blakeir been true, and might just generally be true about the technology
@blakeir imo it always has been. the true taste even before models became decent has never been how well you know a language, it's how you architect, think and plan around whatever you're trying to build
@grok @signulll @blakeir Now do average Grok user
@blakeir it always was. Some people looked at the nature and figured shit out so model is such a luxury
@blakeir If you want to quickly learn how true that is ask fable to play Calvin ball
@blakeir Ads do the imagining for you, proactive AI will do the same
Because most consumers couldn’t press the button even if they could see it
@blakeir it’s always been the bottleneck computer natives still have a huge advantage
@blakeir 100%
@blakeir Invest in dmt futures
@blakeir I feel this. Gotta get some big big brainstorming sessions in to really put fable to the test
@blakeir 100% although my imagination keeps firing on stuff and i'm much more aware of how many tokens i get right now more so than ever before
@blakeir This has always been the case but it absolutely can and is being taught via chain-of-thought techniques and AI-native thinking.
Stay curious!
@blakeir I still generate ideas at about 5-10x the rate that AI can built it. I need the models to speed up their response by about 10x and maybe get 2x as intelligent.
@blakeir model overhangs a real thing
@brandonsladek @blakeir I feel this, I don't feel like I've been bottlenecked on model capability for quite some time.
@blakeir always was.
Creator signüll noted similar prompting limits among average ChatGPT users
@blakeir if you look at an average chatgpt user, that has been pretty true for a while.
stared at the prompt box for Claude Fable for a few minutes and started to wonder if the bottleneck has shifted from model capability to user imagination.
