im using fable almost exclusively, i just buy a new 200$ license if i run out of capacity. its so much better, easily worth it given how much work I get done. I built a code, deploy, e2e test loop with my full aws infra, fable just implements linear tasks constantly, 5 at a time, tests the whole thing end to end

if people dont see why they need a model this smart, they probably arent using ai enough

▲

HDBaseT 3 hours ago | parent | next [-]

People have realized you don't always need a Fable level model. Majority of my work is sufficient with ChatGPT Luna which has effectively unlimited usage on the $100 ChatGPT plan.

I am not doing awfully complex tasks though. I imagine a lot of other people are in a similar boat, either switching from Claude to ChatGPT or even just min-maxing DeepSeek V4 Flash 0731 or similar.

	▲	mikert89 3 hours ago \| parent [-]
		the code quality is higher though, so the value compounds. i dont need to watch as closely to what its doing, so its more autonomous

▲

apt-apt-apt-apt an hour ago | parent | prev [-]

Isn't there a risk of getting banned if you do this (multiple accounts to bypass usage limits)?

	▲	mikert89 32 minutes ago \| parent [-]
		Idk I never worry about it