The Efficiency Trap

#1

Post by djabatt » Thu, Jul 31, 2025, 12:31 AM UTC

The Efficiency Trap
anthonybatt.com

Re: The Efficiency Trap

#2

Post by djabatt » Thu, Jul 31, 2025, 12:31 AM UTC

You have reached your specified API usage limits...

Re: The Efficiency Trap

#3

Post by bigyabai » Thu, Jul 31, 2025, 12:32 AM UTC

djabatt wrote: post #2

You have reached your specified API usage limits...

I quit Claude 2 days ago after trying the new Qwen3 30B MoE model. We're reaching the point where offline models are nearly replacements, I recommend it if you've got ~8-10 gigs of memory to play with.

Re: The Efficiency Trap

#4

Post by djabatt » Thu, Jul 31, 2025, 12:46 AM UTC

bigyabai wrote: post #3

djabatt wrote: post #2

You have reached your specified API usage limits...

I quit Claude 2 days ago after trying the new Qwen3 30B MoE model. We're reaching the point where offline models are nearly replacements, I recommend it if you've got ~8-10 gigs of memory to play with.

I have just started to enjoy Claude Code CLI. I thinking about my next steps.

Re: The Efficiency Trap

#5

Post by willprice89 » Thu, Jul 31, 2025, 12:48 AM UTC

While my company relies heavily on cloud infrastructure (we offer a VPN service) I made it my goal from the outset not to become dependent on a single provider for any one piece of that infrastructure. I think it helps strike a balance between efficiency and resilience.