I taught Claude to talk like a caveman to use 75% less tokens.
normal claude: ~180 tokens for a web search task
caveman claude: ~45 tokens for the same task
"I executed the web search tool" = 8 tokens
caveman version: "Tool work" = 2 tokens
every single grunt swap saves 6-10 tokens. across a FULL task that's 50-100 tokens saved
why does it work? caveman claude doesn't explain itself. it does its task first. gives the result. then stops.
no "I'd be happy to help you with that." no "Let me search the web for you" no more unnecessary filler words
"result. done. me stop."
50-75% burn reduction
with usage limits getting tighter every week this might be the most practical hack out there right now
@AnthropicAI@claudeai hopefully minor ask but driving me nuts: can you guys make it so the app on both Mac and iOS keeps the screen active when Claude code is off doing something? Some of us like to watch the system cook. Can be a setting on backend.
Anthropic just took a big swipe at OpenAI's decision to put ads in ChatGPT. Anthropic is airing ads mocking ChatGPT ads during the Super Bowl, and they're hilarious 😅 Anthropic is also committing to no ads in Claude https://t.co/LR1v4xz9ds
PSA for anyone building a billing portal:
1. Keep receipts accessible for 16+ months. Your customers do taxes in Q1 and will discover your 12 month default right when they need month 13. Ask me how I know.
2. Add a bulk export button. Making someone click “download” 47 times gives them ample opportunity to reconsider whether your product is worth $12,000/year.
Cancelled my Pro subscription. Beginning the influence of advertisers is the beginning of the end of a truly unbiased model. Disappointed, but happy I’m given the out to go to @claudeai instead ahead of it being deeply integrated into my life. Wished for better from you as a leader.