A viral trend has emerged where developers are instructing AI models like Claude to communicate in simplified caveman-speak to dramatically reduce token usage and API costs. The technique works by constraining AI outputs to minimal vocabulary and short sentences, though users report it can compromise usability and response quality.
·Developers discovered that caveman-style prompts reduce AI token consumption by limiting vocabulary and sentence complexity
·The cost-cutting method has sparked debate about the tradeoff between efficiency and AI capability
·Some users report the caveman constraint makes AI responses unusable despite lower costs
·The trend reflects broader concerns about expensive API pricing and AI operational budgets
·The viral nature of the hack raises questions about prompt engineering's role in managing AI infrastructure costs
drawn from Country Living Magazine, 404 Media, Business Insider, IMDb · updated 6h ago