The Great AI Downsizing: Why Cheaper Models Are Suddenly The Smartest Bet
Rising costs are forcing enterprises to shift from expensive frontier models to cheaper alternatives, potentially reshaping the AI economy.Summary
The artificial intelligence industry has long operated on the premise that bigger models are better, but mounting costs are forcing a shift toward efficiency. Coinbase co-founder Brian Armstrong predicts that within 12 to 18 months, 80% of AI workloads will run on models that are 99% cheaper than today's frontier systems. This prediction is supported by real-world tests, such as those conducted by the legal AI tool Harvey, which demonstrated that costs could be reduced by three times without any loss in quality by intelligently routing simpler tasks to smaller models. The emerging cost war is not just between proprietary and open-weight models, but between large-scale and small-scale inference. If most enterprise deployments can be run effectively on smaller, cheaper models, it would dampen demand for inference and raise questions about the justification for the enormous cost of training frontier models. The industry is at a crossroads, facing the choice of embracing efficiency or finding new ways to justify the premium cost of frontier intelligence.
(Source：Home - Bitcoinworld.co.in)