GPT-3.5

Large language models increasingly reply to prompts with a believably human response. Can they also mimic human behavior?

Amid the hype that surrounds large language models, a crucial caveat has receded into the background: The current cost of serving them at scale. As chatbots go mainstream, providers must contend with the expense of serving sharply rising numbers of users.

It wasn’t your imagination: OpenAI’s large language models have changed. Researchers at Stanford and UC Berkeley found that the performance of GPT-4 and GPT-3.5 has drifted in recent months. In a limited selection of tasks, some prompts yielded better results than before, some worse.

Hate talking to customer service? An AI-powered tool may soon do it for you. Joshua Browder, chief executive of the consumer advocacy organization DoNotPay, demonstrated a system that autonomously navigates phone menus and converses...