Seven years ago a representative from Google reached out to me in an official capacity to discuss the company's energy use. I told them it would become necessary for major firms in core economies to reduce energy use in order to enable Paris-compliant decarbonisation. They assured me that their new AI tools were already helping them identify ways to improve the company's energy efficiency, so they weren't worried. They insisted this would soon yield radical efficiency improvements, and lead to energy use reduction without any change to their production activities. "Just wait, you'll see!" they told me, full of confidence. I told them that's not how it works. In growth-oriented capitalist firms, savings from efficiency improvements are generally leveraged to expand production. I predicted that even if their AI found ways to improve efficiency, Google's energy use would go up, not down. Seven years later and this is where we're at: