Blink:

This will be a brief post. I just want to address the relevancy of my September post, when I proclaimed generative AI’s potential as the elephant in the room in regards to planet earth’s environment based on computing centers’ reliance on energy and chilled water for cooling.                   

Read On:

Recently numerous articles about a looming AI energy burden are surfacing. The specialized computer chips utilized in the current AI boom consume immense amounts of electricity which would also impact the world’s carbon emissions if data centers get their power from fossil fuels. From what I have read, projections are all over the map:

• One conservative estimate is by 2027, 0.5% of the world’s current electricity (equivalent to Sweden, the Netherlands, and Argentina combined).
• Most companies have been guarded about their energy and water usage. Another projection is 1 to 1.3 of the world’s electricity based on the sales of the technology (hardware and software) sold by the dominant market leader Nvidia (an estimated 95% market share).

Potential Solutions? A.) More efficient utilization (accelerated computing) of the current AI hardware servers; B.) Increased supply of specialized chips beyond Nvidia’s which has become an AI capacity bottleneck; and C.) Advocate/support the moratorium on the accelerated development of AI to give regulators and companies the time needed to formulate the future safeguards needed to protect us from the risks associated with AI, especially as it relates to planet earth’s environmental resources, and health.

Opinions Welcomed!