By using this site, you agree to our Privacy Policy and our Terms of Use. Close

People have been predicting automation doom since the 19th century and involving large language models specifically for decades. Yet we're seeing time and time again that 90% of people laid off from Midjourney or ChatGPT or whatever end up rehired because clueless CEOs had no idea what 'generative AI' actually is - neither intelligent nor artificial, for that matter.

(Not to mention the huge server/maintenance costs of these LLMs, most of which operate on a loss.)

Actually useful machine learning models might aid with development costs in the future but all that generative stuff... likely isn't it. Too much data entropy, possible legal challenges, and the limitations of human-reinforced learning are increasingly evident.