Member-only story
Technology Sprints, Society Crawls
Lessons from the Industrial Revolution
I don’t know if I’ve ever gotten so much mental whiplash thinking about technology as I get from thinking about generative artificial intelligence.
When I read somebody like Ed Zitron, I come away convinced that it could very well be a “big, stupid magic trick.” Zitron makes a few pretty persuasive points:
- The servers that power AI are really expensive (not to mention energy-intensive), and right now AI companies are not making nearly enough money to cover those costs. When the hype and the venture capital run out, the “AI bubble” could collapse.
- AI can do cool things sometimes, but it’s still far too unreliable to be useful — OpenAI’s state-of-the-art model still has trouble answering questions like “How many states have the letter A in them?”
- The AI companies have trained their models on a bunch of copyrighted material, and they will soon find themselves under the weight of a bunch of lawsuits.
He speculates about an industry-wide implosion:
I am deeply concerned that this entire industry is built on sand. Large Language Models at the scale of ChatGPT, Claude, Gemini and Llama are unsustainable, and do not appear to have a path to profitability due to the…