Did you read the second half of the article? It is all about all of the serious drawbacks to this approach. I don't think I have any "misplaced optimism."

I would also disagree with the idea that current AI systems are not different than what was available 30 years ago. Eliza is fundamentally different from GPT-4.