Didn't you write a piece arguing that OpenAI was history less than a year ago? I didn't buy that argument; but neither do I buy this comparative puff piece which recapitulates the company's own marketing.

What's best in your analyses is that you generally get around to debunking the hype. What's weakest (especially here) is that you spend a lot of time building up the hype before deflating it.

For example, "...What I’ve read these days is significantly more exciting and, for some, frightening."

Why exactly should any knowledgeable person be "frightened" about GPT-4? We should be concerned about underregulated technology companies that market tools without any due diligence regarding social impact or potential harms to individuals, groups, or society at large. This is a question of governance.

But the technology itself? Anyone who is frightened by it is likely to be clueless about how data-driven machine learning works. At bottom, GPT-3 is a statistical model trained on scraped data. GPT-4 will also be a statistical model trained on scraped data.

If there are significant improvements in performance capacity for certain writing (or other generative) tasks, let's find out about them. If these portend significant harms to individuals, groups, or society at large, let's regulate the technology.

In sum: It's nice that you keep your ear to the ground but can you spare us the hype?