Just because something doesn’t display understanding by responding to your expectation doesn’t mean it doesn’t possess understanding . If you ran into my office with these prompts, the response to each would be “what the hell are you doing in my office?” All behavior is contextualized, and GPT-3’s native context is predicting continuous text, not answering questions.

It’s a distracting anthropomorphism to even attempt ascribing “understanding” to a model like GPT-3. An assessment of its useful capabilities should be through an honest effort to get it to do something - and should of course include consideration of the effort/intelligence required to do so. Marcus knows enough to know this set up is inappropriate, so the article reads as disingenuous.