Yeah, and besides, beyond the fact that the claim of "no groundbreaking firsts" was easily falsifiable, I found this one funny:
> took something we've done for 50 years and do it more efficiently.
One can argue that this is "all" that happened in AI in recent years. We took something that was arguably done before, in some similar way or shape, and made it practical. But this is HUGE. It's not a minor accomplishment! It may very well be harder than doing some prototype for the first time.
Indeed, and I'll observe that the significance of this one is often lost on more casual observers.