Fair point. Still a very quick turnaround considering the other labs would have to figure out both HOW to train a Mythos-level model and then do the work (and Grok is the last to catch up), but certainly more plausible than a 2 month window.