> Open models... are still behind GPT-4 by some distance but they are closing in at a rate that leads me to believe there’s not much of a moat there This assum

> Open models... are still behind GPT-4 by some distance but they are closing in at a rate that leads me to believe there’s not much of a moat there This assumes that openai internally doesn't also "closing-in" towards something even more impressive to be released next year or whatnot.