3 ms·
Let’s get a reality check. Current agent performances are no where near that kind of threat, and it’s uncertain when we get there. I think people continue bull
by nis0s 9mo ago
Let’s get a reality check. Current agent performances are no where near that kind of threat, and it’s uncertain when we get there.
I think people continue bullshitting in this domain because they’re worried they have no moat, so they have to discourage via sophistry or bold claims.
Don’t take my word for it, here are the stats https://scale.com/leaderboard https://scale.com/leaderboard
For reference, 25% means getting 1 in every 4 question right.
Secondly, if we ever do get to that point, people have lots of social mechanisms they’ve developed over millions of years of evolution which will kick into place. I am not sure anyone dreaming of this “utopia” is then going to be around to see it.
- disgruntledphd2 9mo agoYeah, it's strange how Dario says this every time he's raising a new funding round. Odd.
- nis0s 9mo agoRealistically, the rate at which both MCP, and Claude products have developed competitors should be concerning for anyone at the helm at Anthropic. MCP was introduced in Nov 2024, and there are already at least 3 competing standards. The same could not be said for something like Tensorflow when it came out, for example, because it took other frameworks/tools a while to “get it” and become something competitive. Yes, I know these are all open source frameworks and not a “real” example of a product Anthropic should be concerned about, but I am making a point about the rate of competition developing. There is real danger here of big players becoming irrelevant quickly, which is why they bloviate like they do.