←back to thread

Grok 3: Another win for the bitter lesson

(www.thealgorithmicbridge.com)
129 points kiyanwang | 1 comments | 20 Feb 25 07:15 UTC | HN request time: 0s | source
Show context
nickfromseattle ◴[20 Feb 25 12:02 UTC] No.43113694[source]▶
Side question, let's say Grok is comparable in intelligence to other leading models. Will any serious business switch their default AI capabilities to Grok?
replies(6): >>43113752 #>>43113762 #>>43114339 #>>43115136 #>>43115333 #>>43119800 #
ssijak ◴[20 Feb 25 12:10 UTC] No.43113752[source]▶
Why not?
replies(2): >>43113824 #>>43114190 #
chank ◴[20 Feb 25 12:17 UTC] No.43113824[source]▶
Because they already have something that works. Why switch if theres no advantage.
replies(1): >>43114057 #
1. tucnak ◴[20 Feb 25 12:47 UTC] No.43114057[source]▶
The API is compatible, and even if it weren't, it wouldn't matter anyway; everybody has been writing OpenAI API-compatible proxies, including Google, for months now. The only thing that matters is availability, throughput, cost per token (Google is ahead of everyone here: Vertex API is insanely cheap for what it does, Batch API at 50% discount, Prompt Caching at 75%, fully multimodal, better performance in multilingual tasks so actually useful outside the U.S. etc etc etc)