I want to stress the fact that no LLM is the holy grail at the moment. We use Perplexity for search, GPT4 for GPTs and pure performance, Gemini advanced for writing.
But each has their downsides…
Gemini’s biggest flaw was unveiled recently in its image generation and it seems the X algorithm is pushing posts criticizing the “wokeness” agenda from Google to try to bash it.
Google has their flaws, but ultimately they possess the data to win in the long run.
So revoking their image generation of ppl is good in the short run, but we’ll see how they respond.
Lots to anticipate this year with Gemini 1.5 Pro (1mil context) and GPT 5.
Have a lovely weekend all.