A speaker argues that no single AI company will dominate search. Instead, different models will specialize by use case: Perplexity Pro, Claude, Llama, and Gemini each excel at different tasks, with Gemini handling massive context windows up to 1 million tokens.
Mike says, I'm a librarian and after using Perplexity Pro for two months and Google Deep Research for five days, I think the days of Googling are limited. I totally agree, but I still think Google is probably going to come out a winner. And I said a winner, not the winner. I don't think in AI there's going to be any single winner. I think they'll all sort of fall into like a bucket of which each one is best at, and people will decide what use case they want to use it for and pick whichever one is best for that use case. But I don't really think like OpenAI is going to win or Google is going to win or Meta is going to win. I really do think that they will all kind of have their reasons to use that one specifically. For instance, Meta is. Llama is open source. You can install it locally. There's going to be a whole bunch of new, like, iterations and spinoffs of it that other people develop on the back of Llama. Claude right now I think is still one of the best at writing code. Although a lot of people have said the new Zero1 model has gotten really good at code. I haven't tested it for code yet. You know, Gemini has the longest context window. Right now you can use Gemini with up to a million token contexts. So right now each of them kind of have their pros and cons. We are going to see a lot of them sort of level out and build features in. But I still think people are going to go to certain tools for certain use cases.
