I think LLMs+search like Bard, Bing, Phind.com or Perplexity.ai are decent replacements to bare search. Average web content is worse than GPT stuff. You can see

I think LLMs+search like Bard, Bing, Phind.com or Perplexity.ai are decent replacements to bare search. Average web content is worse than GPT stuff. You can see that when they compare models trained on raw web vs other sources (books) or curated web. Using bare search means you have to wade through the garbage, a dumpster diving-like experience. You still need to validate everything, on the web or in a LLM window, and that still requires you have sufficient experience in that topic.