But can you inject ads into the answer?
Comparing Google and ChatGPT
21–30 of 280 posts
Re: Comparing Google and ChatGPT
#22Let the market decide, these are cherry picked examples as far as anyone should be concerned.
Re: Comparing Google and ChatGPT
#23Earlier quoted context omitted.
Exactly, i talked to ChatGPT and it gave me a lot of wrong information in an authorative tone. I consider it dangerous as-is.
Turns out humans do this all the time and they actually have real power.
Yes, just look at the twitter thread, most people are not even noticing that the answers are wrong.
Re: Comparing Google and ChatGPT
#24I think these are 2 separate use cases, one for organized knowledge and one for related links. Google doesn't compile knowledge as well, but it does good job on finding related links.
Re: Comparing Google and ChatGPT
#25I wonder what this will do to misinformation. Seems like the next big culture war will be over AI. What seems very Utopian will quickly be framed as dystopian. If AI doesn't promote "opposing positions" it will definitely become the target of politicians ire, if not outright banning as propaganda. For example, what would AI say in terms of the effectiveness of Ivermectin in combatting COVID-19? or Vaccine injury rates? Would AI argue that lockdowns are the most effective measure against a spreading pandemic?
Re: Comparing Google and ChatGPT
#26It's great, until people realize GPT-3 will generate answers that are demonstrably wrong. (And to make matters worse, can't show/link the source of the incorrect information!)
>until people realize GPT-3 will generate answers that are demonstrably wrong
It isn't like google never returns the wrong answer
Re: Comparing Google and ChatGPT
#27Like Google can't acquire this or something similar and include it in its results if it's actually a threat.
It's not that they can't acquire. In fact they invented this tech and have their own models just as good. But the "problem" is you can run such a model on a computer of your own, like Stable Diffusion. And this model could interface between you and the web, doing the final part of question answering. Then you are not forced to see any ads. So the language model has the potential to free us from their ads. You can download a language model, you can't "download a Google".
If you don't think you can run a decent language model on a normal computer check out Google's own FLAN T5 series. Local language models mean more privacy and empowerment for everyone.
Re: Comparing Google and ChatGPT
#28What are the engineering and considerations for serving this sort of model to billions of queries a day? Do the economics of a gpt-as-a-search-engine work?
Re: Comparing Google and ChatGPT
#29It's great, until people realize GPT-3 will generate answers that are demonstrably wrong. (And to make matters worse, can't show/link the source of the incorrect information!)
Google already does this
Re: Comparing Google and ChatGPT
#30Another person who doesn’t realise AI language models are just making shit up. Google results are quite often full of wrong information, but at least it has mechanism for surfacing better content: inbound links, domain authority, and other signals. It doesn’t guarantee correctness, but it’s better than the pseudo-authoritative fiction GPT-3 and friends come up with.