Hi there, interested to know how your Live Crawling feature differs from ChatGPT Plugins such as SEO Assistant. If it is significantly more capable I might be interested in an elite plan while the sale is going on. Anyone wanna give me some opinions on how useful it has been for them in comparison?
if you are not using live current data you are creating fake data!!! seriously people!! you cant prompt all day with fake data and rank for data that came from where??? an AI that is 2 years old and doesn’t give 1000 words on a good day. ok keep doing it hahahhaha
That’s not really correct. You have some solid facts in there, but not a full understanding of them, of how AI works, and what you do have is all a bit jumbled up. That’s kind of what happens when AI gets stuff wrong too.
AI is mostly the result of training data, which is millions of documents, most of them hand-picked and selected and vetted. Everything the AI is is from that training and a core framework. That includes its ability to understand what it is prompted with, or supplied via any kind of data connection or injection.
So, if you think the AI without a live connection is stupid, then you are relying on that exact level of stupid to even comprehend what it gets from a URL. That’s a fact. Looked at that way, I’m sure you start to see how ridiculous the assertion you made, in the way you made it, actually is.
Furthermore, the AI does not learn anything new from what it crawls. It isn’t able to. OpenAI carefully designed every aspect of the AI to resist any attempts to train it further. Instead, anything it is prompted with, or crawls, goes into a temporary memory, just like the context fields. The AI is allowed to read it, to use it as temporary context, but not to learn from it.
The reason for that hard limitation is the same as the reason it wasn’t allowed to access URLs at all at first - to prevent retraining or corrupting the AI’s mind. The AI scientists learned this lesson the hard way with Tay. Tay (chatbot) - Wikipedia
OpenAI CEO and several of their scientists talk about this at greater length in the interview we linked to in Sam Altman (OpenAI CEO) talks about promises, risks, and fears
1 Like
This is a good start, but I’d really prefer a demonstration, the examples are too blurry to see and his explanations are confusing. In fact, several of the comments below the video also ask for a demo.