ChatGPT will likely play out exactly as innovator’s dilemma suggests. To an expert in any given field, it has worse answers. But most people don’t have access to experts for everything, so it actually is a productivity boost to everyone else.
@ronald_istos@solyarisoftware Interesting. I've not seen many examples of it producing personal data so far but considering the size of the dataset there must be plenty in there. I wonder what mechanism they have for preventing (or at least discouraging) it from doing so...
#ChatGPT can not call your custom #APIs.
A workaround is to give it in advance (into the initial session prompt) your custom API's responses (e.g. a query to your system) as a list of facts, letting it to handle user queries.
What do you think?
See the running example:
Why does ChatGPT work so well? Is it “just scaling up GPT-3” under the hood? In this 🧵, let’s discuss the “Instruct” paradigm, its deep technical insights, and a big implication: “prompt engineering” as we know it may likely disappear soon:👇
@daviddlow Sounds similar to co-pilot which also can be integrated into vscode via a plugin. For someone that writing code all day isn't the main part of their job I find it a huge boost but it's probably less beneficial for more experienced developers.
A very cool feature of ChatGPT is if you give it a coding problem and it gets it wrong, you can tell it that and it will try to spot where it went wrong and come up with the correct answer. I got this to work over 6 iterations until it found the right answer!
@markhurrell Yes, that's almost how I'm thinking about it. Tell it the inputs and the result you want and then hit generate until it works. Could be a massive boost for designers who want to own more of the stack.
ChatGPT is the closest model we have that looks like AGI, but still suffers from hallucination and tends to write plausible-looking bs with uncanny confidence. It could be misleading and even dangerous. Can we fix it? There is a simple & intuitive remedy:🧵