What is the best way to find the PDF version? I've seen some free, some paid, but wondering if anyone has a reputable place to download book PDFs? I'm happy to pay for the PDF!
Very interesting and great post. A basic question—is what you upload ever used for training or leave your walled garden? I cannot seem to get a clear answer on this. I ask because I am a litigator and would love to upload a large hearing transcript or depositions and then ask questions but obviously if those are confidential docs I cannot do that if they don’t remain confidential. And, if it does not remain private anyone know of a reasonable way to do this? Thanks all—great to have all of your collective brains (and hope to be able to opine at some point but will wait for questions about poptarts and figure skating).
Just heard about how to do this relative to ChapGPT (each model differs on how you do this, and some automatically do not use your docs for training, as you may already know). It was from the podcast 'Leveraging AI,' and it was mentioned that you need to make your request in writing, and they will notify you when it has been implemented. I would do some searching as I believe there is now a setting that may do this, but I have not found it.
I’ve found adding instructions in the prompt to reference or heavily bias toward the uploaded knowledge files “seems” to work (in ChatGPT at least – it says it’s “Reading documents”, although it appears to have a cap of ~8 docs)
Great idea! I’d be curious to see how the responses differ if you used the same prompts in a standard ChatGPT chat without the PDF upload. Sometimes I feel it doesn’t actually take my uploads into account. Will do some more experimenting!
This is incredibly helpful for people who fear interviews - you can effectively ask to get the hardest questions thrown at you ahead of time.
What is the best way to find the PDF version? I've seen some free, some paid, but wondering if anyone has a reputable place to download book PDFs? I'm happy to pay for the PDF!
Yeah this is a really cool life hack
Very interesting and great post. A basic question—is what you upload ever used for training or leave your walled garden? I cannot seem to get a clear answer on this. I ask because I am a litigator and would love to upload a large hearing transcript or depositions and then ask questions but obviously if those are confidential docs I cannot do that if they don’t remain confidential. And, if it does not remain private anyone know of a reasonable way to do this? Thanks all—great to have all of your collective brains (and hope to be able to opine at some point but will wait for questions about poptarts and figure skating).
Just heard about how to do this relative to ChapGPT (each model differs on how you do this, and some automatically do not use your docs for training, as you may already know). It was from the podcast 'Leveraging AI,' and it was mentioned that you need to make your request in writing, and they will notify you when it has been implemented. I would do some searching as I believe there is now a setting that may do this, but I have not found it.
Much appreciated!
I’ve found adding instructions in the prompt to reference or heavily bias toward the uploaded knowledge files “seems” to work (in ChatGPT at least – it says it’s “Reading documents”, although it appears to have a cap of ~8 docs)
Great idea! I’d be curious to see how the responses differ if you used the same prompts in a standard ChatGPT chat without the PDF upload. Sometimes I feel it doesn’t actually take my uploads into account. Will do some more experimenting!