Wikipedia was useful for me as a grad student because I could look up a topic and there would be a whole lot of citations I could follow. I never used them as a source, but rather as a curated forum of information.
I’ve been doing exactly the same thing with LLMs recently.
"Tell me about "
“What are the big problems their industry is trying to solve?”
“Who are their biggest competitors?”
“What’s the worst/best thing about them?”
Questions like that often give me a great framework to look up specific questions, find relevant articles and get a handle on the sources that are likely to be useful.
Very careful. I never use anything from them directly. I just use them to give me a starting point on what to look for.
For example, if the AI tells me that some company is know for their low latency database, I’ll look around for primary sources on the latency of the database compared to other vendors. I’ll also look for evidence to the contrary.
You’d be impressed by how good I was at finding PDFs of original articles on random sites. Turns out that when you go to grad school in the third world and don’t have access to the journals in the same way as you are accustomed, you learn how to do it for yourself.
This was back in 2010-ish. Don’t know if it works the same way now, there are probably better ways to go about it. Not sure if sci-hub existed at that time. Will remember that going forward.
Wikipedia is like our dear friend. It gives us general information, good advice, and direction in life, but never gets too deeply in it. The choice is ours to make.
Wikipedia was useful for me as a grad student because I could look up a topic and there would be a whole lot of citations I could follow. I never used them as a source, but rather as a curated forum of information.
I’ve been doing exactly the same thing with LLMs recently.
"Tell me about "
“What are the big problems their industry is trying to solve?”
“Who are their biggest competitors?”
“What’s the worst/best thing about them?”
Questions like that often give me a great framework to look up specific questions, find relevant articles and get a handle on the sources that are likely to be useful.
I’d definitely be careful about made up stuff, but this sounds like an interesting idea.
Very careful. I never use anything from them directly. I just use them to give me a starting point on what to look for.
For example, if the AI tells me that some company is know for their low latency database, I’ll look around for primary sources on the latency of the database compared to other vendors. I’ll also look for evidence to the contrary.
Only problem is that half of them are broken :(
You’d be impressed by how good I was at finding PDFs of original articles on random sites. Turns out that when you go to grad school in the third world and don’t have access to the journals in the same way as you are accustomed, you learn how to do it for yourself.
sci-hub
This was back in 2010-ish. Don’t know if it works the same way now, there are probably better ways to go about it. Not sure if sci-hub existed at that time. Will remember that going forward.
Wikipedia is like our dear friend. It gives us general information, good advice, and direction in life, but never gets too deeply in it. The choice is ours to make.