Great use case of AI. Almost 4 years after the sensational release of GPT3.5, the best use case of AI is still being a powerful search engine that can gather information from all corners of the digital world.
This. They still make a lot of mistakes, but their working memory and processing of a massive corpus of data are obviously beyond what any human could ever do. But when they flag something interesting for expert review you can do some really good stuff with it.
Slightly tangential: One way I found of limiting impact from hallucinations was to isolate data from the transformation process. For example, I was making some flashcards for astronomical data and tried to get an LLM to put it all together in Typst, but correcting the hallucinations got to be too much to handle. The solution was to have it write a Typst template and also a Python program which would convert the ephemerides data fields to a Typst file sourcing the template. Worked flawlessly.
I've been using AI to help with my genealogical research, and it's been fantastic. I am loving pushing the family history back and catching mistakes that someone who many others share as a common ancestor have made.
This is a strong application of LLMs IMO. They are idea machines. Historical ways of thinking give us different paths to explore the world through.
I'm guessing it wasn't Claude 5.5, not with the "dangerous" nature of this work.
Great post, thank you! So, let's resolve ancient Near East chronology and determine an absolute date for the sack of Babylon?
Great use case of AI. Almost 4 years after the sensational release of GPT3.5, the best use case of AI is still being a powerful search engine that can gather information from all corners of the digital world.
This. They still make a lot of mistakes, but their working memory and processing of a massive corpus of data are obviously beyond what any human could ever do. But when they flag something interesting for expert review you can do some really good stuff with it.
Slightly tangential: One way I found of limiting impact from hallucinations was to isolate data from the transformation process. For example, I was making some flashcards for astronomical data and tried to get an LLM to put it all together in Typst, but correcting the hallucinations got to be too much to handle. The solution was to have it write a Typst template and also a Python program which would convert the ephemerides data fields to a Typst file sourcing the template. Worked flawlessly.
I've been using AI to help with my genealogical research, and it's been fantastic. I am loving pushing the family history back and catching mistakes that someone who many others share as a common ancestor have made.
I also have been doing this every round of models releasing.
Thanks for reminding me to try this new round of models.
I wonder if there are some good tools/apis that can help here.
A lot of information is often login gated