I do a lot of work synthesizing information. Much of fundraising is taking what exists in the world and connecting the dots of how a person or institution and their philanthropy aligns with an organization's goals and values. This involves a lot of reading and listening and sorting data. It's tempting to use LLMs to outsource this work, because that's what the hype machine says it can do! But it cannot *actually* do this reliably.
Routinely tasking LLMs with summarizing non-important, non-private information, it becomes obvious that there is nothing valuable for me to offload here. You cannot depend on these things without accepting that what they produce is at best, distorted, at worst, completely fake.
I guess that's fine for some — the *appearance* of work, of depth, and trustworthiness is all that matters! — but not for me.
If you have a fediverse account, you can quote this note from your own instance. Search https://social.coop/users/al/statuses/115488518951445692 on your instance and quote it. (Note that quoting is not supported in Mastodon.)