Only they’re not “answers”. AI does not provide information, it tells you what you want to hear. It even makes up fake scientific papers as “references”. So you think your info is backed up by peer reviewed sources. Nope!
The better models absolutely will cite real scientific papers that are relevant to your discussion. It doesn’t hallucinate as a matter of course, it can hallucinate some of the time. You can’t trust what it gives you 100%, you have to verify it. But it very often gives useful answers and information.
It even makes up fake scientific papers as “references”
Broski, it’s not early 2023 anymore. Thinking models with internet access have been able to provide public sources just for over a year now. Technology evolves; shocking, I know…
They do far better today than they did three years ago, but they still get stuff wrong constantly. I’ve had it give me the source for claims, and the source says the exact opposite of what it claimed. This isn’t a quantized to hell model running on my own hardware, but Opus from Claude.
My go to is to get people to ask it something they know a lot about, and see how bad inaccurate it is. If it gets something you know about so wrong, how wrong is it for everything else?
Only they’re not “answers”. AI does not provide information, it tells you what you want to hear. It even makes up fake scientific papers as “references”. So you think your info is backed up by peer reviewed sources. Nope!
One of my favorite economics bloggers for decades started sourcing his information with LLMs. He was such a bright guy. 😭
The better models absolutely will cite real scientific papers that are relevant to your discussion. It doesn’t hallucinate as a matter of course, it can hallucinate some of the time. You can’t trust what it gives you 100%, you have to verify it. But it very often gives useful answers and information.
Broski, it’s not early 2023 anymore. Thinking models with internet access have been able to provide public sources just for over a year now. Technology evolves; shocking, I know…
ok, but they will give you a different answer to a question a day later if you delete previous prompts.
They do far better today than they did three years ago, but they still get stuff wrong constantly. I’ve had it give me the source for claims, and the source says the exact opposite of what it claimed. This isn’t a quantized to hell model running on my own hardware, but Opus from Claude.
My go to is to get people to ask it something they know a lot about, and see how bad inaccurate it is. If it gets something you know about so wrong, how wrong is it for everything else?
That’s just not true.