Yesterday, I wrote about the Apple Vision Pro and how to think about its prospects for success.
Being human, I made one glaring factual error. It doesnโt affect the substance of the argument. It is more embarrassing for me. The sort of brain freeze when you claim Madrid is the capital of Portugal, knowing perfectly well it isnโt.
But every error is an opportunity to learn.
The error was that I said Dan Bricklin invented Lotus 1-2-3. He, of course, created VisiCalc. I even used VisiCalc on the family Apple II. Mitch Kapor developed Lotus 1-2-3, something I should have known, having met Mitch a few times and having developed my first useful, working piece of software in Lotus 1-2-3.
But here lies the power of todayโs LLM-powered query engines. I pasted my essay into Perplexity with the following query:
Please review the following essay. Please identify any factual errors and explain the error citing reliable sources.
Perplexity did a fantastic fact-checking job. Here is the beginning of its comprehensive response.
It identified a couple of factual errors and a few places where it felt suggested the assertions I made needed more evidence. (I am comfortable with what I have written.)
The author states that Dan Bricklin developed Lotus 1-2-3. This is incorrect. Dan Bricklin, along with Bob Frankston, developed VisiCalc, the first spreadsheet program. Lotus 1-2-3 was developed by Mitchell Kapor and Jonathan Sachs.
The author mentions that the original iPhone didn't have apps. This is not entirely accurate. The original iPhone did have apps, but they were all developed by Apple. The App Store, which allowed third-party developers to create and sell apps, was not introduced until 2008, a year after the original iPhone's release.
You can read the rest of its assessment here.
In future, weโll make sure our editorial process includes this additional layer of fact-checking. Why not use the superpowers when they are available at our fingertips?
Thanks to all the patient readers who drew attention to it with the appropriate degree of kindness.
cheers,
A
I LOVE this, thank you SO much for sharing it! I am using it right now to review a big set of answers I'm compiling for the UN's ResilienceFrontiers.org project on key inflection points in the rise of chemical agriculture. I'm making extensive use of ChatGPT ("drill down into more detail, more") and am then running my edited outputs through this prompt in Perplexity. Everything was looking accurate and validated so to test it, I threw in a bullet point claiming that the among the key drivers on a cultural and behavioral level, the popularity of The Muppet Show led to a significant increase in demand for the output of chemical agricultural processes. Perplexity totally caught it and busted me for it! Process validated!