After 20 years, WordPress finally broke my custom theme badly enough that I decided to update it with a modern one rather than learn enough PHP to fix the old one. Just know that the slightly more modern look was thrust upon me–I didn’t want it. I hacked this one to make it look as bad as I could. I’m sad because Marginal Revolution has managed to keep their site looking the same for about the same length of time and they’re just a couple of economists. But there’s a limit to the amount of time I’m willing to spend on a thing that almost no one reads anyway and that Jonah rarely updates because she’s all Facebook all the time, now.
I need to switch the subscriptions to a thing that doesn’t have ads.
Now that ChatGPT does image processing, I wondered how it would do on the Reading the Mind in the Eyes Test. I’m pretty terrible at this test. Even for the ones that I get right, it feels like a random guess. I wish I’d recorded my score the first time I took it, but I can’t seem to find it. Anyway, I was curious how ChatGPT would do. Or if I could even convince it to take the test. It turns out convincing it was pretty easy:
ChatGPT scored a 25/36. The guidance provided with the test is:
The average score for this test is in the range of 22 to 30 correct responses. If you scored above 30, you may be quite good at understanding someone’s mental state based on facial cues. If you scored below 22, you may find it difficult to understand a person’s mental state based on their facial expressions.
I took it again after giving it to ChatGPT. There was a couple of hours in between most of it since I hit the ChatGPT limit and had to wait for it to reset. The test tells you what the right answer is after you choose, so I’m sure that helped. I scored a 20, which I think is significantly better than I did the first time took it.
I think we’re not far away from autism aids where your augmented reality goggles can tell you the emotion on the face of the person you’re speaking to is obviously, “Bored, just stop talking to him!” And then I can stop terrorizing whoever I’m talking to with my airplane stories without them having to be rude.
There’s a lot of focus on how generative AI is destroying humanity–I’m a lot more interested in the ways it’s going to help us be more human.
