On with the scary AI saga. Twitter is abuzz with discussions of the new OpenAI thing that generates text. You can give it any prompt, and it writes up a piece or comes up with an idea. In some ways, pretty scary stuff! E.g. I saw a piece generated by the prompt "Write instructions on how to remove a peanut butter sandwich from a VCR, in the style of the KJV bible." It was actually quite hilarious, and perhaps as good as anything a human stand-up comedian could come up with.
Well, you can go join the fun at https://beta.openai.com/
- it's easy (and free) to sign up and start playing around.
The good news is that if current trends continue, you need not worry that artists are going to be out of jobs while everyone else comfortably coasts along. No, EVERYONE will be out of a job, except some classes of artisans and manual laborers. So at least we won't be alone.
In any event, the software, while scary impressive, is not as miraculous as it appears at first sight. I thought out of the box, and gave the AI the same prompt over and over: "Tell me about the first time you tasted ice cream." Here are its responses:
View attachment 26673
You can see what's going on here: it's essentially the same response over and over. And it's a "stock response" - I would imagine that the most common actual human response to the question would be the same as my own, personal response: I honestly don't remember.
As amateur cartoonist I worry of course that AI cartoon ideas will quickly put me out of a hobby. So I asked it to come up with ideas for single-panel cartoons. Many of these fell flat, and I don't think any were as funny as my own best ones, but the scary thing here is that the AI can generate unlimited numbers of them, and beat a human by sheer brute force. On the other hand, here too there is a somewhat "stock response" quality to the ideas:
View attachment 26674
Still, scary stuff, considering this is but the beginning. Oh, and you will be particularly disturbed to learn that when I set up a hypothetical legal trouble scenario and asked the AI for legal advice, it told me to go to a lawyer, so I fear that while AI might free us of artists, we're never going to be rid of lawyers!
There's a deep irony here. In the 1980s naïve technomaniacs like Carl Sagan promised us that robot technology would "free us of backbreaking labor." It now turns out that the very last jobs remaining to people will in fact be the backbreaking ones. Everything that we thought of as too creative or original or soulful for computers to do, turned out to be the low-hanging fruit! We were promised Star Trek
; instead we're getting Terminator
But of course, one doesn't really know. No one can predict the future.