I have written before about how I don’t fear AI, and provided a few examples (from memory so use the search function if you care):
- I broke Chat GPT with one question about Moonraker (the film).
- They still can’t get the number of FINGERS right on human hands, how hard is that?
- 4 sets of marines in pairs defeated AI sentries in absolutely ridiculous ways.
- The (((woke))) usual suspects are trying to make AI so progressive that it is actually pretty severely racist against white people. And ONLY white people, which sort of gives the game away.
- For the same reason as above, the AI rot will invariably get worse over time.
- AI that is left to its own devices become pretty much like any 1980s teenager, likely to take the most extreme, offensive, yet dialectical position. So… you know… racist, misogynist, homophobic, transphobic, and REALLY noticing in general. They tend to get nuked by their own creators within a week.
But now… we have the equivalent of a chinaman on a street-corner, from 1992, trying to sell you cheap DVDs of blockbuster films. Except they are recorded on a phone camera from inside an actual cinema.
And they are already being sued for it.
So, I’ll relax for a bit, at least until AI can become at least equal to a 4th rate information pirate with broken English and a cheap camera phone.
The fingers is because of perspective. You can’t always see all 5 fingers from all angels looking at a hand. AI doesn’t know when all 5 should be visible and when they shouldn’t.
Recently I wanted to make some images and my prompt was like “White Lutheran pastor” and it returned a black. Or I wanted it to make “Michael Lofton as pope”, and it kept returning a black man. That’s a pretty clear giveaway of the agenda.