The "Grounding" section at the end of this blog post makes the whole read worthwhile. Most interesting combo of life and software writing I've seen in a long time...
Perhaps this is because it's meant as an oversimplification to demonstrate, but the grounding section seems so oversimplified as to be useless to an adult.
Hard disagree. Most normal people do not want to see others’ jobs automated unless it provides obvious benefits. Nobody likes self-checkout lines, but they’re “automated.” I like talking to interesting Uber drivers.
Pharmacists though? Yeah replace them with face-ID prescription vending machines…
I like self-checkout lines, as do I'm pretty sure most people I know. And I’ve never met a person who consistently enjoys Uber driver conversation; in fact I’ve often seen it used as the canonical example of a human interaction people hate.
People didn't like the early self-checkout machines which were horribly slow and inaccurate. Current models are quite slick and many people prefer them.
>Nobody likes self-checkout lines, but they’re “automated.”
I strongly prefer self-checkout. If self-checkout is available, I will use it. Most people I know have the same preference (to varying degrees, I am a bit extreme in that I will wait for self-checkout even if normal checkout is open and there is no queue).
I like self-checkout lines. I also don't want pharmacists replaced with face-ID prescription vending machines. It's possible that I'm not "normal", but I feel normal.
This is the typical AI booster’s game: make vague claims about what AI can already do today without showing one single example, then extrapolate to an imaginary future and say that we’re all too stupid even to debate him.
Instead of writing some post like this, I’d love to hear ONE SINGLE EXAMPLE of AI doing a core work task faster, better, and cheaper than a person can do it, now, today. Here’s the task. Here’s how people did it before AI. Here’s how they do it now. Here’s how much time or money they save.
reply