I swear, the purpose of this blog isn’t to rant against LLMs. It’s just so difficult to escape that these days.
Anyway.
The author escapes me at the moment.
Perhaps you recall – if so, you can let me know on the fediverse.
Someone wrote the other day that AI is always deemed useful at automating other people’s jobs. Our own, not so much – it’s just way too complex and subtle, and no machine would ever understand this.
I’m mostly writing this post here to keep a note of this.
I find a few things interesting about it.
Ignorant
For one thing, it reveals our ignorance of other people’s work.
Of course we know all the ins and outs of our own, and if we find that difficult to grapple with it at times, it’s easy to come to the conclusion that our own job is too hard to automate.
When you think that someone else’s job would be a much easier target, instead of just assuming that’s true, ask yourself this: Am I having this thought because I know way too little about that job?
The answer is always yes, by the way, unless you have done that job yourself successfully and for a sufficiently long time.
Arrogant
The other thing it reveals is our arrogance.
It’s the definition of arrogance to judge our own importance and abilities so much more highly than those of others, and almost always this comes down to deep ignorance.
Antisocial
If the mixture of arrogance and ignorance is not the reason you come to such a conclusion, then you exhibit deeply antisocial behaviour.
Because what you’re really saying here is that your own job security is paramount, and you’re entirely willing to throw the job security of other people under the bus for it.
Let them lose their job, as long as you can keep yours.
Discomfort
There’s a good chance you don’t like something in the above.
Sit with that discomfort.