There was a prophetic podcast episode from the series Plain English a while back that I constantly think about.
In that episode the author describes how the internet is going through a revolution.
Basically 20 years ago, the internet was all about gaining numbers. Companies could operate at a loss if they got people signed up. Facebook, Google, YouTube, Uber, Deliveroo, etc. they were all about getting you in their mailing list or consumer list and who cares what happens then.
Now there’s an issue because that model is not profitable. In order to continue, all the internet is moving towards subscription.
In a sense, I don’t think of that as intrinsically bad. Patreon is a good example. The internet is now filled up with so much shit that people are willing to pay to filter it. So with Patreon, you pay a fee to support an artist to produce the content you want. That itself isn’t a bad idea.
Now that being said, a lot of “bad things” do emerge. The fact that you can no longer buy software like Adobe and it’s all subscription based. That’s shit. But that also inspired software alternatives like Affinity Designer.
These things are interesting for two reasons (to me).
The first is that it seems utterly unsurprising that these inconsistencies exist. These are language models. People seem to fall easily into the trap in believing them to have any kind of “programming” on logic.
The second is just how unscientific NN or ML is. This is why it’s hard to study ML as a science. The original paper referenced doesn’t really explain the issue or explain how to fix it because there’s not much you can do to explain ML(see their second paragraph in the discussion). It’s not like the derivation of a formula where you point to one component of the formula as say “this is where you go wrong”.