Once upon a time I was working for a university, a different one to the one I work for now... we had a conversation with some health spa consultants who were talking about this plan they had for a high end holiday clinic for wealthy folks on a nearby island.

They described at length the technology they were planning on using... a kind of corridor that they said a patient could walk down and while they were doing that a series of sensors would scan them, and then do automated analysis to determine a bunch of things about them that a regular human physician would be unable to. As the technology consultant, I was sort of impressed by the level of detail they described the tech with, because obviously I hadn't heard of anything that was capable of doing... several large parts of this end to end equation, from the detection to the speed of processing, to the type of processing. I asked them (I remember this word for word for some reason) "the technology stack you've described is quite compelling. Where are you getting it from? Who is supplying that?"
They responded "oh we were hoping you could help with that part".

It is possible the author of this article has underestimated the challenges in a similar way. Do not believe the sales folk when they describe what they imagine the technology might be like. They understand narrative, but somehow do not understand that changing the narrative does not change the world.

What do you think "it's premature to handwave" means?

mmm... I think you're conflating two separate types of processes here. Being able to do statistical analysis on token relationships isn't the same thing as surveying chat logs, and even if you are using tools to simplify searching them you're still going to need to know something exists before you can look for it. They still have a limited amount of attention, too.

Regardless, I don't think they would have cared that much if they could have determined it.

every conversation huh

what's the rate on that?

As someone who manages something 1/100th of the size, I'd absolutely believe that they missed it.

Don't get me wrong... I'm not saying you're wrong that they would make this decision given the functional opportunity, I'm saying that they have a vested interest in making you believe they actually vet more than a tiny percentage of conversations and they haven't actually demonstrated the capability.

Well that's an assertion.

We know exactly what they're capable of. They take string chunks and determine the probability of various other chunks, based on their related frequency in datasets they have ingested. They can be made to take in other data in similar ways to show similar relationships to the data they have ingested, even when the dataset being examined is noise.

These patterns are not automatically meaningful.

The operative part of your 2nd sentence is the 'can be'.

It is premature to handwave this, particularly when so much of the market is made up of grifters who have already made impossible claims.

too real

"I have a hammer. Please validate my belief that all your problems are nails and praise me for it."

I'm not sure I understand your question.

Can you explain what you mean? It seems like you're making some kind of emotional appeal, but for it to make sense I'd have to be focusing on the AI more. Not less.

Can you please explain to me what is fun about it?

No, I'm not being facetious. Not deliberately anyway. I genuinely cannot fathom what is fun about competing over which is the best language, and a list like this one just makes it more absurd.
If a person is arguing for using javascript to access an SQLdb, for example, I might simply be confused why they're very obviously trying to use the wrong tool for the job.

What is the point of comparing languages as if they're sports teams?

I have never understood this. It seems like the most irrelevant thing.

Legit reminds me of soda ads trying to vilify 'flavorless' water.

"Designing an algorithm for lockin" when you've decided that you only want customers who aren't capable of understanding your product but think it's important for vague unstated reasons... does seem somewhat counterproductive at the very least.

Before that, people used to write "for media".

Based on your observation though, I now see this as techbros discovering that demographics are an established method of reducing populations for ease of analysis, and predictably confusing the map for the territory.

Absolutely. It's one of the major reasons I think they (and other plans like theirs) will fail, with one of the others being that they think they can just tell people to implement whatever they imagine and it will get done... often despite the thing being contradictory in function or an unsustainable, tautological assertion of a demagogue's beliefs.

They see the world largely through a narrative lens as far as I can tell. Function is someone else's problem, and they reserve the right to get mad and yell if their imaginings aren't made real, etc.