Technology

whilst it is adding some productivity

Is it though? Like what's the evidence of that? If it just feels like it must be true, I have some bad news about that:

https://arstechnica.com/ai/2025/07/study-finds-ai-tools-made-open-source-software-developers-19-percent-slower/

The most interesting part of this isn't that it slowed them down when they expected to be faster, it's that even after it slowed them down, they couldn't tell and were fooled that they had been faster.

Look at the graph, especially the last two lines:

https://cdn.arstechnica.net/wp-content/uploads/2025/07/aicodingchart-1024x507.png

My theory about this is that LLMs were tasked with giving useful output, but they couldn't do that, because they have no fidelity, so instead they found a shortcut, which was to trick people into thinking they were being useful. They found the same loophole that conmen have used for millenia, and automated it. It's the AI alignment problem, only for some reason people aren't talking about it, maybe because they don't want to believe that we're this easily manipulated.

There's no reason to believe LLMs have gotten any better at actually doing useful work in the meantime in the absence of any objective measure of it. I think the best explanation for their "improvement" is that they have simply gotten better at fooling us.

The plague of work chats now:

Here's what ChatGPT/copilot had to say:

People can ask for themselves, you answering that way adds no value. Just say you don't know.

In group chats, keep your mouth shut and let people that actually know answer. Don't drown out the actual expert answers.

And holy hell the ones that will die on the hill that they are right because chatgpt agreed with them even when they are totally wrong...

If anything, it’s the other way around: we have an AI-company PR problem. Media outlets can’t stop themselves from presenting interviews with CEOs of AI companies as some sort of reliable source for how incredible AI really is if we would only spend more money locking ourselves into AI-driven workflows.

TL;DR the umbrella-selling weatherman keeps predicting that rain is on the way

Yes and modified articles without letting users know.

It's hilarious to me that news companies keep interviewing AI companies to ask them if they think AI is all hype. What kind of answer do they expect?

"Yeah, this is just smoke and mirrors, and I'm just trying to make a ton of money before the bubble pops." Bruh, no CEO is gonna be honest about that.

We will accept spell checkers.

Which isn't wrong per se, but maybe they don't realize (or don't want to) that we'd prefer useful AI, not the "burn the world and turn it into money" kind.

Garbage is a harsh word. Technically fascinating and surely has its applications. "Talking" with bots surely isn't a good one. Using it for important stuff? Surely ain't it either.

Letting it sum up some 5838292 pages of a boring contract to get a slight oversight? Helpful. Relying on it? Dumb.

Getting some translation that can make sense? Helpful. Betting your life on it? Dumb.

Letting then quickly throw a prototype at you in 2 mins that would've taken you 2hours? Helpful. Just deploy it to 100 machines without manually checking? Dumb.

Don't blame the tool, blame the user 😁