Your post took off significantly more than mine, but I asked something similar roughly 8 months ago here: https://news.ycombinator.com/item?id=41736600
Figured I'd post since seeing changes in responses over time can be interesting (if any of the typical responses have changed at least).
AI Agents used to be a fantasy, but they're completely the norm today. As in you tell the code to write what you want it to do - it'll make the changes then commit them. Just to show how far things have changed, the term "vibe coding" was coined in February 2025.
Also politics have had a significant shift. A lot of power vacuums opened up all over the world at the same time. This leads to some instability. Wealth parity will often dampen political instability. But that pillar was struck during COVID lockdowns. Crypto messes with IS-LM. AI is messing with productivity-wealth distribution in general.
It's like a small cut that got infected, and somehow that infection is heading towards sepsis. The 'sepsis' point might be bank runs. It can be cured. But are we doing enough?
During power shifts, tech is often the weapon wielded. When Americans fight AI and Chinese force it into the school syllabus, I guess we'll just see what happens.
I think the person you're replying to is being a bit alarmist, but in the context of AI, the models available 8 months ago were honestly significantly worse than what we have now.
Simon Willison did a great write up about the advancements of LLMs over the past six months, and it was a pretty cool read.
8 months ago was such a different world. I think we're starting to see the first hints of societal collapse, so people are taking it a lot more seriously now.