I just don’t see it. All the doomerism about AI seems nuts to me, as I think these LLMs (Large Language Models) and what they are capable of doing is astonishing.

Yes, the amount of software code that is being generated by AI is astonishing now. Adoption among a certain tech-savvy user group since the release of the latest models, Anthropic’s Opus 4.6 and Moonshot’s Kimi 2.5, has been incredible. But even there I don’t see a reason for software engineers to panic.

The software still sucks. It’s buggy as hell.

I’m an enthusiast for OpenClaw, and trying to make your own personal assistant from AI on your own home computer network. This is a tiny hobbyist adventure right now. There will be OpenClaw millionaires soon. But again, they’ll be managing slightly patchy, taped-together software systems.

We’ll all soon be using these AI tools, just the way we all started using the basic internet search. But I think AI is going to look more like the Internet of Things. Most of us aren’t controlling our house blinds and other appliances through voice commands over the internet. But slowly, the Internet of Things is yielding productivity gains in industry and slight changes of lifestyle at home.

© National Review