foo for thought

sigh. I wish it would take them longer to make these models. Claude Fable 5 is, amazing. Not only can it one-shot smaller features, but larger, multi-faceted implementations. Claude Opus 4.8 was already good – Fable 5 is just better and faster. The context is what I think is the key. The most critical part of the AI war. Whoever can loop up the most “context” the most efficiently will win.

Now, why haven’t you heard about the doom posting from Fable 5 like with other models, like Opus 4.6? Oh that’s easy, probably because people are too busy using it to talk about them. That’s the scariest part of all of this actually, is that people aren’t talking about this.

I’m seriously worried about the future of engineering. Not only could this model alone replace millions of jobs, but future, better models could replace millions more. The more context you give these AIs, the more tools + integrations you allow it to have, the more complex issues it can tackle. The layoffs that we’re seeing in big tech I think is only going to go up – and not just in tech. With the rise of applied AI, I think we’re going to see cuts in other entry-level positions in the coming years.

I think it’s accurate to say that, in the words of a Microsoft higher-up I had the pleasure of speaking with a couple years ago, AI shaves off most of the bottom part of the pyramid, a bit off the middle, and none of the top. The best off in this AI-ruled economy are going to be the people who can most effectively utilize this technology, unfortunately. Now, that’s not to say it’s skill-less, in fact I’d argue the people most successful with AI have a background in engineering. This is just because, engineering a perfect prompt is, at its core, engineering. Building something from nothing, buffing out the tweaks to make a final product – that’s what we already do with code, now we just have to do it with our words.

But, and I think this is fascinating – I think a lot of other people are also good with their words who aren’t engineers. Software engineering always had this syntax overhead – you couldn’t code unless you knew the millions of garbage jibberish characters that are the primitives of a coding language. Now, there’s zero overhead. Anyone who can speak a language, regardless of language can be an engineer. You don’t need math or physics either, you just have to know how to talk. Again that’s not to say this isn’t a skill, you still have to be an engineer, it just doesn’t require any pre-reqs anymore.

And another glimmer of hope – AI still can’t do art. At all. This is honestly the one thing keeping me going. I love art, music, everything related. I have this moonshot theory that, if AI takes all the white collar stuff and the economy collapses, we’re going to have an artistic renaissance! Because, if AI can do everything for us, we’re going to need to place value in something else, something that AI can’t do. The same, funny enough goes for blue collar jobs. There’s a lot to be said about physical labor – robotics has gone far, especially in China, but hardware just cannot keep up with software.

So, yeah. I don’t quite know where this is going, either. These are just my unfiltered thoughts on the AI takeover. If you’re in a white collar job, learn to be really good at the thing, good enough to explain it to other people or even a computer – and also keep up to date with these AI tools. Your future robot employer will thank you.