The monstrous AI token costs are finally catching up
The monstrous AI token costs are finally getting to you, and normal people are starting to realize that AI is not magical, it’s definitely not “free” and it doesn’t replace people like we were told it would.
I see many people getting sick of the slop flood, constantly being lied to/fooled by misinformation and the always confident AI hallucinations, not to mention the weekly “apocalyptic” news and narratives. I’m personally sick of all that.
Many intelligent people got fooled. Myself included, many times I wondered: “how could I survive in the future and feed my little son, in a neo-feudalism like dystopia”, hahah it’s funny to think about this now, but I’ve lost some good nights of sleep thinking about this, it hurt.
I see folks realizing they took BAD decisions, like wrong career paths, bad deals and ended business relationships, out of fear from the broken doom promises of big tech sociopaths.
The promise vs reality of AI productivity
They promised work was gonna change and we would do much more while working much less. At the same time, I feel I’ve never worked so hard before, but that extra effort DID NOT translate into more income. I feel rushed all the time, feel that I’m not doing stuff fast enough, now that most people fell for the narrative and think everything should be cheap and instantaneous and everyone should have many roles.
I constantly felt that I (founder of an AI-native company) was not learning the new tools and workflows fast enough.
The realization: nobody knows what they’re doing
Then I came to the conclusion that NO ONE really is. The tech is still so new and changes so fast, NO ONE really knows what the heck they’re doing with it. There are no stablished and reliable AI processes for coding, designing, managing or anything else. No single person figured it all out.
The “Every company should have an OpenClaw strategy” problem
Jensen Huang, CEO of Nvidia, biggest company blablabla, said in a conference that “Every company should have an OpenClaw strategy.” I bet he didn’t even try it himself.
Well, I did and it’s awful, a nightmare to set up, a huge security risk and a total time and money waster, a terrible, non-scalable and wasteful way to automate processes that could already be automated with simpler and cheaper tech like n8n.
I think the better statement Jensen should’ve made is “Every company will need to invest in something that leads to me making money” hahaha
Growing tired of the hype cycle
I’ve grown tired of testing new tools that promised to be paradigm-shifting, and in practice are just soulless and mostly useless sloppy machines, that cost me time and money just to cut through the hype and form my own conclusions about them.
AI is great, but it’s time to be real
I do think AI is the greatest tech of all time, but it’s time we start being REAL about it.
It’s very hard to profitably implement AI processes in a small company like mine, I can’t even imagine what it is like to do that for a company with thousands of people in it, and not lose control.
The fundamentals didn’t change
The current AI narrative is a TRAP. The fundamentals of business, design and development DID NOT change. Pretending they did because of AI is gonna backfire HARD.
AI got good enough to make it look like anyone can build an app, be a graphic designer or write converting copy and content.
On the surface it looks like all that’s true, but it’s not. The vibe-coded apps suck and break, the sloppy design is obvious, noisy and tells your customers you don’t care about them, and — the — AI — copy — is boring and no one cares to read two lines of it.
The danger of skipping expertise
Skipping processes, trusting AI outputs blindly, trying to wear many hats and do other people’s jobs is a recipe for failure.
And the worst part of this is that you end up wasting the time you could actually be using to do the things you’re actually good at, by doing stuff with AI that you can’t even understand or judge the output of.
That’s insane.
The bitter lesson
The bitter lesson I’ve got from all this chaos is:
If you can’t judge the output, DON’T do it, just don’t, it WILL be a waste of time and tokens (and those are getting very expensive).
Don’t try to code with AI if you can’t control/review the output.
Don’t design with AI if you can’t judge if the result is actually useful.
And if you’re writing, at least write it yourself and only use AI to polish it, own your words and let people see there’s a human behind them.
AI should augment and multiply, not replace, let’s collab more, peace!
Side note
I created the cover img with a fluted glass shader because I didn’t want to generate one for this haha