Unit Accord

Unit Accord

@unitaccord.bsky.social

This isn't a series of flukes. There's a pattern.

76 followers63 following326 posts

Remember when deepfakes were a major moral crisis? Everyone was Tom Cruise. And then... people just stopped doing it. That's AI in a nutshell. We think "because we can, we will". We project giant "addressable markets". But there's no market. People don't care.

If your problem can be explained in a short prompt and the solution is correct within the first few attempts, it only means you weren't solving anything novel. The ease of automation is deceptive. It only means you're spinning in place, achieving nothing.

Thought process behind "AI will drive huge growth": 👨‍🦱 "AI writes code in hours that took us months!" 👩‍🦰 "Great! What are you solving with this power?" 👨‍🦱 "Oh we're making versions of the same apps as before, but worse." Same with images. Music. Video. All of it. God help us.

We don't need AI which reads a 10 line prompt & writes 1 million lines of bad code. We need AI which reads 1 million lines of code & suggests a great 10 line patch. "Generative AI" has to go away & be replaced by "Contemplative AI".

A perfect mind is immutable. Input space, time and out comes the right answer, like a math function. There’s no reason to learn, to change your opinion, because your opinion is never wrong. And yet, there is no perfect mind. And our best minds are most open to change. 🤔

A perfect mind is immutable. Input space, time and out comes the right answer, like a math function. There’s no reason to learn, to change your opinion, because your opinion is never wrong. And yet, there is no perfect mind. And our best minds are most open to change. 🤔

Pluribus ain't bad, but it's annoying none of the characters are asking the obvious question why did the aliens send this. Zero chance it's benevolent, as it clearly makes humanity sitting ducks, exploitable & centralized. No one would contemplate "oh maybe it's better this way".

There are precisely two flavors of Internet right now. 1) Talking to a bot, you're always absolutely right. 2) Talking to a human, you're always absolutely wrong. There's nothing in-between. Neither side is correct. Good luck.

Watch what's happening with Tricolor and First Brands. Banks are exposed to billions of losses on these two. The financial collapse has started. First Brands claimed it has 1B cash, but were actually -10B in the red. Reminds you of someone's practices? Tesla's reckoning is near.

Say your strategy is to attract more gullible, therefore less stable people to your cause, because it's much easier to turn them to your side. But those less stable people also more easily turn against you on a whim. You take a big risk. Win big. Or lose big.

It's a genius scheme now, isn't it? You have a lot of unsold Model Y inventory, so you start a small Uber competitor, but dress it up as "autonomy tests". Thus you hide the bad sales, and you add fuel to your empty promises. The problem: camera-only FSD was always DOA.

Notice life is concentrated on the edge where two or more contrasts meet. Land, air, water. Most sea creatures live in shallow waters. Land creatures on the surface & near water bodies. This pattern repeats everywhere: biology, technology, economy. Exploitable energy gradients.

“Why should I hire you?” “I have the power of unnatural luck.” “Prove it.” “Watch as I keep plugging this USB cable in all ports the right way on first try.” “That’s USB-C.” “I manifested that standard into existence.”

People who say "I always thank AI because it may spare me when it takes over", even jokingly, depress me. Because they focus on fear and self-preservation, instead of respect to every other system, which we should have, whether biological, mechanical, electric or otherwise.

The media will keep the OpenAI / Jony Ive collab clickbait coming, but all they can offer is another variation on "smartphone / smartwatch, but worse". The whole point of a smartphone is you can run any app on it. Including AI. It's done. Weird form factors have no value to add.

REST was so bad at "automatic discoverability" that we literally invented *artificial intelligence* (which understands every API by default) before REST "automatic discoverability" could be solved. That's what an utter defeat looks like.