World 1: 40,000 people die a year from traffic accidents, but the driver is always a person. The courts may or may not find them liable, but the victims’ families at least have a name and face, and an explanation of events and motivations that they can rely on to find peace.

World 2: 4,000 people die a year from traffic accidents, but the driver is always an opaque ML model. Nobody is named, let alone found liable. And even if you had a name, there is nobody who can explain the reasoning of the model to you.

I am not sure World 2 is unquestionably preferable.

I think we’re missing something about justice when we only look at how many people die and not also how the survivors are able to make sense (or not) of their loss.

Also, like… the fact that it has a compiler. It’s like saying C is incompatible with assembly because you can’t yeet a .c file at an assembler.

Yeah, that’s uhh… that’s technically true for very strange and specific definitions of “compatibility”… but it wildly misses the point of TypeScript. Not sure what he was thinking there.

It’s easy for billionaires to leave.

It’s harder for them to bring along the land and labor that make them billionaires.

Go ahead and leave. The infrastructure that made you valuable will still be here.

Trusting your own judgement on ‘AI’ is a huge risk

Looking at the state of the world… you might be right.

Of course it is! We’re just animals, after all. Is documenting the behavior of different species of beetles a science? The only difference is that we can replicate behavior through culture, not just genes.

Chronologically, sure. But are years really the appropriate measure of accuracy here? Biological evolution moves a lot slower than cultural or technological evolution does.

https://www.preposterousuniverse.com/blog/2019/09/21/the-notorious-delayed-choice-quantum-eraser/

I think you’re inferring some sort of intense or serious/angry tone to my comments that I did not intend.

I just thought it was funny that this particular AI hype post promotes a way of coding based on vibes rather than understanding, and the way they came up with this very specific recommendation seems to be based on vibes rather than understanding.

(Nothing funnier than leaving two follow-ups explaining exactly what I found funny about it, I know. Also, FWIW, I haven’t downvoted you, in case you’re thinking this is supposed to be a flame war.)

Edit: Nevermind, I checked out your post history and realized I have been trolled. Well, gg.

I do need to chill the fuck out, but that’s beside the point.

The post’s thesis is “buy a Mac Mini and a Claude subscription because that’s the future”, and the supporting evidence includes:

Stack of Mac Minis on a desk. Said he’s using them for AI [whisper.cpp] workloads. … He just asked Claude how to set it up. Followed the instructions.

This example is probably the closest to “Mac Mini running AI plus using Claude”, but the Minis are a prod cluster, not a dev machine, so not exactly.

And the rest…

“AI won’t take your jobs. But guys with a $200 Claude Max subscription and an Apple Mac Mini will.”

This guy had been messing around with ChatGPT on the side. … Two days later, it’s done.

“If you’re not getting a Mac Mini for your @clawdbot in 2026, what are you doing?”

I understand the broader point of “get on AI ASAP”, but it seems like the extent of the Mac Mini recommendation is “I heard it was good for AI”.

Which… it is, if you’re running locally. But if you’re not, then there’s no reason to switch away from your used Thinkpad running Linux.

And if you are running locally, the Claude subscription may be irrelevant. So why advocate for both in conjunction?

Uhhhhhhh

Uhhhhhhhhhhhhhhh

Claude is not local AI though?!?!

Dude is like “Buy a Mac Mini specifically for running local models. And also pay $200/mo for hosted models” but shows zero understanding of this.

While saying “Don’t worry about understanding things”!

Is this satire?

Edit: Also he keeps saying 799, but that’s: 599 one-time purchase of hardware + 200/month subscription. Those are two different kinds of payments. Why would you add them?

I think it’s satire? I can’t fucking tell anymore. I’m tired, boss.

A little crazy, but not a lot crazy. ARM adoption may provide the spark necessary to ignite this fire.

Big agree.

But also: people seem to only focus on the output side of the task of writing code, and forget that the developer also receives input from the codebase in return.

Even if you end up with exactly the same code artifact after completing a work item, you’ll have a better understanding of the codebase without delegating swaths of it to AI. But bosses tend not to consider this.

Tech bros have successfully convinced people that mental states do not exist, or at least do not matter — for laborers, anyway, cuz they’ll happily claim that their superior thoughts are exactly why they deserve to be billionaires.

Code is a liability (not an asset)

Related: Jevons Paradox

As Cory Doctorow says: code is a liability, not an asset

I was once a fool like you :)

Mike McShaffry’s book “Game Coding Complete” is a good guide to the practical side of using a game engine IRL to get things done.

It’ll give you a good idea of how things should be shaped in order to be useful, and some things you can “skip ahead” to. Off-the-shelf engines have to be extremely general in order to be flexible enough to be useful to many customers, so game devs have to put in the effort to make them more specific. You’ll have to start off by being specific, if you have any chance of actually finishing something.

Eberly’s book “3D Game Engine Architecture” deals with the nuts and bolts, the rigorous academic engineering stuff. It’s pretty solid, but it’s aimed at making a general-purpose engine, which is beyond the scope of a one-person project.

Backing up though… You don’t have any language or library opinions? You might need 5-10 years of experience doing general programming (or game dev) before you can sustainably tackle this, or else you’re likely to paint yourself into a corner.

Edit: Probably the biggest PITA with game engine dev is testing. If you’re not already an expert in setting up test harnesses at multiple levels of detail, you’re gonna find it impossible to keep moving after a few months.

Good luck!

AI is destroying the human-to-human connections that produced the artifacts that made AI function in the first place.

Fuck AI.

Headline: X disables CSAM generator

look inside

X monetizes CSAM generator