Security guard is one. Had a friend in college that basically got paid 8hrs/night to do 2hrs of actual work and 6hrs of building his portfolio. It can definitely work well for some folks.

Got Megabonk working on my Retroid, and can’t stop playing it. I thought I would try getting some other games going, but I just play Megabonk instead.

Opus Magnum

That game scratches my brain in such a satisfying way

Sure, but do you need a discrete video card if you’re gaming on an ARM SoC? And we’ve seen from the struggles of x86 iGPUs that graphics APIs pretty much have to choose whether they’re going to optimize for dedicated VRAM or shared memory, cuz it has inescapable implications for how you structure a game engine. ARM APIs will probably continue optimizing for shared memory, so PCI-E GPUs will always be second-class citizens.

Nvidia does not care about the ISA of the CPU at all.

That’s kinda my point. They’re stuck communicating over PCI-E instead of being a first-class co-processor over AMBA.

Fex

1. Nvidia abandons x86 desktop gamers
2. The only hardware that gamers own are ARM handhelds
3. Some gamers stream x86 games, but devs start selling ARM builds since the x86 market is shrinking
4. AI bubble pops
5. Nvidia tries to regain x86 desktop gamers
6. Gamers are almost entirely on ARM
7. Nvidia pulls an IBM and vanishes quietly into enterprise services and not much else

Mostly agree.

But I think their advice falls prey to the “only a Sith deals in absolutes” problem, when they start contrasting “concrete advice” vs “generic advice”. They are offering “generic advice” with this post, aren’t they?

They hedge against that hypocrisy by offering some special carve-outs where generic advice is still “allowable”, but Idk. I think this post could’ve stuck to the 60% of the topic that was a slam-dunk instead of trying to take on the entire topic of design principles.

After all, I think you could argue that when experienced designers appear to contradict design principles, it’s because they understand the underlying logic of the principles and are recontextualizing them for this specific problem. That argument prioritizes concreteness but also doesn’t paint design principles as unimportant.

As Picasso or someone once said: first you must learn the rules, and then you must break them.

Make computers do stuff for what purpose?

I joke to my family that I just name things for a living. When you take away all the incidental stuff like files and pointers and ports, that’s really all it is. “This sequence of events with these properties is called <this>, and when you ask our system what to do about it, it does this other sequence of events with these properties which we call <this other name>.”

It’s kinda like those ancient stone tablets that are the first example of writing, and they’re just like “Ramses owes Jeremiah 5 chickens” or whatever. It’s just how we manage abstract concepts moving around our civilization. Yeah there’s math involved, but every endpoint is a human being in one way or another.

Yeah…

Human mistakes tend to 1) look like mistakes, and 2) are surrounded by lots of hints that the author had trouble with that section of code.

AI mistakes tend to 1) look like regular code, and 2) look just as confident and effort-ful as the rest of the code.

Tech bosses have well and truly lost it.

Consider that:

• Code is primarily to communicate from human-to-human, and only incidentally for computers to execute
• A codebase that is 30+ years old has an absolute shitload of learnings incorporated into it, much of it very subtle
• Languages are, in fact, different. So some things cannot be directly translated with exactly the same semantics, so devs will need to fully understand the intent and resolve ambiguities
• A million lines per month is a lot of text for someone to successfully interpret and translate without losing any subtleties

Why?

The language itself is… fine. It has some bad decisions baked into it, but what language doesn’t? And it has a pretty mature security model, which is a big help for embedded devices.

And given that so many embedded devices these days are talking to cloud services — that are probably running JS, but are at least communicating via JSON — being able to share some cross-platform code or tooling can help things go quicker and/or safer.

Edit: This is probably not a route for teams that are chronically tight on memory and choose to solve that by spending hundreds of dev-hours on optimizing code rather than adding BOM cost. But for teams that could stand to increase BOM cost as long as the savings on dev-hours make up for it… it could be an option.

They’ve been telegraphing this for a while — long before the RAM crisis

Why would you expect that?

Molly White’s coverage:

…maintained that they were merely developing privacy-preserving software, and that they were not responsible for criminal use of the software. Prosecutors have argued that the developers actively intended the software to be used for criminal purposes, pointing to marketing aimed at “Dark/Grey Market participants” and those engaged in “Illicit activity”.

Judge Cote cited a letter to the court in which Rodriguez continued to say that he was merely motivated by a desire to protect financial privacy and not “a desire to facilitate criminal activity” as evidence that Rodriguez “has not come to terms with what he did. … The letter indicated to me that you were very much still operating in a world with moral blinders on.”

https://www.citationneeded.news/issue-96/#samourai-wallet

iPhone’s design is more secure than Android (partly because of OS+hardware integration that just isn’t practical in a multi-vendor space), but they still have plenty of zero-days in their implementation. iPhone 7 is old enough that official security patch support is EOL, though Apple has still shipped some critical fixes past EOL.

if you want to, say, stop LLM server if available mem is under 8GB and start it again when it’s over whatever LLM needs (in my case it’s 64GB):

So this is the guy that bought all the RAM.

Yeah, we need to be careful about distinguishing policy objectives from policy language.

“Hold megacorps responsible for harmful algorithms” is a good policy objective.

How we hold them responsible is an open question. Legal recourse is just one option. And it’s an option that risks collateral damage.

But why are they able to profit from harmful products in the first place? Lack of meaningful competition.

It really all comes back to the enshittification thesis. Unless we force these firms to open themselves up to competition, they have no reason to stop abusing their customers.

“We’ll get sued” gives them a reason. “They’ll switch to a competitor’s service” also gives them a reason, and one they’re more likely to respect — if they see it as a real possibility.

Eh, it’s fine. It has some bad choices baked into it, but what language doesn’t? And JS in 2025 is miles better than JS in 2005.

I wouldn’t choose it for every project, but it’s a reasonable choice in many cases.

• loads more slowly on HDD
• now small enough for your SSD