HumanOstrich

I think you're missing the point of the MIT License and its history. If you want a proprietary license, "specializing" the MIT License for that is silly. My grandmother could specialize as a bicycle if I added wheels to her.

Yes, I'm a secret AI agent here to try and stop your powerful new licensing idea. /s

So what if Claude Code makes a mistake and tears up the sandbox? What happens to all the persisted state (aside from the container image)?

The linked fly.io article discusses why containers aren't a good fit for sandboxes that need persistent state and how sprites.dev addresses the challenges.

If you want to add restrictions to the MIT license, just don't use the MIT license or name it that way. Your change makes it no longer an open or permissive license so it's bizarre to keep the name. It would be like creating the MIT But-You-Have-To-Pay License.

How do local LLMs help?

Not really. One of the primary features of sprites.dev that I don't see anywhere on exe.dev is a fast way to create and restore checkpoints, like a git repo for your entire VM.

This is needed for sandboxes if you don't want to throw them away and start over when something goes wrong.

With sprites.dev you can create an additional checkpoint and then turn Claude Code (or your preferred agent) loose to do anything. Even if it burns down the sandbox you can just restore a checkpoint in about a second.

You can trace what's going back and forth over the wire between Claude Code and the model in use. That's going to be more insightful than their huge blob of JavaScript using React to render a terminal GUI.

I think the statistics definitively prove you wrong that wearing seatbelts doesn't matter. Same goes for there being non-zero risk with skipping all permission checks in Claude Code.

People aren't saying "I'm never using Claude Code again" just like they aren't saying "I'm never driving again". That's not what the post or most of the discussion is about.

Leave CC permission checks on or run in a sandbox. At a minimum, understand there are risks even if you haven't personally encountered them. See https://en.wikipedia.org/wiki/Faulty_generalization

They're still 90% of the way to their goal. And there's only 90% left to go.

This is like saying you've never worn a seatbelt and still haven't been in an accident. So you'd like to know the series of turns that led to someone else's accident.

Thanks for mentioning this. I always find it unsettling when I've researched solutions for something and only find a better option from a random HN comment.

Site: https://kubb.dev/

Yea that's a common theme of excuses for both Rust and Nix. Wrong though, because most anyone who can use a computer at all can learn the basics of Vim.

Seeing that flake.nix badge of complexity lets me know a project will be a nightmare to set up and will break every other week. It's usually right next to the Cargo.toml badge with 400 dependencies underneath.

Now you have to set up Nix first and deal with that nightmare of a learning curve. Just to auto-install some dependencies for a script.

Might as well rewrite all your scripts in Rust too while you're layering on unholy amounts of complexity.

The point is to not make a copy at all for the shared data of the same book. That's even what the page claims is happening (but it's wrong). That's the whole point of the flyweight pattern[1]. Instead, the example returns the same book instance for the same isbn in `createBook`, then blindly copies the shared data like title, author, and isbn every time in `addBook` to a new object.

> The shallow copy saves memory....

Versus not making a copy? A shallow copy of primitives still copies them and uses additional memory. Maybe there's some low-level optimization that makes it more efficient than that, but it's not relevant here. And there isn't any deep cloning happening in the example.

Might as well get rid of the complexity and instantiate a new Book class every time. And maybe stop shallow-copying class instances into new plain objects. It works for the example but has footguns.

Conveniently, there are other patterns on the site that would help you avoid creating an amalgamation of random merged class instances and objects.

The whole example is a mess.

[1]: https://en.wikipedia.org/wiki/Flyweight_pattern

The site is not "promoting" the singleton pattern. In fact, there is a "Tradeoffs"[1] section that calls it an anti-pattern in JavaScript.

In spite of that, there are plenty of reasonable use cases for singletons in many languages, including JavaScript. For example, ES Modules are effectively singletons. If you import the same module in multiple places, it only gets evaluated once.

Let's not turn the singleton pattern into forbidden knowledge.

[1]: https://www.patterns.dev/vanilla/singleton-pattern/#tradeoff...

I was glancing around and landed on the page for the flyweight pattern.[1]

It looks like `addBook` is using the spread operator, which always creates a shallow copy of the book instance properties, thus nullifying any benefits of the flyweight pattern. It also attaches extra arbitrary properties, but still assigns the result to a `book` variable. I don't think this is a great example.

[1]: https://www.patterns.dev/vanilla/flyweight-pattern/

Edit: I forgot to give you kudos for all the effort it must have taken to create all this content, and I appreciate that you're making it available for free.

"AI" isn't a verb, and most people don't care to cargo-cult their designs just to appease other members of the cult.

That's like saying Comcast doesn't have a monopoly because you can get crappy mobile Internet at 1Mbps speeds by tethering to your phone. Or saying DSL is "competition" at 12Mbps speeds. By your definition, there are no ISP monopolies left on Earth now that Starlink is here.

BTW according to the study I linked, less than half of Starlink customers get service that meets the anemic FCC definition of broadband.

> It's massively changed the market dynamics. And I suspect Elon will push the pricing down further and further.

This is hand-wavey conjecture, not an argument. He might lower the pricing if he gets enough welfare for his companies from the government. But only if that's a requirement, otherwise no he's not going to lower any pricing. The monthly pricing for Starlink has gone up, not down.

> There's literally an option for "auto", "light", and "dark" on each page.

That's not defining a color palette, which would probably go in CSS. You're describing a UI selector. It doesn't "define" anything related to implementing the functionality.

You're still missing the point of this entire thread. Go back and look at the top-level comment. I think there might be a language barrier here or something, so I'm going to bow out. Good luck.

Seems like a self-inflicted problem. You can turn off Light Switch in PowerToys. I don't know what screensaver software you're using, but if it's hijacking the setting maybe consider uninstalling it.

I don't think you know what you're talking about. That looks like the setting you would change specifically on the Wikipedia site. There would be additional code and CSS behind the scenes to enable that functionality. For example, where is the color palette defined?

The discussion is about site owners making light/dark mode configurable. Your first reply reads like you're saying that Wikipedia has an article for how to implement configurable light/dark mode on a site. I now think maybe you just meant "Here's how to change the setting on Wikipedia, it's so easy!" which still misses the point and is not easy for most people.

Maybe if you link to the Wikipedia page you're talking about that would help. I can only find variables like the above in documentation about templates and themes. That wouldn't be helpful anywhere outside of Wikipedia.

Insider trading by politicians is not really what's being discussed. Rich people and corporations lobbying politicians is different.

The residual value in the failed startup will be such a small fraction of the funding. You aren't making a convincing argument.

There are many light mode themes just like there are many dark mode themes. They even have colors.

While I agree that it's not as complicated as some make it out to be, let's not pretend it's just one variable change.

[Citation Needed]

I also use Dark Reader and read your articles often. Some sites look terrible with any Dark Reader settings, but yours looks fine.

I'm sure there are still plenty of people who would appreciate a site toggle for dark mode with a theme of your choosing.

Providing both is not a complicated thing, especially with support in CSS, browsers, and frameworks. That is, unless one designs websites that are so complex and buggy already that supporting a second color scheme costs "a lot of engineering time".

Starlink is not a substitute for fixed broadband. It might be adequate for most use cases, but it's not equivalent and it never will be. Your "99%" figure is nonsense. More than 1% of people will see the difference in download speeds between 1Gbps and 300Mbps. You realize some apps and games show you information about latency and bandwidth?

Also, just because they advertise "up to 305Mbps" doesn't mean everyone is getting that. A friend of mine with Starlink in the midwest gets about 100Mbps during off-hours. See <https://www.ookla.com/articles/starlink-us-performance-2025> - median speeds are typically less than 150Mbps.

Starlink also costs about 2x-4x as much per Mbps as 1Gbps service (at least where I am). I doubt they're going to suddenly offer 1Gbps speeds in the next year without changing their pricing. They'll add a new, more expensive plan.

Even 1Gbps Starlink (which would be more than 3x the current max speeds) is going to have other differences such as increased latency (they mention 30-40ms in one place, 25-100ms in another), more jitter, and lost signal sometimes during bad weather. Starlink also uses CGNAT which eliminates a bunch of use cases and introduces its own problems with certain apps and games.

They've had issues with capacity before where they wouldn't accept signups for some areas. Adding capacity involves launching more satellites.

Starlink isn't the ultimate solution to everyone's Internet access problems.

I could do without the irrelevant "AI agent" tip that implies it won't even work.

These days, any non-AI project that doesn't attempt to interject AI somehow is refreshing.