Yeah. I do wish there was something that was like Clojure with a TypeScript or Go-like nominal typing, but I do feel myself missing types a lot less with Clojure compared to other languages.
Type annotations mix poorly with s-expressions imo. Try an ML, which answers the same question of "How do we represent the lambda calculus as a programming language?"
There's already type annotations in Clojure and they look fine (though are a bit noisy). There are type algorithms that don't need annotations to provide strong static guarantees anyways, which is the important part (though I'm not sure you can do that with nominal types?). I think TypeScript and Go's syntaxes are a bad fit for s-expr but the idea probably isn't.
I doubt that all of the providers hosting open models on open platforms are losing money on serving inference. They have the benefit of not having to pay for training, but the models are open and aren't going away anytime soon.
Sadly, I have not been able to find any open model that comes close to Opus 4.6. So while they are much cheaper to deploy, they also aren't good enough for unsupervised agent execution. But you need a model that can run unsupervised for the claim "Code is Cheap" to become possible.
I don't really think so. Maybe it's because the systems I build need to be reliable and understandable to humans, but I don't think Opus 4.6 is good enough to be unsupervised. I've spent a lot of time using it and I have to tell it no semi frequently and rewrite by hand/iterate with the model frequently. It's saves me a lot of time when used this way, I think, but I still have to give it overall structure and keep it scoped to small changes to prevent it from going down wrong paths and generating tons of unnecessary code (which is how you end up with unmaintainable slop). Less code is pretty much always better and these models made it really easy to ignore that until it's too late. This is on a healthy mix of greenfield and brownfield projects.
To that end, I've actually found Kimi K2.5 to be "good enough" for a lot of that, not quite as good at Opus 4.6, but good enough that it gets me like 80% of the value for a fraction of the cost and more speed.
It looks like it depends on how much you use ACLs. The new ACL cap seems strictly worse than before. Otherwise it seems the new free plan is mostly the same as the old Personal Plus plan for most users.
It's honesty sad that it took like 20 years of bullshit to get to the point where we can freely discuss this stuff without being shouted out of the room for being sexist or some other -ist/-ism. The pursuit of equity is one of the single most toxic things to happen to modern societies.
1. I don’t see how that’s better in any real way. You can infer the exact same information as querying the range and it makes dynamic behavior based on age range (ex. access to age restricted chat rooms as an obvious example) completely impossible.
2. Is it meaningfully more identifying than User-Agent? There’s dozens of other datapoints for uniquely identifying a user. If we get a few high profile lawsuits because advertising companies knowingly showed harmful ads to children, I’d consider it a win. Age is not that interesting of a data point.
I wouldn’t focus on whether it’s “identifying” but whether it’s revealing. Young teenagers are a very high-value target for advertisers. They are very impressionable, and they provide a proxy for advertisers for their parents’ money. So this law essentially makes it mandatory to share that information with advertisers. And also by proxy, predators.
It also makes it explicitly illegal to do use it for such purposes. While I agree on the point, I think in practice it changes little. I also think it could be a net positive, because now there’s no plausible deniability about the targets age, opening up a decent amount of liability for exploitative practices targeting children specifically.
It's so much better. In the one case, the OS is leaking age information (even if just an age range) to every service it talks to. In the other case, the OS isn't telling anyone anything, and is just responding to the age rating that the app/service advertises.
How would you implement a feed of mixed content? Say you're YouTube and some videos are about puppies and some videos are about guns? How would you hide only the gun videos from the homepage when the user is under 16?
I'm not even talking about entire sections that feature blatantly pornographic or perverted content, some of which are clearly aimed at a younger audience who might accidentally stumble upon it through keywords you wouldn't expect.
1. Depends on how it's implemented. It won't identify you to individual platforms if the OS filters on a per-app or per-website basis. And yeah, there would be no dynamic behavior based on age, as that would enable tracking based on age. I don't think any kind of API is the ideal solution though, it's just better than the malicious one being mandated in the Cali bill. Instead of an API, it's simpler and more effective to just have an app installation lock (like sudo on Linux) and a firewall for website blocking with a nice UI in the phone's settings, locked behind a password/pin.
2. Other data points like User-Agent are not required by law, and browsers already spoof user agent by default. I agree that there are other data points we need to address, but the problem in this specific case is the slippery slope of legally-mandated data points. And I don't think winning high profile lawsuits is a real "win", it just exposes problem which we already know in this case. Keep in mind those people can get away with the Epstein files.
Just because you have never hit the issue doesn't mean it doesn't exist. This particular issue will only really show up on notched devices with a small screen and a lot of status bar icons. It's highly dependent on what model mac you run on.
reply