Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
w10-1
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
34 ms
·
511.
▲
by
w10-1
2y ago
Is this post from 2020 being revisited now because the situation seems even worse as Swift is building out isolation regions, complicating function coloring with stuff like per parameter inference?
512.
▲
by
w10-1
2y ago
I agree with the semantics, but restrict usage to the very narrow case where the variable has no other use than to convey a value from a visible outer scope to a visible inner scope, where both together are less than a few lines. In that c
513.
▲
by
w10-1
2y ago
Poison pills are much more effective at fending off takeovers. I think the public benefit form mainly helps with workforce and customer alignment to dampen disruptions from transformative and invasive applications. Peter Gassner converted
514.
▲
by
w10-1
2y ago
As a refresher... These exploits promise to be the rule, not the exception -- and not (just) because this company might have to comply with its national imperatives. Assuming companies get paid for deploying hackable devices, it gives them
515.
▲
by
w10-1
2y ago
I'm sorry the OP has to endure this. Grief is long. The story reminds me that busy people don't seek health care enough (not saying it would have changed anything in this case). We wait until something is so bad that we can'
516.
▲
by
w10-1
2y ago
Two models to address how/why/when AI companies make sense: (1) High integration (read: switching) costs: any deployment of real value is carefully tested and tuned for the use-case (support for product x, etc.). The use cases ty
517.
▲
by
w10-1
2y ago
--- There's actually a ton of money at the top of society looking for somewhere to go, which ideally would be routed to more worthy ventures than it is, but that's a whole topic in itself --- Forget "ideally": the experi
518.
▲
by
w10-1
2y ago
Where does one get part-level information to qualify a supply chain as "juicy"? Otherwise the first advice looks like a 1st-world solution that independent 3rd-world developers can't deploy. Could early-stage companies even
519.
▲
by
w10-1
2y ago
--- (2) “Critical harm” does not include any of the following: (A) Harms caused or materially enabled by information that a covered model or covered model derivative outputs if the information is otherwise reasonably publicly accessible by
520.
▲
by
w10-1
2y ago
The disparity between size of the promise and the ambiguity of the business model creates both necessity and advantage for executives to leverage external forces to shape company direction. Everyone in the C-suite would be seeking a footho
521.
▲
by
w10-1
2y ago
I agree with the business model of focusing on vertical integration with specific partners instead of DTC. There will be inevitable product quality trade-offs, but if you can select the partner context where those trade-offs work you can m
522.
▲
by
w10-1
2y ago
> It needs a layer of quasi-stable well documented subsystem APIs I think the Rust developers weren't even asking for that. They just want the C developers to sign up to some semantics. But the C developers know the function int
523.
▲
by
w10-1
2y ago
Paper title: Biomolecular condensates regulate cellular electrochemical equilibria Summary of abstract: protein aggregates form electronegative buffer, with corresponding field extending to cell wall. Tested in e-coli: induced aggregates g
524.
▲
by
w10-1
2y ago
It's not new, but it's more usable, which makes new transactions and productions possible by lowering information costs. That lower of costs is the ONLY basis for thinking AI is good for all. It's to the detriment of people
525.
▲
by
w10-1
2y ago
> to put AI into the hands of as many people as possible, we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips) Not really. The problem is that learning requires scale, mostly of data.
526.
▲
by
w10-1
2y ago
> a defining characteristic of the Intelligence Age will be massive prosperity That's the sales pitch, that this will benefit all. I'm very pro-AI, but here's the only prediction for the future I would ever make: AI will a
527.
▲
by
w10-1
2y ago
> If I can in any way purchase something without the web, I'mma do that To get to the milk you'll have to walk by 3 rows of chips and soda.
528.
▲
by
w10-1
2y ago
OK: This has not been a detailed step by step walkthrough on how to do things, by design. You are meant to go and explore; this is simply a way pointer to invigorate your curiosities Sorry, but because I came looking for
529.
▲
by
w10-1
2y ago
Unclear if this has been in the works longer as the graalvm LLVM build of python discussed yesterday[1]. The first HN discussion is from 2022 [3]. Any relation? Any comparisons? Funny I can't find the license for graalvm python in th
530.
▲
by
w10-1
2y ago
> the compiler is asking me to help it out Yeah, this sucks, but it's also mind-boggling. The SwiftUI body result builder returns "some View" which means the compiler has to infer the specific type that the function produ
531.
▲
by
w10-1
2y ago
I think the complaint is the global runtime namespace, not source modules. Statics live forever, and extensions in any library on a type apply to all uses in the runtime (with no guarantees about conflicts). Mostly that's minimizable
532.
▲
by
w10-1
2y ago
It's super nice that they support moving up incrementally. Moving to Swift-6 mode with full data-race safety checks can be daunting. They wrote a separate post on that, and Holly telegraphed that they're still actively reducing n
533.
▲
by
w10-1
2y ago
> story for web development Under the hood, Swift-NIO and async Swift is a pretty powerful basis for writing performant servers. Aside from Vapor, there are other small/fast containers like hummingbird. Not mentioned (surprisingly)
534.
▲
by
w10-1
2y ago
Swift throws are results, but now with specific and generic and specializable types. To illustrate, Swift has a nice `rethrows` feature that helps with function composition. If your function takes a function parameter that can throw, it can
535.
▲
by
w10-1
2y ago
It's a fair point, but radically different 8 years on. Swift type inference is two-way, so you can often improve things by annotating types instead of using inference. Also, it very easy to have complex builders and nested closures tha
536.
▲
by
w10-1
2y ago
I haven't seen embedding using graal/vm, or inter-op using the native JVM FFI. There is (active, 2K stars) https://github.com/pvieito/PythonKit and I've heard of people being able to deploy apps with pyt
537.
▲
by
w10-1
2y ago
Yes, but I was under the impression that graal-level inter-op was limited to packages the graal toolchain could compile. Thus, while swift and graal both depend on llvm, they use different variants and there's no real way to make inter
538.
▲
by
w10-1
2y ago
It's not just that Tim Cook wants to help humanity, or more accurately avoids the false positive products. Nor should Apple aim to deliver "I'm so special!" feeling of a new product category. The key is: Tim Cook underst
539.
▲
by
w10-1
2y ago
Their content-addressing hash would seem critical, but the "combineUnordered" hash they use just adds each byte. The API is clear that this is only as good as possible, and I'm not sure I would rely on it for data used for
540.
▲
by
w10-1
2y ago
I hear the complaint but it's a bit of a trap to just seek protection. Yes, the law on point is permissive. That goes with the evolution of law. But assuming for the moment that we want not just avoid injury to ourselves but to create
More ›