It's still considered an experimental feature, just as it was in 5.38.x. Also the linked documents aren't 100% reflective of what's been implemented so far, a better URL would probably be https://perldoc.pl/perlclass.
I've been surprised. I have a coworker who has a deep industry knowledge but none of it was in Perl, he was reaching for ChatGPT a lot early on and while sometimes it wasn't how I would have written it it wasn't wrong. So I did some experiments and I asked Claude to do things like generate apps using the latest Class syntax which is only 3 years old at this point … it didn't have a problem.
Edit: well it occasionally had a problem where it would confuse things for Moose or Raku … but by and large it wasn't wrong it's just the syntax is new. I know Perl developers who have the same issues.
I'm not going to dispute the data, but I don't think it means what you're saying it does. Honestly, I think you've made the same mistake that most people make in this conversation and equating NPR to Public Radio. NPR is Public Radio, but not all Public Radio is NPR.
TL;DR: Based on the document you linked to the lions share of the money goes to _public radio stations_ in unrestricted grants to pay for their operations. Of the money that does trickle through that to program providers, NPR is only one of dozens of program producers and distributors, and NPR's own reporting on this seems to accurately portray the reality of how they are funded from independent stations.
Looking at the second link you provided theses are the expenditure programs that look related to public radio broadcasting in the us: Radio Program Fund, National program production and acquisition grants, Community service grants, and Public broadcasting interconnection system. There may be expenses in other categories but they don't appear to be directly related to public radio.
According to that document, these programs are:
Radio Program Fund represents expenses for the development and production of high-quality, new and innovative radio programs that might not otherwise be supported by the marketplace.
National program production and acquisition grants are restricted grants made to qualified public radio stations that must be used for the production, acquisition, promotion or distribution of national radio programs that are of high quality, creative and reflect society’s diversity.
Community service grants are unrestricted general operating grants made to qualified public television and radio stations.
Public broadcasting interconnection system [...] a new interconnection system to be used by both public television and radio stations to transmit and receive programming feeds [... as well as] “other technologies and services that create infrastructure and efficiencies within the public media system.
The line item breakdowns on these services for 2022 are
Radio Program fund: 7,145,925
National program production and acquisition grants: 24,579,081
Radio community service grants: 72,446,561
Public broadcasting interconnection: 2,871,313
Total: 107,042,880
So it looks like the CPB spends slightly over 107M on _public radio_ in the united states.
Note that the "National program production and acquisition grants" specifically go to qualified public radio stations. NPR isn't a public radio station, and so doesn't directly get any of that money. That money would go to a local public radio station (here's a [list](https://radiostationusa.fm/formats/public)). A station like my own WMFE then would use that grant money to purchase a radio program like All Things Considered from NPR. But they might also use that money to purchase a radio program like Marketplace from APM, or This American Life from PRX, or be used to produce it's own programs like WBUR's Here and Now. So that money _might_ go back to NPR, but it might not.
The same thing applies to the Radio community service grants, they're unrestricted grants to _stations_ so NPR _might_ get some of that money but much of it probably go to pay local expenses, like the salaries of the radio engineers who keep the towers up and running.
What about the Radio Program fund? That looks like it's pretty much targeted to NPR. Certainly not 100% of it is going to NPR's new programming. Even being generous all de-funding this will do is specifically cut programs "that might not otherwise be supported by the marketplace". I suspect most of that money goes to things like The Splendid Table, and Growing Bolder (neither are NPR shows btw). Unless you want to argue with Rupert Murdoch and Ted Turner that news isn't a marketable product?
The Public broadcasting interconnection is paid to _stations_ to make physically acquiring content easier and a deal with NPR, APM, PRI, PRX and others reduced that cost _significantly_, so that line item dropped to 2M in 2022 from 78M in 2021.
Nothing in your document I believe NPR _claims_ they get about 13% of that portion of a station's budget (based on [WMFE's 2021 Audit](https://drive.google.com/file/d/121U2UFSI8eNYUg59TnZtELEhBL-...) "program acquisition" was about 16% of their total budget so NPR seems credible here) … so about ~$12M of the CPB funding, add in a chunk of the Radio Program fund let's say 80% or ~$6M … we're still at about $18M of the $107M … plus another 1% direct funding (or is that 1% the Radio Program fund?) and we're around 20% of the total CPB "public radio" budget ends up in NPR. Since the CPB's "public radio" budget is only about 25% of the total expenses from the CPB … is 5% of the CPBs expenses considered "significant funding"?
De-funding "public radio" won't mean that Morning Edition goes off the air, it means that the farm report for Ida Grove, IA goes off the air … and the wildfire reports for Billings MT … and close to home for me it means that Brendan Byrne won't be able to keep doing Are We There Yet and reporting on the rockets that keep going up funding my local community with something _other_ than tourism.
The numbers aren't wrong, your definition of "public radio" is wrong when you assume it's just NPR.
I actually find Go harder to work with JSON with for exactly those same reasons. With Go if I want to parse JSON into something I can call methods on, the JSON has to conform to something I can build a struct around. If the JSON is complicated in anyway that doesn't line up with Go's sense of structs I'm forced to drop to map[string]interface{} (or something like it). With Perl I can parse into the hashref equivalent of map[string]interface{} and still be able to bless it and call methods on it.
This isn't to say that Perl is better or Go is better, they just have different affordances. Perl is designed for loosely structured data and that causes all kinds of problems when you want to enforce highly structured data. Go is designed to handle strictly structured data and the ramp down is hard when you don't have that.
