My hobby projects have 100x more tests than they used to, because LLMs are great at writing tests. And my subjective experience is that the net quality has increased as a result.
YMMV, but it’s certainly a common belief, and for me at least a lived experience.
Did you just move the goalposts from “you can’t run arbitrary code today” to “hypothetically, in the future, Apple could prevent running arbitrary code”?
As with Google accounts, it's not hypothetical, it's a risk. People do occasionally get locked out of being an Apple developer for reasons they cannot foresee.
I could have a stroke that leaves me unable to program. Does that mean I am not truly free to program today?
Those are risks, but they do not change the on-the-ground reality today, and the claim was that users, today, cannot use these device as general purpose computers.
We can use them today as general purpose computers, if we make a large effort to do so.
In my Linux and Mac, I dont think twice to quickly write a script to automate some pain-in-the-butt issues. But with my phone, it is pain-in-the-butt to write anything. It becomes not worth the effort.
Moreover, we can argue if technically it is a general purpose computer for whole day long. But that's not the point.
The point is that we are allowing gradually the big organizations to restrict general purpose computing, the internet and other previously free systems. It is happening slowly, where we can still give them the benefit of doubt. We are the frogs in the kettle where we are arguing that the temperature is just one degree more than earlier, so it is not actually boiling. We can keep on arguing about the temperature or step back and see the big picture where it is going.
No it's not. I need permission by a third-party to be able to program a device I supposedly own. I need to give them money, I need to give them my identity, and I need to tie my identity to any distribution of the software I make if other people are to be able to install it.
This is not a rhetorical sleight of hand, this is just saying that I am not truly in control of the device that I have bought.
IDK, not really a fan of redefining computer to make a rhetorical point.
It seems counter-productive to tell people the computing device they think as a computer isn’t really a computer. It’s like saying my car isn’t really a car because I can’t adjust spark timing. Someone could make that semantic argument but it’s hard to imagine anyone would care.
I don’t see the value in hypotheticals like that. If the claim is that a computer is not really a computer unless every user can do any low level operations they want, is it also true that a car is not really a car unless every user can do any low level operations they want?
Manufacturers are taking away right to repair too! I think you picked a bad example. Back in the 60s you absolutely could change every low level component on a car.
Point taken. But I think we can say that smartphones and tablets are definitely not "general-purpose computers" because they are not programmable, at least not freely so.
Is it still a "car" if it only takes you to the train station and forces you to use public transport? "shuttle" might be more appropriate? Is it still your car if it leaves during the day and carries other people?
The meaning of words drifts when the situation changes.
Yes. Roads are subsidized; the true cost of building and maintaining roads comes from general funds, not just from vehicle registrations and gas taxes (which of course Waymo doesn’t pay, being righteously electric).
So you pay Waymo, they pay a few hundred dollars a year per car in registration, and you benefit from billions of dollars a year in highway funds from both state and federal sources.
Good point about electric. Maybe a tax on tires would be more fair, but that would lead to some dangerous behavior.
Waymo and I pay a lot in state and federal taxes. Shouldn't that work out that we're paying for a shared resource we use even if the proportional accounting is not exact?
But in any event, cars and roads are massively subsidized, such that drivers get far more than they pay for. See for instance: https://www.cbo.gov/publication/59667
Roads are paid for out of the general fund, meaning that even those (few) who don’t use them pay for them, which I’d call a subsidy (as opposed to self-supporting). That’s not necessarily a bad thing; the same is true for many programs that support low-income people, and I think that’s great. But it’s still fair to call it a subsidy.
It’s an agreement between a public company and a highly scrutinized private company. Several of the provisions will change what happens in the marketplace, which everyone will see.
I imagine the thinking was that it’s better to just post it clearly than to have rumors and leaks and speculations that could hurt both companies (“should I risk using GCP for OpenAI models when it’s obviously against the MS / OpenAI agreement?”).
And they aren’t being objective and rational about the polls, they are funding and cherry-picking poll data that tells them to do what they want to do.
There’s no principle, no strategy, no goal. We’re living in the political version of Cube, and just like the movie: it’s a headless blunder operating under the illusion of a master plan.
The only polls they care about is their sales numbers and their sponsor dollars.
It really doesn't matter how popular or unpopular a candidate is, what matters is if their listeners are still willing to overpay on snake oil. Or if their oil barrons are still giving them a few million dollars for whatever message they want to sell.
AJ is probably the worst in this space. One of the things leaked in his emails is if you give him $20k he'll gladly bring you on the show and talk about whatever it is you want to talk about and sell. You could probably get him to shill for a book about the benefits of communism.
You don’t see any issue with insecure drivers for obsolete hardware, exactly the kind of thing that is most prevalent in an industrial control type applications?
Stuxnet should have been a wakeup call to everyone: the boring, obsolete, “safe because nobody browses TikTok on it” hardware is exactly the highest risk.
My hobby projects have 100x more tests than they used to, because LLMs are great at writing tests. And my subjective experience is that the net quality has increased as a result.
YMMV, but it’s certainly a common belief, and for me at least a lived experience.
reply