I have to use more than one Python type checker because there is not a single one that works. Not only different tools catch different issues. They also have different bugs, and different configuration requirements. Different teams have different preferences.
It's a nightmare. If Python taught me something about typing is that a language that doesn't have a clear definition of types in the reference implementation, it will never get it fixed with external tooling.
The worst part for me are the false positives. I frequently need to get into the spam folder to discover emails that Gmail thinks they are spam, even though there is absolutely no reason for it. I have been thinking about leaving it.
> The total addressable market — every emerging-market saver who would prefer dollars to their local currency, every cross-border worker sending remittances home, every small-business operator outside the conventional banking perimeter — is, by any honest estimate, several trillion.
Emergent market savers are already buying dollars. Nobody in Argentina saves money in pesos, for instance, so there is no much demand there left (and Argentina is the country that holds the largest amount of physical dollars after the US).
It is a fantasy to suggest that there is a magic, hidden demand of dollars that can be measured in trillions (coming from emergent countries, of all the places). At least, not one that would not be controlled by local authorities, just like they do today. Governments can control who tries to buy some random crypto with the local currency. Not when it's done in small amounts, but anything sizable will be noticed. The idea that crypto or stablecoins can somehow enable capital flight measured in trillions without governments being able to do anything does not make much sense.
I could imagine it being promoted as a tool to launder money in some way, though. Just buy this stablecoin via some tax haven, we will make it equivalent to holding US debt, and we won't ask where the money comes from. I would bet that's where the tether/usdt money comes from. Drugs, arms trafficking, perhaps that could provide some of the liquidity that the US Treasury needs.
All this Mythos talk reminds me of fuzzing. The number of vulnerabilities discovered greatly increased. But bugs were patched (often, without even assigning CVEs), systems were upgraded, life continued. Outside of the tech world (and even within it), nobody noticed. Many software companies still have not adopted fuzzing.
If Claude finds security issues, so what. Some systems might not be updated, but these systems are vulnerable to even a single security issue, they were unsafe already. Systems that are upgraded don't have that problem. In fact a high rate of security fixes will make them safer, as zero-days kept in secret by government security agencies for long time become patched.
With fuzzing we didn't see articles talking about the geopolitics of security vulnerabilities. Investors didn't finance fuzzing startups with hundreds of billions. There is way too much propaganda.
Well, that's exactly what the article is about. Wayland put all together into one process I order to avoid unnecessary context switch. This protocol aims to keep the performance advantages of Wayland without giving up on separation of graphics c server and window manager.
Foreign operators are mandated by the EU, they can't be banned. Spain has been one of the first countries to allow foreign high speed operators (unlike other European countries that did attempt to delay their entrance as much as possible
Have you read the article? He didn't just ignore her. He combined periods where he ignored her with periods of caring only to hurt her in dark ways.
> Once, she says, as Jobs groped his wife and pretended to be having sex with her, he demanded that Brennan-Jobs stay in the room, calling it a "family moment." He repeatedly withheld money from her, told her that she would get "nothing" from his wealth — and even refused to install heat in her bedroom.
I'm surprised that governments didn't take this problem more seriously. Obesity is a huge problem, people have been ignoring it only because improvements in medicine have been offsetting the general health decline. Without the medical improvements that save the life of obese people, life expectancy would have decreased. I don't expect the Trump administration to make the best decisions but at least they are taking it somewhat more seriosly.
There are certainly other use cases. git bisect was enormously useful when it was introduced in order to find Linux kernel regressions. In these cases you might not even be able to have tests (eg. a driver needs to be tested against real hardware - hardware that the developer that introduced the bug could not have), and as an user you don't have a clue about the code. Before git bisect, you had to report the bug and hope that some dev would help you via email, perhaps by providing some patch with print debug statements to gather information. With git bisect, all of sudden a normal user was able to bisect the kernel by himself and point to the concrete commit (and dev) that broke things. That, plus a fine-grained commit history, entirely changed how to find and fix bugs.
As someone who uses D and has been doing things like what you see in the post for a long time, I wonder why other languages would put attention to these tricks and steal them when they have been completely ignoring them forever when done in D. Perhaps Zig will make these features more popular, but I'm skeptic.
This makes me sad. AI is a product, these days being mentally healthy should imply having the emotional ability to be aware of that. If you become emotionally attached to a commercial product you should seek for help.
I would say that the lesson here is that cross-vendor replication is more important than intra-vendor replication. It is clear that technology can (largely) avoid data losses, but there will always be humans at charge
Yep, for me it confirms all the reasons why I think python is slow and not a good language for anything that goes beyond a script. I work with it everyday, and I have learned that I can't even trust tooling such as mypy because it's full of corner cases - turns out that not having a clear type design in a language is not something that can be fundamentally fixed by external tools. Tests are the only thing that can make me trust code written in this language
The problem is that even when you give them context, they just hallucinate at another level. I have tried that example of asking about events in my area, they are absolutely awful at it.
I have to use more than one Python type checker because there is not a single one that works. Not only different tools catch different issues. They also have different bugs, and different configuration requirements. Different teams have different preferences.
It's a nightmare. If Python taught me something about typing is that a language that doesn't have a clear definition of types in the reference implementation, it will never get it fixed with external tooling.