_b

If ordinary laws could redefine words in the constitution, then every constitutional right and protection you care about is undercut.

Washington's constitution says: "The word 'property' as used herein shall mean and include everything, whether tangible or intangible, subject to ownership". That is extremely broad, and there is 90 years of precedent affirming income is include here as something intangible and "subject to ownership."

Washington State’s constitution limits a tax like this to no more 1% and requires it to be uniform; this law meets neither requirement.

I am uncomfortable that supporters of the income tax are so unbothered by it being unconstitutional. So few are insisting we amend the constitution to allow or not do it at all, on the grounds that violating the constitution (or flexibility construing it to match our desired ends) is bad.

I think this is right, in part because I've been told exactly this from people who work for Google and their job is to sell me cloud stuff- i.e., they say they have so much internal demand they aren't pushing TPUs for external use. Hence external pricing and support just isn't that great right now. But presumably when capacity catches up they'll start pushing TPUs again.

> There are obvious conflicts of interest here.

There are almost always obvious conflicts of interest. In a normal startup, VCs have a legal responsibility to act in the interest of the common shares, but in practice, they overtly act in the interest of the preferred shares that their fund holds.

I once worked somewhere where I found some evidence that the offsets we were considering buying were probably junk. The CEO agreed, bought them anyway, and proudly proclaimed to all staff how we were now a net-zero company. Junk offsets are cheaper.

Making advanced LLMs and releasing them for free like this is wonderful for the world. It saves a huge number of folks (companies, universities & individuals) vast amount of money and engineering time. It will enable many teams to do research and make products that they otherwise wouldn't be able to.

It is interesting to ponder to what extent this is just a strategic move by Meta to make more money in the end, but whatever the answer to that, it doesn't change how much I appreciate them doing it.

When AWS launched, I was similarly appreciative, as it made a lot of work a lot easier and affordable. The fact AWS made Amazon money didn't lower my appreciation of them for making AWS exist.

We're adult enough to have discussions like this in public. They are healthy to have. People make mistakes. Kudos to the original authors for releasing the source code so people could inspect and replicate their results.

That is a good point. Are there modifications to the voting strike idea that would make it effective?

For instance, what do you think about an "inverted-voting" strike instead? Everyone on strike does things roughly as normal, but swapping most of their down votes for up votes, and visa-versa? That way, people on strike will continue to vote about the same amount as before, and in the same subeditors as before, so it would be harder to identify them. The hope would be that although Reddit could do work to clean up the voting data, it would be annoying & take time, and presumably still end up as significantly lower quality data than they had before.

I think a "down-vote everything" strike could be highly successful. As long as enough users participate, it could cripple subs whose moderators didn't choose to participate. And it is fitting. Reddit is trying to charge its users for content and data created for free by its users, so why shouldn't users make that data junk for awhile to make a statement?

DOCSIS 3.1 has been out for 10 years, and I'd be happy if Xfinity just offered me those speeds. Compared to max speeds of the standard, Xfinity offers up to 12% the download speed and 3.5% upload speed.

Loser pays causes higher legal fees, because on average, one only has to pay half of one's own fees, and also because losing becomes more costly making it worth it to spend more to try to avoid losing.

Hart says "Amazon served me with a lawsuit,"[0] but it was Amazon who was sued by California citing Hart as a witness to support the lawsuit. Hart is having to go through discovery as part of that.[1][2] Amazon is allowed to do this discovery, on the theory it should be able to scrutinize claims being made against it, particularly when billions of dollars are at stake.

Hart's estimate for the fees Amazon charges him seem approximately the same as the fees Amazon estimates it would charge for products like his.[3] Given that, perhaps he shouldn't matter much for the lawsuit. But the law is vague, often saying something is prohibited only when it is bad for competition. So vague claims of what helps vs hurts competition matters, which is why Hart's testimony matters.

Most stores choose not to feature items that one can readily buy cheaper elsewhere. So what exactly is the state's theory about when that is legal or not in such a way it is illegal for Amazon to be doing -- and is the state's theory here supported by the laws?

The problem really is the laws around what is anti-competitive are vague & subjective. As a result, everyone has to spend a ton of money on lawyers (and lobbyists) to figure out an outcome. Instead of this, the legislature should revise the law to be clear. I assume this is not done because the lawyers and lobbyists like the flow of money to them created by vague laws, but it also might be because vague laws are easier for the legislature to write.

[0] https://twitter.com/Molson_Hart/status/1663582603906088961

[1] https://oag.ca.gov/system/files/attachments/press-docs/2022-...

[2] https://twitter.com/Molson_Hart/status/1663601424226189315

[3] https://sellercentral.amazon.com/revcalpublic look up "B06XJCKS46"

Yup. AWS never raises prices, assuming one ignores that one time they added the per-transaction fee to S3 shortly after launch, but I think everyone thought was a rather understandable change.

Conventional wisdom is that AWS is more reliable to trust to build on top of. My interactions buying cloud services is that the AWS account managers are more trustworthy and reliable (although they spend less time building personal rapport and small talk). And this kind of thing just confirms that.

Cloud providers raising prices is unsettling. Teams invest many person-years into building software on top of cloud services assuming certain prices. There is a lot of lock in there. Although it is hard to predict to what extent they will pass along future cost savings, or if new cheaper options will come available, it has been safe to assume they won't jack up prices on any given workload.

By violating that trust, it will make teams wary of building on top of Google's higher-level services where there is the most lock-in, which unfortunately for Google, is also where the highest margins are.

People aren't interchangeable. The company might want to layoff 10%, but they don't want to layoff just any 10%. They will want to layoff people in positions and teams that they want to do away with, and to the extent some people are interchangeable, they would like to layoff those that are the least productive.

If you announce layoffs before telling people who will get cut, the people who tend to leave voluntarily are often those most able to get jobs elsewhere because they are high performers in positions that are hard to staff -- so exactly who the company wants to keep.

Having a 7% unregretted attribution target is pretty different than a 7% layoff. Many (most?) of the folks who make up that 7% leave on their own, because both the employee and the company are on the same page that it isn't working out.

A 7% layoff affects some people who are good employees, did nothing wrong, and are safe from a PIP -- but their position just no longer makes sense for the company to keep.

Companies do layoffs in a dehumanizing way largely to avoid lawsuits. If individual managers communicate it, some might say something wrong and invite a lawsuit. Even if they don't say anything wrong, some of the people being laid-off may misremember what they were told, and without a clear record of what they were told, they might win a lawsuit. In contrast, a mass email laying people off has the advantages that lawyers can review it beforehand, and it is clearly document exactly what was told to the employee in the course of laying them off.

The other part is they want to lay everyone off at the same time, so people aren't finding out at different times and wondering if they are safe or not. An email, sent to everyone at the same time, achieves that much better than trying to rapidly schedule a lot of simultaneous meetings.

Perhaps there is a better way, but that is the conventional wisdom that I've heard from folks planning layoffs.

How does that follow? Unionized companies also do layoffs.

I'd be interested to see "C compiled with Clang" added as another language to the benchmark games. In part, digging into Clang vs gcc benchmarks is always interesting, and in part, as Rust & Clang share the same LLVM backend, it would shed light on how much of the C vs Rust difference is from frontend language stuff vs backend code gen stuff.

Don't write constructors that can fail unless it is failure that would be appropriate to crash for. That works out a lot better than it might naively sound. It is hard for people to reason about the possibility of constructors/destructors failing, so actually rather nice to just forbid it.

I think the dilemma for C++ exceptions is that if the programmer actually thinks something should never happen, it is almost always better to just crash. But if the programmer thinks it might happen sometimes, then it is risky to predict it is will be super rare, as often this code will be called in different contexts in the future, so it is safer to use normal returns and flow control. As a result, throwing an exception is basically never the best thing to do.

20 years of taking pictures, but we'll be able to study the data it gathered, and benefit from discoveries it enabled, for an indefinite amount of time.

I checked gov.cn, and see it using "Mr. Xi" and "President". For example: https://www.fmprc.gov.cn/mfa_eng/wjdt_665385/zyjh_665391/t11...

But maybe I'm not really understand your point. Are you saying there is another English word that would better describe Xi than President, or that English should add one? Are you thinking "President" is insufficient to capture his power, and something like "Dictator" would be better? That has such negative consolidations that I would be hesitant to use it.

Most people in my life with a Phd choose to go by "Mr" as opposed to "Dr". All major news organizations I just checked, including Bloomberg, WSJ, and the BBC, call him "Mr. Xi" and not "Dr. Xi". Do you any evidence he chooses to go by "Dr" as opposed to "Mr"? Because otherwise, I am inclined to assume all the major news sources have it right.

You might be conflating justifying the ostensible purpose of a policy with justifying its likely effects, and perhaps justifying its means.

I think reducing inequality is great. But do you think Mr. Xi's initiative will meaningfully reduce inequality, or will it mostly be used to bolster Mr. Xi politically?

Do you have any worries about Mr. Xi keeping companies on a tight leash with the threat of regulatory hassles if they don't appease him? I think it is a mistake to view these coerced donations as independent from Mr. Xi curtailing the independence of tech companies. Mr. Xi views keeping party leadership over all forms of work in China as essential to the survival of his politically repressive regime (see a summary of "Xi_Jinping_Thought."). No doubt having the headline cause making the tech companies submit being reducing inequality is more politically palatable than, say, further crushing criticism of the government, although reduced corporate independence seems likely to lead to that too.

The unexpected $15.5 billion in spending on Xi Jinping's "common prosperity" initiatives was unsettling. It was only 10% of their net income for 2020, so not terrible as a one-time hit, but really drove home for me that the government might unexpectedly expropriate large amounts of shareholder value in the future. Baidu, Tencent, and others also had to pay up.

What? Everything in the STL works fine without exceptions. Only wrinkle is if one wants to handle out of memory in way besides crashing, although there is generally nothing better to do. Besides, the kernel will often kill processes rather than return malloc failures (see the OOM killer) anyway.

That is true only if making lotteries illegal made them nonexistent. In the United States, the legal lottery is arguably better than the widespread (illegal) "numbers game" it replaced.

By hashing data already in the processor's cache.