bleepblorp

Have enough details about TSMC contracts ever been released/leaked to know they don't do this?

I don't follow the semiconductor industry closely enough to know anything about TSMC's business practices, but these kind of contracts are far from unheard of in other sectors.

Pandemic-induced growth in bike use for commuting is mostly at the expense of public transit ridership. People aren't trading their cars for bikes, they're trading the risk of dying from a transit-related COVID-19 superspreader event for the risk of dying in a vehicle vs. bike crash.

A mode shift towards bikes won't happen in most of the US because most of the US doesn't have viable public transit, and therefore, there's no need for Americans to switch transport modes.

I believe bike use has increased in relation to public transit ridership in greater New York, but this is absolutely the exception to the rule as NYC was (prior to COVID-19) a transit-dependent city.

GPU drivers are essentially operating systems for the GPU.

The problem with monitors is much more with the display industry's preference for finding new ways to scam customers instead of commercializing display technology improvements.

The history of LCD monitors and TVs is a history of the industry increasing prices while finding new corners to cut by removing functionality that reviewers and users didn't know they needed to test for.

When users first realized that IPS was vastly superior to TN for most purposes, the industry responded by making low-quality 6-bit IPS panels, which the industry happily described as 'IPS' without disclosing that the panels can't display color gradients--and often flickered badly due to FRC. Early 6-bit panels were worse than TN in everyday use.

6-bit IPS is now essentially unavoidable in smaller, non-4K, screens.

Another fun scam the display industry concocted was releasing one production run of good quality 8-bit IPS monitors, to get positive professional reviews and user word of mouth, then making subsequent production runs under the same model number with 6-bit fraud-IPS, PVA, or even TN panels.

The low-frequency PWM backlighting scam (which probably saves all of ten cents per display) is now known well enough that reviewers will test for it, but the industry has developed other ways of preventing users from purchasing good-quality hardware at a fair price.

Dynamic contrast is a particularly insidious problem. This piece of joy from the display industry changes the backlight intensity and RGB pixel values depending on screen contents. As a result, the color tint of the display changes depending on screen contents. These constant color shifts make dynamic contrast monitors unusable for even non-professional design/photo/video work and prevents color calibration entirely. This technology is even applied, with no user option to disable it, to monitors sold as sRGB pre-calibrated and marketed towards mid-range color work. Officially, dynamic contrast is intended to save energy, but the actual intent is likely to force people who want stable colors to buy professional displays at ten times the price.

A new scam that's emerged since the start of the pandemic is color gamut restriction; it's now very difficult to get displays--especially laptop displays--that support more than 50% of the sRGB gamut. Displaying actual red is now something the industry has decided to exclusively paywall into professional panels targeted at the design market.

The problem with displays isn't technology; it's an industry that's built on the premise of ripping off consumers.

People who are too poor to buy things won't buy any fewer things just because they get free things. They'll just have more things that they can resell for more money, which they will likely use to buy things they actually need.

Giving money--even indirectly--to poor people makes economic sense because they'll spend it on goods and services that keep other people employed.

Punishing the poor out of some moral imperative is economically destructive.

I don't see why this is a problem for reusing precompiled headers in the way the parent post suggested. If the compiler records the value of every #define when it precompiles b.h, if b.h is #included again with the same #defines set, the compiler can reuse the same precompiled version of b.h.

Indeed, if the compiler-buildsystem detects that b.h has been #included with a different set of #defines, it might be a good idea to issue a warning because that could be an indication that the codebase is about to footgun itself.

My assumption is that support for C strings (null terminated) was a design requirement.

Note that just replacing separators and newlines with nulls isn't enough to parse CSVs because the file format supports escape sequences and optional quotes around values. If you're modifying a buffer in place, you'll still be stuck copying data to replace (longer) escape sequences with their (shorter) final values.

I don't understand the line you've quoted from the article regarding why this code is slower than csv2. The csv2 code seemingly does copy from the original buffer rather than manipulating in place. Indeed, the csv2 parser copies data from a buffer into a C++ STL container object, which I would expect to be quite an expensive operation. Something does not add up here, either in my understanding or in the mechanics/description of the benchmark.

Only Zucker---- can speak to the rationale for Facebook's nudity policies. However, I've seen it speculated that FB's policy is driven by the fact that advertisers don't want to risk the blowback if one of their products were to be seen advertised next to any content religious conservatives will get upset about. Other sites that used to post various forms of semi-nudity have run into this problem.

Facebook itself may also be unwilling to risk losing advertising from conservatively-aligned companies by allowing material that their conservative clients would find offensive.

Religious conservatives can cause tremendous brand damage if a megachurch pastor happens to 'find' something blasphemous that he can use to motivate his flock to up their donations that week. For most companies it's not worth the risk of stirring up these people and it would not be surprising if this is also the case with Facebook.

Going back even further than that, however, James Burke's Connections warned of the risks of weaponized data mining in 1978, decades before Facebook commercialized the use of data mining to sell behavior modification as a service.

Dystopian futures aren't hard to predict. The hard part is getting enough people to listen to the predictions to prevent them from coming to fruition.

I'm speaking from remembering things I read over the past few decades on Usenet (yes, that far back) and on blogs. I'd be amazed if any of this is still online.

If a company goes into chapter 7, the entire company is disbanded and its assets are sold off. In that scenario, Red Hat, as a division of IBM, would cease to exist.

A lot of people--authors, technologists, public intellectuals, and others in what was the broad spectrum of 'nerd culture' at the time--saw this crap coming decades ago but lost the fight to stop it.

The writing has been on the wall about the dangers of social media for a very long time. It's just taken this long for people with the power to even consider doing anything about the dangers to start taking it seriously.

Productivity growth at the bottom end is very low, but wages at the bottom end still need to keep up with the cost of living or people who are in minimum wage jobs will no longer earn enough to survive.

I vehemently agree with this .... however I think platform intervention against potential incitement is the best of a very bad set of choices available for how to handle the next six months.

The ability of social media platforms to spread malicious propaganda intended to incite division is incredibly dangerous. The moment where a US president has refused to commit to a peaceful transfer of power should he lose the election is absolutely not the moment to test how well society can withstand the use of social media as a propaganda platform.

Society needs a long term solution for social media--my take is that it should be shut down completely--but this is a short-term emergency and now is not the time to let the perfect be the enemy of the good.

Being cautious is not a virtue when erring on the side of caution may lead to tanks in the streets.

There's a little more to preventing real wage erosion than it simply being the right thing to do from a humanitarian perspective.

The interests of individual firms are not aligned with the interests of the broader economy. It's in the interest of every firm to pay its workers as little as possible, but it's in the interest of the broader economy that most people can afford to buy goods and services. Every employee's wage is ultimately some other firm's income; keeping the game going requires that enough people are paid enough money to be able to spend it on things other than absolute essentials.

Even minimum wage workers have to pay for the basic necessities of life at prices that are largely set on the basis of what the median worker can afford. If minimum wages erode, minimum wage workers can't afford to live. This is a both a humanitarian concern and an economic problem because starving people don't make good employees, and people who don't have money can't buy goods and services from other firms. Every person who is pushed out of the economy in this way is an economic loss as they are no longer contributing to economic demand or producing goods/services with their own hands.

In a society with a social safety net, under-paid workers are also a cost to government as they will not pay taxes and will be due social assistance payments/transfers to help cover their costs for food and shelter. Further, underpaid workers may also pose a crime problem (in part because starving people will steal food), which imposes direct costs on crime victims and costs to government for police, prosecution, and corrections.

Underpaying workers benefits firms that do it, but the practice harms the overall economy and imposes externalities on others. Preventing this is why minimum wage laws exist and should keep up with the cost of living.

Along these lines, here's a graph from the Federal Reserve's data service which plots median household income in contrast to GDP per household over time.

https://fred.stlouisfed.org/graph/?g=bOfK

Note the increasing spread between GDP/household and median income. That's the US's entire socioeconomic problem--from rising consumer debt levels, rising deaths of despair, and the rise of reactionary politics--shown in one chart. The median person is falling behind.

Hey, hey, we're not talking about USB-PD here. :-) PoE actually works.

I'm not sure if there are any hollow moldings still around. The closest equivalent I'm aware of is Legrand's Wiremold[0] line, which includes cable raceways that are functionally equivalent to hollow moldings. However, these parts are designed for the commercial electrical market and would be difficult to fit into residential decor.

[0] https://www.legrand.us/wire-and-cable-management/raceway-and...

RedHat revenues likely aren't enough to keep all of half-IBM afloat. It's vaguely possible that IBM could go into Chapter 7 if things go badly enough. That's the worst-case scenario I was speaking to.

Since IBM bought RedHat, the parallels between IBM and Boeing here are too close to comfort. RedHat is potentially as vital to the Linux ecosystem as Boeing is to the (pre-pandemic) civil aviation sector. If IBM falls apart and RedHat vanishes, the impact on the Linux sector could be almost as bad as Boeing collapsing. Since Linux runs everything that's not a desktop computer or an iOS device, IBM-RedHat failing would be a very major problem for everyone.

Where to draw the line is a very hard question for which I don't have a a definitive answer.

Legally, there's precedent that RTLM[0] (Rwandan hate radio) went too far, as the people who ran it were convicted of crimes against humanity. How much lower to set the bar to protect the public while still preserving individual freedom of expression is a question I feel society isn't yet ready to answer.

I personally would set the bar at spreading false information. Most of the more vile efforts by demagogues to stir up violence are driven by fabricated propaganda. It's difficult to use facts to incite violence because reality is usually too mundane for people to get violently upset about. Non-factual propaganda, however, can be as lurid as it takes to make people get violent. Keeping outright fabrications out of public view ought to help keep some of these risks contained.

It's distasteful, for example, to spread valid statistics that connect immigrant populations with crime (in large part because such statistics, given without context, ignore integration and economic equity issues), but I feel spreading facts is less likely to trigger persecution/violence than feeding vulnerable people with a constant information diet of lies along the lines of 'immigrants eat babies.' As a result, my take on where to draw the line on political speech is 'say what you want as long as it's true.'

[0] https://en.wikipedia.org/wiki/Radio_T%C3%A9l%C3%A9vision_Lib...

Historical precedents from other societies that have collapsed into despotism make it very clear that the US is in a very dangerous place. The fact that so many Americans are in denial (or ignorant) of the risks just increases the danger.

"It'll all work out, it always has" isn't a prudent response to a political environment that has proto-Brownshirts running loose, looking for ways to kidnap elected officials.

This kind of thing doesn't usually end well for anyone except for the people who lead the Brownshirts.

The US is a two party system[0] because it uses first past the post elections. First past the post electoral systems can't support more than two viable parties.

[0] Nominally. In practice, the commonality of ideas between elected Democrats and Republicans, combined with structural advantages that give the GOP a disproportionate hold on power, mean that the US is more of a 1.5 party system.

In the case of Brexit, the sudden swing to 'leave' happened very shortly after leave ran a massive campaign of deceptive dark ads on Facebook. This advertising campaign pushed factually untrue narratives (such as that the EU was on the verge of forcing the UK to be overrun by Syrian migrants)[1] to vulnerable people and, from all indications was able to turn the referendum in leave's favor.

The sudden swing towards leave does not seem to represent a failure in polling (or media) but rather a demonstration that Facebook's dark ads platform alone can subvert the ability of a high-functioning democracy to make an informed, fact-based, decision on a question that will impact generations of people.

That should terrify a lot more people than it does.

[1] https://www.joe.co.uk/news/brexit-facebook-adverts-192164 ; https://www.channel4.com/news/factcheck/factcheck-vote-leave...

A problem arises when the appetite for 'bad ideas' aren't merely trivial things like whether Rust is better than C++ but rather consist of ideas over which of ones neighbors should be discriminated against, persecuted, imprisoned, or killed.

Providing a platform where people who subscribe to these kind of ideas can self-organize, and connect to demagogues who would lead, them is not a virtuous defense of freedom of expression; it is a real threat to the liberties and lives of the people these ideas target.

Submerging people in sealed bubble of misinformation that only reinforces their prejudices--as Facebook does through a combination of advertising, recommendations, and control over user timelines--does lead to radicalization and that radicalization can be weaponized in ways that are fatal to others.

Just as the right to swing your fists ends at someone else's nose, the right of 'bad ideas' to be amplified by platforms ought to end before people who believe in those 'bad ideas' are turned into political weapons by demagogues and would-be autocrats.

That is my usual practice. However, the pandemic has increased the prices of used/refurbished notebooks so much that it's no longer worth the risk to buy used.

The ability of an optical system (a lens, mirror, or combination) to collimate the output of a light source into a narrowly diverging beam is directly proportional to the diameter of the optical system's final element and inversely proportional to the size of the light source.

Bigger optics = less beam divergence = more useful range.

This is a function of the diffraction limit[0] and is a property of physics.

If you want to match the beam divergence of a 30" optical system, you have to use 30" optics; there are no shortcuts.

[0] https://www.edmundoptics.com/knowledge-center/application-no...

Downvotes are used in practice for 'I disagree' rather than just 'this content is inappropriate.'

Platforms that support downvotes (including HN) are prone to groupthink and vote bullying. Going against the popular view--either site-wide or against the popular view of the people who access a given comment thread--will result in downvotes even when posting material that's factually accurate. This tends to create a self-amplifying process where established views become more entrenched and, possibly, more extreme.

For example, the majority view among people who access COVID-19 posts on HN is that the disease is not a major risk. Anything posted to the contrary is highly likely to be be downvoted and/or flagged for removal. This drives away factual information in favor of groupthink and creates nothing more than a single-sided echo chamber.

HN is no better than Reddit in this regard; it's just that the groupthink hot buttons on HN are different from those on Reddit.

As this view itself is a view that runs against the universally accepted wisdom on HN, this post itself has already been downvoted and will likely be downvoted further.

As a matter of personal ethics, I only downvote material that is socially hazardous (e.g. COVID-19 misinformation), seemingly inorganic, or seemingly dishonest. Being wrong in good faith isn't a karma crime and shouldn't be punished. Unfortunately this is far from a universally accepted perspective.

With regard to why both-sides-bad arguments are not well received in the context of US politics, I feel newen covered most of that in his/her reply to you. However, I will add that 'both-sides-bad' has a really long history of being used disingenuously by the right wing, both directly and by paying left wing fringe candidates who use 'both-sides-bad' messaging to split the Democratic vote.

Both sides are indeed bad (see: e.g. Obama's execution-by-drone policy), but one side is far worse than the other (see: e.g. Trump's children-in-cages policy) and 'both-sides-bad' is usually used by the right wing as a way to trivialize Republican abuses. People who pay attention to US politics tend to be aware of this and will be less inclined to give 'both-sides-bad' comments the benefit of the doubt.

Further, on the question of respect for democratic rights, the factual history of voter suppression in the United States is so black-and-white (literally: it's a history of white conservatives preventing black people from voting) that anyone who does a 'both-sides-bad' take on this is either grossly uninformed or dishonest. Neither option will be well received by people who pay attention to these kind of issues.

It's primarily for homelab rather than pure home use, for people who are doing professional computing at home. The underlying use case is the same as in a professional environment e.g. SAN fabrics, clusters, etc.

LED car headlights do use this principle; it does have some applications.

However, the diameter of each LED optic would need to at least equal the diameter of the single optic used for a point source light. The result would be an enormous searchlight with a very broad beam.

I can't find any solid information on the quantitative light output of the AN/TVS-3, but xenon arc lamps typically produce 30 lumens per watt. Therefore, the 20KW AN/TVS-3 should produce around 600,000 lumens.

Readily available high-power LEDs are capable of 1700 lumens, so equaling the AN/TVS-3 would take--as a very rough ballpark that does not take into account the different light distribution between LEDs and discharge lamps--350 LEDs and optics. The AN/TVS-3 uses 30" optics, so those optics would also need to be at least 30" in diameter.

350 30" optics is a lot more like a wall of light than a searchlight. Aiming that many separate optics would also be a nightmare.