Well, if the usage needs are normally distributed, then 2 SD covers 95% of users and 3 SD covers 99.7%.

My take is that if no PC manufacturer has ever come close to producing a machine that fits my needs, and that's in an industry that produces bizarro machines on occasion that probably sell hundreds, I think I can safely deduce that my needs (okay, my wants) are so far out of the norm, that it would be stupid for a company to try an fulfill them.

Rather than feel feel unhappy or resentful, it's far more realistic to realize that we're simply not worth catering to.

If these new M1 devices really do get a factor of 4 increase in battery life, then given a reasonable mix of compiling and editing code with an hour of meetings and lunch, they *might* *sometimes* make it through a single workday without charging. We really still need another factor of two on top of that to do so consistently.

Indeed, having needs that are 2-3 standard deviations outside the usual is always a tough spot to be in.

If Apple (or anyone else in the industry) could get that kind battery life, they'd immediately cut the weight or cost until they were back down to 15-16 hours for normal users and 2-3 hours for those of us at the tail end of the power needs.

To do anything else would simply be producing sub-optimal machines for the vast majority of their customer base.

I remember someone complaining bitterly about the quality of their team mates and demanded match-making that put him with equally skilled team mates.

I agreed (as I wanted like equally incompetent team mates), and pointed out the nice thing was that if it was done well, *everyone* would have 50% win rates.

He was appalled and said he would drop the game instantly if they took measures that dropped his (impressive) 65% win rate.

That was the instant the light finally dawned on me as to why they hadn't implemented skill-based match-making.

Because Disney employs entire legal firms, and has the muscle to effectively do whatever they want.

The one counterpoint is that entire legal firms cost a lot of money to employ.

I wouldn't want to be the executive who has to justify spending 3/4 of a million dollars in legal fees in order to save $50,000 worth of royalties.

There's a reason that companies often settle nuisance suits of no merit. And this suit, with a semi-competent lawyer could go very badly in front a sympathetic jury.

My guess is that Disney is playing a game of chicken, and will fold if the attack seems determined.

But of course, it's not my money at risk if I'm wrong., so I certainly don't blame Foster for choosing all the cheap means at his disposal before deciding whether Disney is willing to lose a lot of money to protect their privilege of robbing him of his royalties.

Also it's a dangerous tack for Disney to take. "No, we only bought the Disney t-shirt design from that licensee - we certainly didn't buy their liabilities to Disney..."

Where extremism is concerned, I think this claim is wrong far more often than it is right. As we've seen with social media, extremism is more like a very addictive drug. Many may have a predilection towards extremism but simply not be exposed to it in regular society and live productive, happy lives.

We like to think people who engage in acts with serious consequences are truly dedicated to their task. What we find (and I'll admit I find quite frightening) is how fickle certain people are about such cataclysmic decisions. Read interviews with various Muslim terrorists: "Well, I was either going to blow up America, or become an Air Conditioning repair man, and I was sick on exam day for A/C school." (not literally). The vast majority of minds that tend toward extremism are capable of doing great damage, but aren't particularly strong.

Certainly we see this with suicide. You would figure there can be no greater, more significant decision than to take one's own life. Yet the simple expedient of making it difficult to commit suicide from a recognized landmark has saved thousands of lives because those suffering just aren't that determined to carry it out if their preferred method is thwarted.

Now, this of course brings us to what should be censored, and I don't think this is either easy or safe to determine. But the idea that censorship doesn't work is more a much desired fantasy than fact.

Also, like a virus, extremism needs both a receptive host and an effective means of spread - communication. Traditionally in American society, there hasn't been a large receptive body. But I think in the last 20-30 years, the US has grown subpopulations that *are* receptive hosts, most likely due to the declining status of poor, less-educated whites with respect to various visible minorities. (Essentially a repeat of post-civil war turmoil when some of the southern population turned to extremism like the KKK to defend their status against freed blacks. No matter how far down the totem pole of status you were, there were always visible minorities below you.)

I think in another 20-30 years, when we reach a new equilibrium, I don't think the call to extremism will be quite so strong, but right now, we have a lethal cocktail - a lot of people primed to embrace extremism, and vastly easier communication mechanisms that can expose people to such extremism, be it flat-earth, or white nationalism.

My personal experience is that DR tests are responsible for more outages than actual catastrophes. Any system that affects everything, but is only rarely used, is a huge risk factor. The old comp.risks mailing list was a gold mine for the danger of having and testing DR.

You can't win, you can't break even, and you can't get out of the game.

Usually, you lose a lot more as a result than doing it right the first time would have cost you. Greed, stupidity and amateurism at work.

I'm no longer so certain about that. Over a 40 year career, I've seen a steady degradation of the "number of 9s" required by management for system stability (or security - as similar tradeoff).

This is caused by a realization in management that customers don't value security and stability. Of course they say they do, but if you give them the a choice between stability or security vs. price, they'll choose price every single time. (Working in QA, this was a somewhat bitter experience - being told to do your job less thoroughly turning out to be the right decision...)

Now, my hyperbole aside, you can get decent stability/security at a fairly reasonable price. The inflection point on the price curve before price starts going exponential has improved tremendously over time. But the interest in going beyond that inflection point by customers you figured *really* cared about stability has dropped. Pretty much all the firms I know that specialized in those fields have gone under. New ones pop up, last a few years, and then die.

And I'd complain more about people not valuing stability and security enough, but then I look at my life. Nowhere am I willing to spend beyond the inflection point for either stability or security. I'll just live with the risk and enjoy the substantial savings. And it turns out that successful firms have made the same decision.

And those cheaper locations may be in India, Philippines, South Africa, and other places where enough English speakers can be located.

I've been outsourced twice, and while I definitely was not overjoyed about it, I couldn't really come up with a moral justification why the company should not have replaced me with 2-3 cheaper people outside of "My birthright as part of the nobility of the modern world (i.e. being born a citizen of an industrialized country) is that I am inherently worth more than others who are not".

That said, both outsourcing projects eventually collapsed in tears, but even so, there is no inherent moral reason why I deserve to be paid more based solely on where I live (and thus indirectly on my citizenship).

As a member of the professional class, I've benefited greatly from the huge influx of cheap goods supplied from overseas. It would be rather hypocritical for me to complain as that same process now moves to my industry.

In the end, as I've warned my sons, the great flattening is inevitable and that period when we've been kings of the world for no particular virtue of our own is likely coming to an end. With luck, we raise everybody's boats to a level that is acceptable to us in the Industrialized world. But we're reaching the point where we need to be prepared to compete on an even footing with the entire world, not just our those in our city.

Yikes! What industry do you work in?

In 40 years in the work force, I've yet to encounter a peer I'd consider either a psychopath or sociopath (and while tempting, I won't make judgement on someone I don't work closely with). Some people have required a bit of carefully worded encouragement to modify behaviour that other team members found difficult, but in general, people want to meet the expectations of decency and good behaviour that others have of them.

Or maybe I'm just a liberal that hasn't been mugged yet. (Well, that and I do actively avoid companies that have a reputation for discouraging collegiality.)

Fair point. I have yet to attend a meeting with more than 6 people that was useful for all concerned.

For a conference of 10+ people, teleconferencing is probably a plus, as it allows people to actually do real work while they appear to be attending the meeting (aka "Sorry, bad connection. Could you repeat that question again?")

I'll push back on this. There is a particular geek culture (heavily white) where you can have a tech meeting that is *very* efficient. People get cut off all the time (with the implicit "I understand your point now") with the possibility of raising your voice slightly and continuing ("no, you haven't"). When everyone is sympatico, a 2 hour meeting can be finished in about 30 minutes with barely anyone having gotten a full paragraph out. It's almost a verbal ballet that requires a level of mutual understanding and respect.

You can't do that over teleconference. It requires a *lot* of reading each other because you cannot afford to interrupt someone unless you are quite certain you've understood their point, and you can't afford to just raise volume and continue unless you are *certain* they don't. (Sure, there are idiots who always escalate, but it's amazing how quickly they get forgotten to be invited into such meetings.)

Now, the counterpoint to this is that it requires a very specific geek culture. Even pre-covid, I've moved away from this meeting format because as teams grow more diverse, not everyone prospers under this meeting style as there are many who are not comfortable cutting off their peers or bosses. Sure, the meetings last 3 times longer, but it means people unwilling to fight for their ideas still get heard and that's a net plus if the alternative is good ideas being lost.

However, I can see in a place like Netflix, which may have a strong geek mono-culture, people like Reed would hate the fact that the density of ideas in the meeting has significantly decreased (and he sounds like the type who conflates "value of an idea" with "how much you're willing to fight for an idea", which is very culture-specific).

> Stating a problem to be solved is easy. It's when you add all the different circumstances and what you want to do in those circumstances that creates complex - and useful - software.

Absolutely this. I remember interrogating a client about automating a manual task and what should happen in all the cases, and sub-cases, and sub-sub-cases, and he eventually became quite annoyed. How was he supposed to know what should happen for situations that almost never happen. When I asked what did the people on the ground do now, he said they "winged" it.

When I pointed out that programs can't wing it, and that he would have to specify an uambiguous policy for every possible case, no matter how rare, he cancelled the project on the spot.

Thing is, I actually agreed with him. This was a task that occasionally required a human being to basically use their judgment. Trying to quantify that judgement would take tens of thousands of lines of description, which was out of the question for a 1 man-month project.

Which means that more often than not, a complete,unambiguous description of a problem with even a modicum of complexity looks like a thick book written in lawyer-ese which no human can parse.

Which leads me to my thesis: In many cases, the source *is* the most parseable, complete, unambiguous description of a problem.

Second, I don't believe the 150k number. The US probably isn't doing well. But that much worse population-wise vs every other country? Hardly believable. We have a better medical system, we have a relatively healthy population. We are technically advanced enough to perform lots of remote work. We can sustain hardships better than most, since we have more money and our economy can absorb A LOT.

Another example of the American "fairy dust" attitude where results aren't related to specific actions, but just to "being American".

"The results can't be real" not because the USA took appropriate action but because "we're American and thus tawdry things like disease don't happen to us".

America has a *ton* of advantages, some of which you list. Sadly, they aren't all that useful (although they mitigate a little) if you don't manage the basics on how to control a highly contagious disease. And to be honest America *did* do a lot of lockdown. However, simply looking at the evidence, it wasn't enough to prevent the epidemic from growing.

Now, a debate can be had - better that people die than endure what might be necessary to actually defeat the virus. I'm cold-blooded enough to be able to argue both sides. But the idea of pretending that the facts on the ground don't exist in order to support one's case - that's obscene. Acknowledge the costs of one's policy, whether it be a huge number of deaths or crushing debt. But don't pretend that the costs don't exist.

The banking industry really needs to standardize this, one way or the other.

On the contrary. It is *so* important to get the amount right, one standard is not enough. We need at least two. And perhaps, just to be safe, three.

"beats" is doing a *lot* of work here. There are literally hundreds of definitions of "beats"

And if "beats" = "better long-term run-time performance of specific high-memory, high-CPU tasks when the on-GC program is written by high-skill programmers who have had a year or two to find all the memory leaks", then I absolutely agree.

But my experience has been that situation is fairly rare. Granted there are a bunch of other very specific definitions of "beats" for which your statement could be considered true.

More to my point, I have found that claims of "A beating B all the time" often reveal a lack of understanding that problem domains are incredibly diverse, and the chance of a solution being the optimal one in all (or almost all) problem spaces is essentially zero. Being pro or anti-GC is as sensible as being pro or anti-hammer.

(And sorry for picking on your turn of phrase, but I have dealt with far too many projects where damage has been done by those who have morphed processes or tools or technological approaches into near dogma. I consider that tendency harmful and thus I found your sentiment worth of comment.)