obviouslynotme

Yeah, that is why I said TS (or something similar). TS made some decisions that make sense at the time, but do not help compilation. The complexity of its typing system is another problem. I'm pretty sure that it is Turing-complete. That doesn't remove feasibility, but it increases the complexity of compiling it by a whole lot. When you add onto this the fact that "the compiler is the spec," you really get bogged down. It would be much easier to recognize a sensible subset of TS. You could probably even have the type checker throw a WTFisThisGuyDoing flag and just immediately downgrade it to an any.

I have thought about doing this and I just can't get around the fact that you can't get much better performance in JS. The best you could probably do is transpile the JS into V8 C++ calls.

The really cool optimizations come from compiling TypeScript, or something close to it. You could use types to get enormous gains. Anything without typing gets the default slow JS calls. Interfaces can get reduced to vtables or maybe even straight calls, possibly on structs instead of maps. You could have an Int and Float type that degrade into Number that just sit inside registers.

The main problem is that both TS and V8 are fast-moving, non-standard targets. You could only really do such a project with a big team. Maintaining compatibility would be a job by itself.

What do you mean, "So?" This is maybe one step below being removed from the ability to have a checking account. He has been removed from Internet commerce with Zero Recourse and even gaslit by the company harming him.

This will keep happening because people like you will just flippantly continue saying, "It's not my problem." It will become your problem soon. As companies continue ignoring completely screwing a slowly growing percentage of people, that demographic will do what he will likely do next, call his local congressional representative to grab your company by the metaphorical ear.

Internet companies have incredible legal privilege right now. As expected, they are now abusing it as they have grown large enough. I'm glad these companies are filled with people like you because you don't even mask yourself. That is what will make the boot come down faster and harder.

It happens in all fields, at all levels. You cannot trust any papers without reproducing them. The intentional fraud is rampant. There is too much reward for publishing junk as quickly as possible and too little punishment for misbehavior.

Apple is just as sleezy as Google but doesn't have as much power and reach. If I don't want to, I don't have to interact with Apple at all. Google is much harder to avoid, impossible in some domains like email.

Except large code bases do the same. They regularly die when their ability to be understood drops too low. Even with well organized code, they are pushed to add features until they aren't understood at the deepest level. Once you hit millions of lines of code, even when you spend decades in that code base, you still forget changes you made even if you have an overarching picture. That's ignoring other people working on it all the time. The understanding gets reduced to contracts, types, and interfaces.

And most importantly, humans are more complicated than code. With enough time and knowledge, I can accurately tell you what any piece of code does on a single expression or statement. Humans regularly do things they don't even know for purposes they don't understand.

I explain it in the third paragraph but to illustrate it further: Consider a function. A function that is 1 line long is immediately understandable. At 10 lines, it is readable within a minute or two. At 100 lines, it is maybe legible to someone who lives in that function. At 1000 lines, it is a black box. Human organizations are the same way.

You might suggest refactoring, which is what companies do too. They create departments, promotion ladders, org charts, and mission statements. The problem is that abstractions leak by design. As your abstractions accumulate and change outside your view, your ability to understand the entirety reduces.

It's less about risk aversion than it is about position, size, and complexity. As these things grow, the incentives change and the ability to understand what the organization even is becomes impossible.

A startup starts at the bottom. It begs investors, customers, and employees from a position of optimism and humility. The organization enthusiastically changes itself to find a good balance between those three or it dies. As the organization grows, it starts demanding everyone else change for them instead. Google's interviews are an example. Its famous customer service is an example.

Then we get to size and complexity. Thanks to Dunbar's numbers, we know that there are numerical limits to a human's ability to know people. This makes sense. I can know everything about 6 people, most things about 50, and keep track of about 250 well enough. As the organization grows, your ability to know it disappears. You begin making abstractions. Instead of knowing exactly what Susan does, you say she works in X Department, for Y Initiative, doing Z position.

Google is so big that one person can't understand it anymore. The inevitable reduction to a corporate abstraction occurs and then people treat it like the X Company, which is just like Y Company but makes X instead of Y. Short term revenue and expenses are the only measures at the end.

And in this faceless abstraction, the professional parasite class infests and extracts resources and morale. Eventually the C-suite stops fighting it and joins in on it until only the sheer size and momentum of the company keeps it going. Maybe an investor group will come and force a rework of the company, but not before the company is just a shadow of the shadow of its former self.

Freedom not only allows jerks and kooks, but it requires them. We need sovcits and police auditors, because they define the boundary and remind the police and prosecutors, whom are apt to forget, where that boundary is.

Everyone demands victims be kind and normal, but that is rarely the case. Kind and normal people stay far away from the boundaries by nature, but the boundaries always shrink when there is no push-back. We need the kooks out there getting tazed so we don't have to.

Here's the actual problem. Kids are cute and annoying. You love them but they will drive you nuts if you don't drain their energy and attention.

The previous solution to this was that parents booted their kids out of the house and made them play outside. This has somehow gone from normal parenting to abusive behavior that could get you arrested in some places with a possibility of losing the child.

So instead, we throw tvs, ipads, and phones at them just so we can get a modicum of quiet. Not surprisingly, each one of those is bad for the kid. The developmental problem is screen time but the scary one is strangers on the Internet. Elsa Gate upset a lot of people, and that's not the worst thing that happens.

Policing the Internet is a bad decision. Any Internet Safety Committee will eventually be staffed by the people you would least like to be there. That's especially problematic because one person's bad guy is another person's hero a lot of the time. Although this is the inevitable future due to increasing clamor by Internet companies for regulatory capture, it won't even be effective.

Parental controls never work. Your kid has way more time and motivation to figure out workarounds, even if the parents and the company set everything correctly. Everyone also knows that there will be at least one friend whose parents don't care. This is true whether the parental controls are local or done by a central committee.

The only solution to this mess is:

* Remove the insane laws and policies that prevent parents from letting their children play outside, especially unaccompanied.

* Educate parents on the importance of moderating and monitoring electronics use.

* Educate parents on how to educate their children about the dangers of the Internet, especially social media, pornography, and pedophiles.

* Accept that some parents don't care and will neglect their children no matter what you do. Instead, empower the parents that do care to easily and safely raise their children.

>Code which was committed to repositories didn't match the actual running code

I remember the first time I saw this at a small company. The head programmer was the only one with the actual production source, despite a working source control system. He complained about how people didn't appreciate how hard it is to make builds work. I should have left when we lost three man months of code to losing email. I only left after screaming at the head programmer for refusing to stop making the repository unbuildable at all.

As I said, RDD is the dominant paradigm because it is rewarded. The career trajectory of someone who hops companies versus someone who doesn't is wildly different.

I also pointed out that the problem isn't with the developers but the leadership, starting at the top. The programmers are just acting rationally in the environment they are given. The value of a programmer is exponential with time in the code, and yet, they aren't compensated for it. What they are compensated for instead is decorating their resume for their next job.

Microservices are Resume Driven Development, the leading paradigm of our industry and others. People will always optimize for the incentive structures of their environments.

The most important person in every company is the CEO. He sets down the culture of the company and creates these incentives either directly or indirectly. The only solution to RDD is for the CEO to be tech oriented, avoid hiring people who jump companies every couple years, build a great work environment, and compensate people well.

IQ is an incredibly useful metric and why we sort our society on SAT scores. It's just not the only useful metric.

Most people are totalitarian in a few ways. Some people hate drugs. Everyone hates crime. The homeless cause social issues. Some people want their religion enforced. Some people hate foreigners.

Most people grant these powers to the government hoping to get their pet peeve taken care of, and are surprised when the government not only doesn't do what they promised, but turn those powers onto its own favorite victims instead.

The one time I was asked to be searched while driving I declined. They immediately called a drug dog, yanked its leash, and called it a drug hit. I then hung out for an hour while they confirmed I was a poor college student with zero drugs.

Driving with any amount of cash above a few hundred will get you robbed by the police in asset forfeiture. It happens everyday around the country to small business owners that deal with a lot of cash.

Sizing values is also great for reducing bandwidth footprint, either network or memory. The really good optimizations come from this.

All governmental prosecution should be before a court with the protection of rights. Even in your contrived example, the defendant has the right to face his accuser, cross-examine, attorneys, judges, juries, and the many things we throw in the government's way of harming people, justified or not.

When the USG tells anyone to do something, chances are they will comply, legal or not, just because it isn't worth the pain and suffering of fighting, especially for someone you don't even know. We have relearning what it is like to have your personal life ruled by people you have never met in places you have never been. The USG has stepped too far and the overreaction to public/private partnerships is coming.

If management cared, not updating documentation with every merge would be a strike, with multiple offenses leading to firing. They don't care. All large development is just slapped together and shipped. By the time the reaper of long term problems comes, the problems are too big to fix.

That wouldn't fix anything at all. You still get that PhD to do research and publish. The only fix for this is funded and career-advancing randomized replication. We should lean harder into the scientific method, not withdraw from it. Absolutely nothing else will work.

It's worse that that. Most companies give career incentives to new projects. Google is probably the most well known offender, but they aren't the only one.

Very few companies understand that the most valuable asset of a programmer is his mental model of your system. You need people who can do the hard work because they understand the system at an intuitive level.

There are already many energy storage systems in place and many more being installed even now. The problem is that energy storage and recovery is a hard problem. Anyone who solves it will be rich overnight.

Power generation comes in two classes, baseline and peaker. Baseline units are on 24/7. They are typically very efficient, but can take days or weeks to start. Peakers can turn on in minutes of notice and are usually better at following demand.

Power demand follows daily sinusoidal cycles. By having an economical consumer of electricity that can choose operation times, you can have them use power at night in order to flatten your sinusoidal demand, thus allowing a higher percentage of efficient baseline units to run instead of peakers.

My attorney friend has explained to me that SLAPP is very hard to get, even when it is obvious, but especially when experts are involved. It's the civil equivalent of malicious prosecution, which is also notoriously difficult to get. Judges have a lot of sympathy for the attorneys that are in their courts everyday.

This is only a premonition of the things to come. Our universities have been hell bent on destroying themselves for the past few decades. They stole as much as they could as the traditional gatekeepers to the upper-middle class, but even the most clueless are realizing the missing value proposition. College enrollment has been dropping from its peak in 2010.[1] As the amount of freshmen and state funding drops, the parasites will fire everyone except their fellow parasites until the hosts are consumed.

1. https://www.statista.com/statistics/183995/us-college-enroll...

It is funded by the US Government. The only shills it will find are Russian, maybe the Chinese these days, the politically unpopular, and social media marketers.

Software productivity is all but impossible to measure and insanely detrimental. We tried lines of code, story points, key process indicators, and others, but all they do is pervert the process into an exercise in cooperative stupidity. Without metrics, ratings are popularity contests. With metrics, ratings are popularity contests skewed by metric gaming. You cannot get rid of popularity contests. You can only ensure that you punish even more the people whose drives to work on important things is more powerful than their drives for success.

As someone who maintains a large Fortran codebase actively maintained from the 50s, I can say with 100% confidence that syntax, compiler, and other tools aren't even 10% of getting up to speed. It's some of the worst code you will ever see. A lot of it predates "GOTO considered harmful." It also comes from an era where different common blocks and subroutines were moved into and out of memory using a custom virtual memory system.

The demand for Fortran/Cobol experience has nothing to do with training. We need to make sure you are masochistic enough to trudge through the sludge.

It has little to do with the look and feel. Loading a "new" reddit page feels like going on a 2010s news page without an adblocker. The page takes 15 seconds or more to load. Elements are shifting around during that time. If I'm not logged in, I get a screengrabbing beg ad to install their apps.

Reddit code is trash. The old version is just much simpler and therefore less annoying. If it weren't for the fact that the smaller subreddits have lots of value, no one would put up with it.

>It may be possible that there is no English sentence you can utter about anything which is both explanatory and fully accurate.

FTFY. Philosophy aside, humans understand complex things by building simpler abstractions for them. I think the key to NN understandability centers on building simpler NNs focused on specific tasks that build into bigger ones.