"An LLM is just trying to say things that sound good.  That's it, full stop.  The fact that it so often, unexpectedly and surprisingly, says things that are factually correct is an interesting phenomenon that needs to be studied, but it's still just a "reasonable continuation" calculator.  It was designed to be a calculator and we've discovered that it somehow−in ways that we really don't understand−makes great hollandaise sauce.  That's interesting and we should study how it can be that a calculator can make sauce, but in the meantime, the people in charge want to say 'We should use this sauce-making calculator to administer all food on the planet, and our health care system, and...'"

People that talk of its hallucinations have a misapprehension that "providing answers" is somehow part of its brief as a "reasonable continuation" calculator.  I would argue that ChatGPT is almost never wrong.  It was designed to produce statements that sound like something a human would say and nothing more. (Not to minimize the achievement; that's not a trivial accomplishment) I am just as impressed as everyone else by its uncanny productions, but when it produces statements that are believable yet not reflective of reality, it is still operating 100% as designed.  NFE...PEBKAC

That said, I have found ChatGPT to be a useful tool in a number of scenarios.  It is frankly astonishing how well it is able to deal with abstract concepts and metaphor in conversation.  I have used it to synthesize new ideas out of existing, but incompatible ones.  I have had several, very educational conversations and I especially like using ChatGPT to solve the "blank page" phenomenon and create JSDocs and test suites.  I always need to go back and clean up after the fact, sometimes a lot, but having it generate the starting point is pretty handy.

As a sidenote: much has been written about its weakness in math, but I think what the authors of those pieces were actually describing was "calculation", not "math".  ChatGPT very quickly and regularly falls flat when asked to calculate something, but mathematics is not about calculation.  Mathematics is about ideas and ideas are constructed with words−ChatGPT's bread and butter.  One of my academic "White Whales" has been to have an intuitive grasp of Fourier transforms.  I was able to make very concrete progress as a result of a conversation with ChatGPT.

But as useful a tool as ChatGPT can be at times, I find that it is too unreliable and unpredictable for me to consider it a first-class citizen of my toolchest.  For every conversation that astonishes or illuminates or assists, there will be two others dealing with problems seemingly less difficult where ChatGPT has no helpful input.  It is not clear what the difference is that leads to an astonishingly insightful conversation or one that is Eliza-esque nonsense.  Until that difference is made clear, I do not intend to rely on any LLM, to any degree.

I remain highly skeptical of LLMs, but my skepticism is and always has been primarily focused on humans and how this technology will be applied.  I believe the most suitable role for LLMs is as an interface to actual computation engines, ones engineered by humans that function in understandable and knowable ways.  I believe that LLMs−especially with the current state of the art−have no business being incorporated into decision systems, on any level, in any capacity.

I've never liked the  approach of lying to a child because it is believed they would not  understand the true answer, and so telling them something that's false.   So I thought for a bit and came to this explanation.

My son and I had been talking about fractions a lot, and had been drawing lots of number lines.  To give the concept some context, I explained that fractions were numbers that allowed us to talk about "parts of things".  Without fractions, using only integers, the best we would be able to do is to say things like "some" or "a little bit of".   So fractions were a new kind of number that allow us to talk about the exact amount when dealing with parts of things.

To answer his question about calculus, I drew a number line and marked off 1, 2, 3..., then asked him to mark the number that comes right after 1.  Since we'd been talking of fractions, he didn't go for 2, but instead marked right beside 1.  I circled that mark and drew a call-out to "zoom in" and show that, between those two marks was more space, that they weren't right next to each other after all.  It took  only 1 or 2 more times of zooming in for him to catch on that we could play this game forever.

I explained that calculus was a way of creating a new kind of number that allows us to talk about "the real number that comes right after".  This new number is a differential, dx.

Integration is simply a "sum" of differentials, differentials generated by a differentiated function.

Since you took the time to follow-up I will respond.

I have interviewed with Amazon 6 times over the years and have come to the decision that I want to never work for this company.  Unfortunately, Amazon seems to have no system to prevent their army of recruiters from contacting me on a near daily basis.

The final straw came with their recent practice of automating their technical pre-screening exercise.  The last time I interviewed, I did not complete the exercise and did not ever get the chance to speak with a human.  I solved the problems relatively quickly with exception of a single, hidden test case.  Was this test case a valid condition that needed to be handled?  Was it a matter of opinion with no closed-form "right" answer, open to discussion?  Was it something potentially obscured by the problem wording, that might've easily been explained by a simple interaction between humans?  Was it a bug in the test itself?  I'll never know because Amazon decided, based on this test alone, that I wasn't worth even speaking to.  That helped me finally decide.

I'm sure that Amazon gets inundated with job applicants, but I am not a commodity.  If I'm not worth even speaking to as a candidate, then why would I expect to be treated any better as an employee?

I understand that you're only one of many recruiting for Amazon and likely have no control over this, but if there is any way to remove me from Amazon's "You might also like to hire" list, I would be grateful.

There is a point of view that says "plenty of people pass the automated test, so it must work". It feels very much like a lot of teams do not value raw intelligence, but rather hire those whose point of view happens to coincide with their own, those whose natural inclination is to press the buttons that the company wants pressed. It's as if Amazon et al. are simply administering a typing test and then hiring the monkeys that happen to produce Shakespeare.

At any rate, I wasn't ever very keen on working for Amazon anyway.

There is an art to making a good tool. If you don't make the right tool, you can end up making the task more cumbersome rather than easier.  Automation is especially vulnerable to this.  Automate the wrong thing and you can make a process more restrictive instead of more flexible.

A good tool is in its essence an abstraction.  And abstractions as most people conceive of them are hand-wavy, wishy-washy concepts.  I prefer to think of abstraction as having a precise definition (copped from abstract algebra):

A collection of entities combined with a set of operations defined over those entities.

The virtue of such a formulation is the synthesis of a new thing: the abstraction, something that exists of itself and permits one to interact with it without regard to the encapsulated details.

I love tools and I seem to have a knack for making good ones.  But I've used a lot of bad tools, too.  I've given a lot of thought to tools and to why some are good and why some are so bad and I think I can summarize the problems.

Many tools have no net benefit.  They simply replace one action with another.  Modern "language designers" commit this sin with every C-like-syntax-transpiled-in-a-novel-toolchain hobby project that gets attention because the author works at one of the Big 3.  Simply replacing one action with another does not constitute a good tool.  And, in the extreme, such a tool can make things worse when the old abstraction is hidden by a leaky abstraction, forcing the user to care about both (similarly to the next sin).

Other tools actually add steps to a process, increasing the total work.  In the worst case, the original action must still be performed, in addition to using the tool.  This seems to be a failure unique to computer tools as no sane person would use a hammer that required hammering something else before you could hammer a nail.  But plenty of users seem to think it perfectly normal to have to manually edit input text files, character by character, before feeding them into a tool to process the text.

But worst IMO are the tools that actually restrict the user or prohibit full use of existing capabilities.  From hence is born the all too common lament that "the system won't let me do that."

All of this technology, and yet so much "can't".

Obviously, a lot of restrictions are deliberate power grabs by unchecked corporations, but many still are the result of poor tool design.  Such tools say to their users "stop what you're doing and do it this way from now on".  Bad tools are very much the tail wagging the dog, demanding care and feeding from their users while providing reduced or restricted benefit as compared to not using the tool.

Good tools should be genuinely useful, interesting to learn, and should delight their users while using them.

How an interview code submission that wasn’t even submitted changed our process

I recall being in assembly class in school and a classmate said "I can't wait to be done with this so I don't have to do it again" .  I was taken aback—I enjoyed assembly class (and still occasionally write assembly for my rpi).  I didn't understand why he was there if he wasn't interested in the subject matter.  Was it just for the future paycheck?  That's valid...I guess.

But as it turns out, it's been my experience that he is representative of a large portion of the programmer population: people that don't enjoy programming.  They don't care; they just want to get to the next ticket, keep their sprint velocity up, collect their paycheck, and then go do what they really want to do (which is not programming).

Everything is a computer now, which means that software engineers are really everything engineers.  Most of you reading this are not software engineers and so don't see the sausage being made, but it's been my experience that most software engineers are not engineers.  The employees that are writing the code that controls more and more of the physical world around us don't have that deep spark of interest that drives them to get it right, just right enough to close the sprint as fast as humanly possible.

When you sacrifice good code for good enough code, you leave holes in what you've built.  And they know this; they've come up with a cute name for it: technical debt.  The implication is that technical debt is OK; it's just the cost of their sins which will one day be paid for.  But like many other kinds of debt that day never seems to be today.  They're accumulating tech debt.  They're weakening the foundation to build just a little bit faster.  We're living at the top of a very tall Jenga tower and as long as the software engineers can keep their sprint velocity up, can keep building the top of the tower higher, they will feel justified in continuing their practices.

"As I told you yesterday, that assignment was due last Friday. If everyone else is required to hand in their work on time, I see no reason that you should be the exception…"

"Who do you think you are?," he interrupted, infuriated.  "A college professor who thinks he's smarter than the rest of us?"

"…and, as I've repeatedly tried to explain, Mr. McConnell, you aren't a student at this school."

Rather than just list a series of references, I'd like to answer your question by explaining a little of my history with math or: How I Got Addicted to Mathematics.

When I dropped out of college for the second time (no money), I eventually went to work in the public school system as an instructional aid.  At that time, my math education went no further than some high school algebra—which I didn't complete—and some college algebra—which I also didn't complete.  I had my own computer and had been learning to program and, of course, wanted to try making my own games.  As I began reading about various graphics techniques, more and more math came up and, before long, I was at a stand-still, spending hours agonizing over single sentences, trying to understand the math.  One technique involving camera positioning mentioned something called "quaternions".  I actually went to the university library and checked out the original book on the subject. I didn't make it past the first few pages.

Meanwhile, at work, it quickly became apparent (even to the students) that my teaching ability far outshone the teacher's and I was not asked to help with the instruction anymore.  I became quite efficient at making copies, so I found myself with lots of free time at work.  Then I made one of the most important realizations of my life.  I still had the text from my college algebra course and I realized: this is a book.  I realized that books are made to be read—there's text in there that explains the subject.  I've discovered that my school experiences differ wildly from others', so I should explain that we never read, or were assigned reading from, our math books in public school.  They were simply the source of the exercises.  I thought, "I don't know how far I'll be able to get on my own, but I bet I could at least get past the first chapter or two."  So I opened to page 1 and started reading.  I finished that book, snagged a free geometry book from a cart on its way out the door to the dumpster, worked through that, borrowed a book on trigonometry from one of the only math teachers at that school that was worth a hoot, and worked through that (I was really fast at making the copies now, so I had lots of time).  This took about a year and a half, after which I joined the Navy.  On a side note, towards the end, as I was studying and would try to ask for help from the teachers (it was a school after all), there came a point where one of the teachers said outright, "I don't know how to do the math you're doing. I can't help you."

Not to worry though, that was in the 1990's. I'm sure the standards have gone up since then.

My math progressed a little bit slower in the Navy.  I also had a great deal of free time while out to sea, but I spent a lot of it learning to program.  I taught myself calculus.  I got as far as the equivalent of first semester calculus, which I CLEP'd out of and dabbled a little in other areas—probability, counting theory, number theory, linear algebra, etc.  I left active duty and immediately rolled right into full-time school.  First semester I took calc II—and hated it.  Remember, this was the first math class that I actually completed after junior high school.  I did not enjoy any of the math classes that I took while a full-time student.  The pace of the courses precludes any attempt at reflection or digestion—there's no time to play with any of it.  They just shove it down your throat and want you to regurgitate it at the exam.  I'm so glad that I'm done with that.  Ironically, I think that I'll go back for a Master's in math at some point, but first I want to spend a few years studying on my own again, if for no other reason than to remember what I like about math.

So, how would I recommend learning math?  Don't take it too seriously.  Do it for fun.  Enjoy it (and maybe find a job where you can get paid to sit in the back of the room and study).  If you go the solo route, then find resources to ask questions of, but you must also learn to self-assess.  It's easy to think the words "I understand this", but it's much more difficult to actually determine whether you do or not.

Figure out what you're wanting to accomplish and why.  If you want a degree in math or a job that requires such a degree, you'll have to go to school.  If you simply want to understand more about math—stay as far away from school as possible (I'm only partly kidding).  Here are a couple of excellent articles on math education:

A Mathematician's Lament

Whatmath

Why do so many people have such misconceptions about Mathematics? The great misconception about mathematics -- and it stifles and thwarts more students than any other single thing -- is the notion that mathematics is about formulas and cranking out

Robert H LewisAcademia.edu

Either way (school or no school), you'll want a book.  Mostly, the textbooks are homogenized these days, so I don't know how much it matters which one you get.  Personally, what I look for in a textbook is generality.  If I'm just starting a subject, I don't want something very specific or niche.  I want a broad treatment.  A good rule of thumb is the fewer words it takes to describe it the better.  College Algebra with Applications to Social Sciences for Art Majors should probably go back on the shelf.  The book that I have is below, but there is nothing special about this book aside from it being the book required for the course that I never completed (I do like the book though).  Heh, I just realized that I never did take a college algebra course.

College Algebra

Providing Strategies for Success: This text provides numerous strategies for success for both students and instructors. Instructors will...

Mark Dugopolski

Finally, if the goal is only to be able to understand the Mandelbrot set, that is a mountain at the end of a road that is much longer than you might think.  As I said, for all of my study, I can only glimpse some of the pieces; I can not claim mastery.  A book you should probably check out is:

Chaos

A work of popular science in the tradition of Stephen Hawking and Carl Sagan, this 20th-anniversary edition of James Gleick’s groundbreak...

James Gleick

This book was my introduction to the Mandelbrot set and what inspired the first version of my plotting program.

I've taken the machine gun approach to answering your question; I've thrown lots of stuff out there. Take what you find helpful and throw the rest away.

I've always wondered how differential gearing worked, but never taken the time to go study it.  This is an awesome video series.  What's really striking though is the pedagogy.  Take particular note of the lack of PowerPoint slides, interactive demonstrations, hands-on practice, group discussions, or any of the myriad of supposedly indispensable instructional enhancements on which modern educators waste time and money.  Was the instructor actually teaching us using a non-dynamic, unilateral, throat modulated flow of information?  GASP!  A lecture!  The lowest form of instructional technique, tantamount to torture!

Take note of how clear and concise the ideas form in your mind.  Note how easy the concepts are to grasp.  Note also how easily the concepts rest in your mind.  Even days later, you will still be able to recall the basic operation of the mechanisms.  Remember that this is from 1953.  Note the technology used as instructional aids and then compare this to what's happening in our modern education system.  People get up in arms about lots of stuff.  The economy, health care, X's rights.  Not education.  There are little brush fires from time to time about education, but which lack the scope and intensity warranted by the problem.  There should be a four alarm fire in every state, city, and home.

Every civilization collapses−name one before the current that hasn't.  I don't believe history will say ours collapsed from war, poverty, health care, or a deficit or surfeit of X's rights.  The economy might beat education to the punch, but I suspect that's because the two are related.  We will fail because we bought Johnny a calculator instead of teaching him how to add.  We will fail because we bought Susie an iPod instead of teaching her how to read.  I think history will say we failed as a civilization because we failed to teach the next generation how to keep the lights on.

I recently acquired two cheap laptops for my children to use for school. Upon first boot, I was greeted by the normal configuration screens and then the ubiquitous End-User License Agreement screen.  How I hate them!  Except that there were two EULA's.  I noticed that the first was the familiar Windows EULA that had a note "Required to use Windows". Ok, that's normal—obnoxious and inappropriate—but normal.  The second EULA was from the manufacturer (Acer) and had a note "Required to use the computer".  An End-User License Agreement to use the computer?!!  I'm not an End-User of the computer—I AM THE OWNER!  I am not being granted a license to experience the wonder that is Intel integrated HD graphics; money has changed hands—I AM THE OWNER!  There are no terms to which I must agree—I AM THE OWNER!  We may have been tricked into drinking the "Intellectual Property" kook-aid, but hardware is still something we OWN!  Have we become so inured, so dependent on the corporate masters of our technology, that we have forgotten the concept of ownership entirely? The rubbish arguments and pop philosophies that accompany "Intellectual Property" are apparently taking a toll, not only on the development and use of computer technology, but on our collective psyche as well. "EULA's? Sure. They're all over the place. You can't get around them, so just click through it. Stop making it such a big deal!"

This is how rights and freedoms are lost; whether of laziness or of faith, the blind eye we turn to daily inequities or to lasting legislation does not see the washout gully until home and croft are swept away by the slow erosion of the little things that we didn't sweat.

1. The terms being dictated by the licensor include the right to change the agreement at will.  An agreement that one side can unilaterally change is not an "agreement" by any definition that could pass any sort of "reasonable person" test.
2. The terms of the agreements are completely vague.

Reading one—as a reasonable person—it all sounds like an acceptable deal.  Putting on my "lawyer hat", I start to think about things like "Who are these 'select partners' that are exempt from the restrictions the licensor has placed on the agreement?"  Direct from the new LinkedIn User Agreement:

"The profile you create on LinkedIn (including Slideshare) will become part of LinkedIn and except for the content and information that you license to us is owned by LinkedIn."

That sounds reasonable to me.  What's mine is mine, but the profile is theirs, right?  That sounds reasonable, but then I wonder what is a 'profile'?  If you started to mentally answer that then you've fallen into the trap that it seems most people do: you thought to apply reason in trying to understand the law.  The two have nothing to do with each other; there's zero overlap between reason and law except by sheer coincidence.  That's not a statement of sarcasm or cynicism; that's a statement of pure observation.  Where does it say that laws have to make sense?  Indeed, not all do.  It may be that we want laws that make sense and that we want the people who make the laws to make laws that make sense, but once a thing has become law, that's it—it's not law because it makes sense; it's the law because it's the law.  Attempting to apply reason to understand a law is non-sense.

So what is a LinkedIn "profile"?  It doesn't make any difference what you think one is or what you think "makes sense" for one to be.  The sole standard by which it can be judged is what the definition of a LinkedIn profile says a LinkedIn profile is.  And, should you decide to go read the LinkedIn User Agreement, you'll notice—there is no definition of what constitutes a LinkedIn profile.  That means that the definition gets to be invented, on the spot, in court, to be whatever suits LinkedIn. (I suspect this is a common reason for settling out of court.  Not to avoid cost, or avoid the publicity, but to avoid using this one-time, blank check)  You can rest assured that when Facebook purchased Oculus no such ambiguities existed in the agreement.  If these ubiquitous contracts are going to be legally binding on individuals, shouldn't they be required to match the same level of legal "completeness" as is expected by business-to-business agreements?  Oh, right.  Normal citizens aren't supposed to make such a big deal out of this stuff and most don't care.  Please disregard and return to your regularly scheduled spectacles—and here, have this loaf of bread shaped like an iPhone.  Don't forget to accept the user agreement.

"You should push some of those buttons and do more of that technical stuff.  There, I did the hard part for you."

Somehow that's loftier and more important than actually knowing how to do the work.  It's amusing to think of the common, comedic plot depicting a young, up-and-coming individual who's just displayed their talent being swarmed by a mob of shady, quick talkers all wanting to be their "manager".  Yet when applied to a person with a suit and tie, the term suddenly becomes laudable, if no closer to any actual productivity.

"But someone has to set a direction.  Someone has to say which button to push.  Someone has to manage."

Well, of course someone has to say which button to push, you blithering twit!  The fallacy behind this—and, indeed, that which underlies all of management apologetics—is that it assumes that you do not have to be intelligent enough to know how to push the button yet could, mysteriously, still be intelligent enough to say which button to push.  And, conversely, that those who are intelligent enough to know how to push the button could not possibly be able to tell themselves which button to push without having someone less competent than them set a "functional direction".

"But we have to have some kind of structure for society.  What other option is there?"

What if those that actually did the thing got paid the most?  What if we turned the structure on its head and exalted those that actually produce the thing over those that merely say it should be produced?  Sure, having someone around to keep track of the insidious "big picture" does free up the producers to do more producing.  If you listen closely to that statement, however, you will hear the inherent priority.  Managers have their place, and that place is in support of the producers—not living large on their backs.  I'm not suggesting that every assembly line worker is actually qualified (or even desirous) of running the company.  I'm just imagining a world where the proficient are not enslaved to the incompetent and where managers and other support workers aspire to—one day, if they work hard—become able to contribute something more tangible than "direction", instead of the other way around.

"Good luck with that."

People are beginning to awaken to the absurdity (as they always do) of the top-heavy structure created by putting the least competent in charge.  As this happens, we will see executive stubbornness brought to a new level and words like "vision" and "entrepreneurial ability" desperately touted as salvation.  As the people that truly execute against the plan start to question the decisions made by the people whose only product is decision-making, we will see the decision-makers stubbornly reject all opposition and oversight.  When the true executives fully realize that the managers' decisions are no less random, no more informed, no better than their own, the societal dialog will shift to the question most feared by those in power: why do managers—those that actually do the least—deserve so much more than everyone else, than those actually doing the thing?

Think of every, single power structure in society.  At the top, you will find the distributors, the middle-men.  I think it is one of history's greatest scams, the ultimate pyramid scheme, that they were able to insert themselves into this position.  Ask any highwayman-turned-robber baron and of course they will want to talk of how necessary they are to the transfer of goods and services.  But how is it that the person who facilitates the transfer of goods is more important and better compensated than either the producer or the consumer of those goods? This observation also extends to government.  I don't refer to any particular government, but to the very notion of a governing body.  Are they not the middle-men in the contract of society, facilitating the transfer of its benefits among its members?

Whatever the role of the middle-men in the past, the idea of the Internet has within it the potential to make them completely obsolete.  I no longer need the RIAA in order to have access to a wide selection of music.  Moreover, if the RIAA didn't exist I would have access to no less music than I do now.  Representative government was a compromise accepted during an age when a full gathering of society's members would not have been feasible and free communication would not have been possible.  Just as a foreign body triggers an immune system response,  the capability of direct connection is an existential threat to the middle-men and they are trying to kill it.  The ability to connect directly means I don't need to pay someone for the privilege.  The ability to connect directly means that We The People have new possibilities for how we might provide for our future security, possibilities for greater freedom as a society that didn't exist before.  That's why we're seeing the governing bodies issuing edict after edict, limiting what can be done with OUR! technology, negating the existence of the technology by prohibiting the capabilities it provides, so they can keep things as they have been.  I'm trying to look at the long term direction of technology and I don't see greater freedom, I see greater control and I see that it's too late to do anything about it.  In our consumer-based economy, the only private action we are permitted is to vote with our dollars.  Are any of us willing, or even able, to choose to not use the Internet?

The Internet was only ever a dream.  Now it is just the Net and we are all caught in it.

Why do we still need ads at all?

Perhaps in the past there was a role for advertising.  In the past, one could conceivably have resided in the same region as a widget maker and the two not have had any knowledge of each other's existence.  If no one knows you make widgets then how can they buy them?  But in an age where information is so easily accessible, however, advertising feels as useless and out-dated as a phonebook. (Incidentally, have you looked at a phonebook recently? They're mostly ads now.) If I want a widget, I have the ability to see if there's a widget maker next door, or in China—or anywhere else in the world—that will sell one to me at a reasonable price.

Capitalism requires an educated public that can make informed, rational decisions.  Advertising is the picture-of-a-french-fry-on-the-button version of capitalism.  It is likely that advertising is now inextricably inter-woven into our economy and is here to stay, and this offends me.  Basic economics courses talk about "economic inputs" and discuss one, cleverly named "entrepreneurial ability".  This term is used to justify the offensive salaries "warranted" executives.  The notion is that these well-compensated individuals earn their summer homes by "thinking big thoughts".  We have made some pretty cool technological advances, but technology provides capabilities; it is how we apply those capabilities that matters.  The first TV commercial aired in 1941.  Since then, executive compensation has skyrocketed and our technology would be viewed as near mystical by their standards, yet the focus of all our technology is still putting ads in front of eyeballs. Has not one of these "executives" had any better ideas in the past 70 years than show ads to humans?

I see advertisements all over the place, but somehow our leaders seem unable to pay for the same level of society as existed before this technology was invented.  These are the "big thoughts" that are worth a 7-figure salary?  Does every new technology represent merely a new channel over which advertising may be delivered?  Is this all we can expect from the future—more and better ads?  I do not reject advertising because it is annoying, or intrusive, or invasive; no, better ads will not sate me.  It is the very existence of advertising that offends me.  I believe we can do better things with our technology.

I don't have cable or an antenna, so I don't see the current shows or commercials.  Once or twice a year, I'll be in a hotel somewhere and have nothing to do but watch TV, so I'll flip it on to see if I can find anything good (it's more an exercise in re-confirming my disdain).  Recently, this was the case and I was appalled at the totality of the suck.  Most of you—yes, you, the brainwashed, TV-watching morons—are not able to appreciate the downward plunge because you are the proverbial frogs being cooked alive in your own gradually boiling pot. Because you are constantly inundated by it (TV's, phones, iPads, etc.), you don't see just how awful it is.  It would be dismissive to assume that the industry is unable to do better.  The industry knows the shows suck—and they know that you want to watch them.  Quality costs money.  Why spend the money on quality when you're just as likely to watch reality crap.  They were feeding you crack before, but now they know that you'll be just as satisfied with drain cleaner.  Don't be so naive as to think that they are pocketing the profits.  They are re-investing the profits in buying laws to ensure that things stay just as they are.  They know that you don't possess the intelligence to comprehend that they are taking your freedom in order to keep you happy with the crap they are feeding you.  They are selling your own fat asses back to you.