The problem is really one of supply and demand. Whatever SV talking heads say is a post-hoc rationalization on top of this basic fact.
We have too many PhDs (I say this as one). It's never been easier to get one. Most PhD topics are incremental and derivative whereas they should be seminal and ground-breaking.
Unfortunately, with credential inflation, this cycle will escalate. Soon people will complete two just to qualify for an academic position.
I would blame the monopolization of the economy. A few corporations purchasing big chunks of the industry control the job market create a bottleneck where supply of jobs is controlled by a few corporations. Once all jobs are controlled by a few decision makers the precarious work conditions, diminished salaries, abuses, etc. come naturally.
> Unfortunately, with credential inflation, this cycle will escalate.
Even if everybody had high education, companies would still compete for the best employees. There is no competition for employees because large corporations have agreed to not do so.
Apple, Google, Intel... discussed no-poach as a way to keep salaries low. Has anything changed?
Do you think the boundary of science isn't pushed forward incrementally? Not every person can be an Einstein, hell, not every generation has an Einstein. And Einstein couldn't have done what he did for science without the foundation of those "incremental and derivative" advancements.
This nonsense falls apart at the barest inspection. Science IS BORING. And it should be.
Take for example a muscle building study that found that the biceps grew significantly more when tension was maximized in the stretched position. Science based lifting people hawked for years that the "stretch mediated growth" was king. All based on that one "seminal and groundbreaking" research. Years later when a "incremental and derivative" study was done on the hamstrings found no stretch mediated growth effect. Without the boring work, we wouldn't know that some muscles grow faster when tensioned under stress and some don't. And we still don't know exactly why. The current leading theory is it's something to do with the balance of fast vs slow switch fibers that make up the muscle, but we don't know without more derivative and incremental research.
Hell even under your criteria, if the stretch mediated effect wasn't found in the original study you'd probably classified it as incremental and derivative too.
Want another example? How about this one, a scientist was studying which tricep movement produced the most growth. It's obvious right? It's the one that lets you load the most weight onto the triceps, or at least the one that lets you load the most weight onto the most heads of the triceps. Boring. Derivative. Incremental. Except this study found that despite "common sense" it was actually the overhead tricep extension. You can't load it the heaviest, it's mainly targeting just one head of the tricep, it makes absolutely no sense. But science has proven it to be the case. Later "incremental and derivative" research has proposed a theory that since it's overhead, the muscles go slightly hypoxic during the lift and that triggers a stronger growth reaction, and in fact, applying a band for vasoconstriction around the arm and doing bicep curls was found to lead to more bicep growth than doing it without the vasoconstriction. All of this is incremental science. All of this advances our knowledge of how the body grows.
Science is slow. Science is advanced unpredictability. Science is boring.
Anyone can cherry pick examples to support that science is incremental (or not). The current structure of academic science struggles to reward creative thinking, struggles to support eccentric thinking, and struggles to move outside of their ivory domain based towers. It’s both a bureaucratic issue and one of hierarchy and power within science itself. I have seen far too many physicists resist changing how they teach because they have already figured out how to educate how dare you question them. I have seen far too many seismologists refuse to use non acoustic data sets because why wouldnt seismic data be enough? These are often even young people who refuse to step outside of their domains point of view perhaps from fear that they will never secure a faculty position. Additionally it is often times driven by university politics and finances. For example, Most R1 universities large revenue source is grant overheads, and yet most faculty have little say on how those overheads are spent because university democracies and leadership have been replaced with administrators building bureaucracies. I say this as a scientist for 15 years whose published over 30 papers, won grants, advised phds and postdocs, etc. the system would do well to change if only to give more time back to scientists to do science they find interesting instead of what can be keyworded in to grant applications.
> Anyone can cherry pick examples to support that science is incremental
This is not a rebuttal of what I stated. You dismiss my data and provide no data of your own, just feelings. I appreciate what you're trying to say, but bring data or else we can't discuss it meaningfully.
Nietzsche argued that genius is more frequent than we think, but that something else is missing for its realization ("the five hundred hands"):
> In the realm of genius, might the “Raphael without hands” — the term understood in its broadest sense — be not the exception, but the rule? — Genius is perhaps not so rare after all: but the five hundred hands it needs to tyrannize the καιρὁς, “the right time” — to seize chance by the scruff of the neck! [0]
“I am, somehow, less interested in the weight and convolutions of Einstein’s brain than in the near certainty that people of equal talent have lived and died in cotton fields and sweatshops”
This chestnut gets trodded out every so often but it's frankly absolute and total nonsense sold by anti-intellectuals and bought by people from all walks of life.
Science is and always was incremental. The breakthroughs come from truly unforeseeable places. It takes seemingly niche and unprofitable and incremental research like studying bacteria living in volcano vents, for us to have PCR.
VCs expect a sliver of their companies to become Unicorns, we understand it to be a numbers game. That grace is given to entrepreneurs but scientists need to grovel for cash and endlessly show that their research is "translatable" or sufficiently impactful.
Sorry, I've heard this one too many times before. Thanks for your contribution to our world's knowledge, I hope you value it as much as I do.
This is exactly the problem - early on there was a lot of "low hanging fruit" in science - entire new areas where our tools and capabilities for discovery and analysis got way better very quickly. Think of everything that better telescopes, scanning electron microscopy, and computerization allowed.
Complaining that "Why doesn't progress go fast like before?!" when the newest tool-side improvement is a slightly faster CPU or a new clanker model.
I think there's this group of folks who are like "Why don't we have flying cars?" and eventually realize the problem is physics, but have to somehow blame people instead.
That's still an immense amount of code for a chat interface essentially consisting of a text box and a button, which any OS (mobile or desktop) can usually throw up in a few lines of code.
Academia is no different from any other profession or sport. Holding it to a higher bar than say, medicine, engineering, law or accounting, doesn't make sense.
As an example, let's take soccer: All players will tackle if they think they can get away with it. Even Messi, Ronaldo, Mbappe do it. Those who are caught receive a red card and are sent off the field. Do red cards stop tackles? No. Players just try hard not to get caught.
That same professor will happily take money from the student's startup to conduct research assuming it is successful and has funds to spare. That should tell you right there how the incentives are aligned.
If there's one problem that LLMs have solved, it's language. While an LLM may hallucinate, it does so in grammatically correct English sentences. Additionally, even the local version of gemma-4-26B can seamlessly switch between languages in the midst of a conversation while maintaining context. That's perhaps the most exciting part for me: We have a bonafide universal translator (that's Star Trek territory) and people seem more focused on its factual accuracy.
Language is not about just grammatically correct sentences, it’s about expression, intent, and communication that goes beyond the spoken, written or even motioned word—not one of these things is in the realm of possibility for current (and dare I say even future) AI.
Your Star Trek comparison is also incorrect. Following your logic, we’ve had a “bonafide universal translator” for a while now with websites like Google Translate (and so on). But none of these websites or AI tools are capable of learning languages on the fly purely from context and with minimal input data (that’s the magic of Trek’s UT, what they call linguacode).
Kind of a tangent I guess, but the coolest thing about Star Trek’s universal translator to me was that it could translate new languages mid-conversation with an extremely small amount of data. Makes me wonder how close we might be able to get to that eventually
I don't know. I think the translator functioned perfectly. It accurately provided a translation of all of the information that the alien intended to communicate. The translator can't translate things that are not intended to be communicated via spoken language.
Like...it doesn't work on cats. It doesn't translate smiles.
If "nafnfowmfowl grtakkan ssshelpik" means "Temba, his arms wide" then the translation was correct. The fact that the crew mostly faileld to understand the communication is on them, not the translator.
This can also be seen in Data. Just because he can speak all sorts of languages and understand a majority of communications, he doesn't really know Riker's love of Troi because his translation matrix isn't designed for that in the same way that the ship's isn't designed to interpret intent.
You don't even need to go to Darmok to see this limitation, really. The fact that the translator provides human translations of klingon speech is not enough to understand their cultural intricacies. The same could be said of "pon farr" being a "time of mating". That knowledge alone could not have prepared Kirk for the trial or even hinted at there being a trial.
The translator is neither mind reader nor cultural liason. If it were we wouldn't have needed Troi.
I’m curious what you think an abstraction is. Even running “ls” involves several layers of abstraction: a shell, a process (abstracts memory), a thread (abstracts CPU)… you think it would be simpler if you had to deal with all that to list a directory (another abstraction)? Even bits are an abstraction over analog voltage levels.
You're taking it out of context. I'm specifically referring to abstractions introduced in the codebase to maximize code reuse, as per the OP's comment.
I don't think these things are as different as you think. I started at "ls" and worked down. If you work up, you get things like a "socket", an "object" within a programming language, a "linked list" in a standard library, an "HTTP client" within an application-level package. You can keep going up and rattle off lots of useful abstractions in application-level code.
There are certainly _bad_ abstractions that ought not to exist, which I think is what you're getting at. There are poorly built abstractions, and leaky abstractions. But abstraction itself isn't the problem -- abstraction is what allows us to build anything at all without being crushed by the sheer complexity.
You're conflating system level abstractions with code-based abstractions. As a counter-example, introducing a factory constructor to handle object creation makes the codebase harder to understand.
the thing about abstractions is that nothing implies that they aren’t leaky abstractions, which may be worse than no abstraction for future bug hunters
We have too many PhDs (I say this as one). It's never been easier to get one. Most PhD topics are incremental and derivative whereas they should be seminal and ground-breaking.
Unfortunately, with credential inflation, this cycle will escalate. Soon people will complete two just to qualify for an academic position.
reply