throwawayjava

Image processing, text processing, and the next wave of robotics/manufacturing. A.k.a., "AI".

Public labs are nice for society because they're extraordinary cheap relative to their corporate counter-parts (grad student = 30k/yr with no benefits, research scientist = 200+k/yr with full benefits)

Or to put it another way, to the extent that we interpret local and federal and so on formally enough that they retain any connection to information theory, we're no longer talking about politics.

The words either have their political meaning or their formal meaning, but pseudomathematical sophistry that conflates formal and informal statements in the same breath leads to quackery and nonsense, which is exactly one of the article's critiques of libertarianism as a framework.

The CAP theorem occupies a similar place in pseudomathematics as Goedels theorems. An important result, but almost never for the reasons that someone attempting to use it to make a non-rigorous mathematical argument thinks it is...

For example, sections 2.1 and 2.2 resonated with me because, as a mathematician, because I can plainly see how the algebraic structure chosen by Rothbard for representing utility is going to severely impact the reasonableness of his models.

There are several other compelling examples, where anyone who has done a bit of mathematical modeling can see that the neoclassical school of thought is simply providing better tools for achieving the desired outcomes.

He's not.

To take your example, Section 3.1 critiques The Use of Knowledge in Society via a critique of Human Action (which was published by Mises nearly a decade after 1942). Many of the observations in Section 3 (and 3.1 especially) are directly relevant to The Use of Knowledge in Society.

This is certainly the major difference.

But yes, there are a lot of jobs out there that even Stanford CS graduates aren't qualified for.

Note: the patent bar's eligibility rules have special exemptions for CS degrees that are not ABET accredited but do satisfy other criteria, so many Stanford CS graduates will automatically qualify for the patent bar based upon those alternative standards. But that doesn't mean the patent bar doesn't care about accreditation.

And it's worth noting that any institution with a half-decent reputation has some form of accreditation. Again, this was the second point of my post -- accreditation is actually does do a very good job at over-approximating the set of educational institutions whose degrees mean something.

Was this with room and board? FWIW, there are many college degrees in the USA that are in the $7k/yr range, and many of then were around $20k total prior to the last round of state-level austerity in the USA.

Getting a US college degree for under $20k is still very possible if you a) commute and b) do the first two years at a community college (typically closer to $2k/yr, or $18k-$21k for 2 years cc + 2 years state college).

With a bit more public subsidy (or differently targeted subsidy), many US colleges could be well within the $20k USD range for commuting students. And some courses of study already are.

This is false because many fields do care very much. E.g. medicine (and not just mds), law, the patent bar, and many engineering fields care about accreditation. Even degrees from highly acclaimed universities aren't enough if the degree isn't accredted by the right body (eg ABET for the patent bar).

This is also not relevant because, in practice, places with good reputations are typically accredited.

Note, however, that you won't get e.g. OPT from an online degree.

Like it or not, in the USA cash is king. It is what people optimize for.

If you FIX a highly-rewarded measure for a nontrivial time period, the "take best comers" approach kicks ass. As long as your fix apriori and them measure. That's what we see here. GDP was defined before current immigration policy was set, after all.

If we have faults, they are in our system for doling out rewards. Not in the color/creed/religion of the people we let compete for those rewards.

If you want better social cohesion, design social systems to REWARD that cohesion. Then, communities will offer what you seek. To your great surprise, I suspect, those communities will be highly diverse.

Otherwise, you end up with sub-populations that severely under-perform on actually measured metrics and go out to search for bullshit alternative (under/arbitrarily)measured metrics by which they win. Again, that's what we see in your comment.

I want to suggest that perhaps those students in CA really were that much better. Not in terms of raw talent, but in terms of the previous 10+ years of their life.

I learned a lot in PhD school, but the most important thing I learned was that others' high school experiences really did prepare them for academic rigor and business success in ways that mine decidedly didn't.

XKCD isn't criticizing the positive mindset of naive-but-gifted researchers; it's criticizing the attitude of experts in one field that their expertise necessarily gives them unique and powerful insights into solving problems in another field.

Perhaps a better counter-anecdote would be all of the mathematicians in the 1950s - 1980s who solved hard program design problems with little or no experience actually programming computers. That really was a counter-example to the XKCD mentality: a bit of mathematics was far more useful than knowing the instruction set of the latest IBM mainframe...

In particular, the quote is not meant as some sort of claim about the metaphysics of cancer or anything ridiculous like that. Rather, it's a (perhaps not so eloquently stated) way of pointing out that the mathematical and biological models underlying the current (and again, concrete) moonshot attempts fail to account for certain important issues and that this strongly suggests that those models are not going to result in cures.

The rest of the article summarizes some of Brivik's warrants, which are well thought-out. The claim is that the models currently used are fundamentally incapable of explaining how certain environmental factors effect cancer and the rest of the body, and that these environmental factors are important for understanding how to "solve" cancer because cancer evolves over time in often unexpected ways. The article points out, for example, that there are modeling assumptions baked into Microsoft's approach that are empirically falsifiable.

Finally, I'll just re-iterate that the argument made in the article is made in the context of current proposed/ongoing moonshots, so various comments about nanobots at the like are totally off-topic.

The article is not about science fiction. The topic is not whether maybe one day in the far future we might have enough knowledge about cancer cells to take a purely mechanistic approach toward curing cancer. Whether, the article is about the allocation problem in real science, today, and makes an argument against the efficacy of the approach taken by some current, concrete, well-funded moonshots.

[1] https://www.med.uio.no/imb/english/people/aca/jbreivik/

That seems like an ineffective mechanism. Just litter and pick it up in the very rare cases where you get called out.

If I were council dictator for a day, I would make the fines stiff and immediate but require that 100%^1 of the revenues are refunded to all residents at the end of each year.

That way there's still a stiff disincentive to littering, but no significant incentive for abusive/dishonest enforcement.

[^1]: (perhaps sans some low and more importantly fixed amount to cover enforcement costs)

Surgical robots -- at least the ones that actually exist and work today -- are designed more like traditional safety-critical robotics platforms and less like a one-off case study for a nice visualization/table for an end-to-end deep RL paper. There is no "the model".

> $90k (9 month salary) is the bottom 10th percentile for a starting assistant professor in CS.

That's only because your data includes professors at R1s and also not including instructors at R1s. This is doubly misleading, since those professors you are including spend most of their time on research and those instructors you aren't including are the ones doing the bulk of the teaching.

And outside of R1, 90k is damn good for ast professors.

If you are teaching for a living, as opposed to doing research, $90k is in the upper 10th percentile at the asst prof rank, and $100k is a ceiling for tippy-top candidates at the asst prof rank.

> $200k (9 month) is the 90th percentile for a full professor with 16+ years.

...and is also bottom 10th percentile for wet-behind-the-top- ears top 10 PhDs, who don't have 16+ years experience on top of their 5+ year phd...

> This obviously varies by school and region, but $60k-$70k is what you would expect for an instructor at a very low rank school, which is not a position PhDs from top universities would be applying for.

Sure, but outside of top 10 $150k is still pretty typical industry salary for phds.

The same problem exists at every scale.

Your assertion is also fairly confusing; lots of smart people decide to teach at places where they can have the greatest impact, which is typically not a large-endowment private school...

A lot of ideas are obvious even without hindsight (again, especially in applied mathematics, and again, maybe that's just my bias). None-the-less, the person who said it first gets their name on the important idea.

There are lots of "<person's name> operator" and "<person's name> theorem" in big-but-not-core fields of (again, esp applied) mathematics where this phenomenon occurs.

I've been party to a lot of conversations that go something like "<person's name> idea or <person's name> technique is just '<fairly basic thing> applied to <fairly commonplace thing>' and it's stupid that it's named after <person's name>".

I always remind the speaker that this is good news -- one route to science fame is to just state the obvious loudly enough and early enough ;-)

(Even at 400K, "highly paid administrators" doesn't pass the smell test.)

I'm talking about places with words like "south eastern" and "central" and "state" in their names. And SLACs with endowments below ~100M. In other words, I'm talking about the vast majority of universities and colleges -- the ones that aren't (literally) Harvard or MIT. Places that don't have phds programs, and maybe don't even have sizable masters populations.

Education is not about what works for outliers.

There are literally hundreds of liberal arts/teaching colleges...

Also, Colgate definitely isn't in the tuition-dependent, take-all-comers mold of the vast majority of teaching-oriented colleges and universities in the USA. Their acceptance rate is below 30%, whereas most teaching-oriented colleges take 70+% of applicants (and draw from a much less well-prepared population of applicants to begin with).

Mathematics is as close to "humanities" as the STEMs get in terms of compensation and employment options for PhDs, and even top 10 mathematics PhDs can make 2x in the private sector with at least 2 years fewer opportunity costs.

Basically the only fields where what you're saying is true are the humanities and social sciences, but those aren't expanding programs or cost centers.

Work-life balance is one of those things that you would think would better in academia, but is often actually not.

The broader point is that outside of R1, prices mostly reflect the fact that for most students education doesn't scale very well and experienced human labor is more expensive than ever. Those are self-reinforcing, for obvious reasons. Case-in-point: the would-be educator who can't afford a 200k pay cut because of the loans they took out for their own education.

...and look at the instructor's salaries. PhD's from top universities who could easily make 300k+ in the private sector but instead take positions that pay 60k-70k. Even with tenure, that's a labor of love.

Blaming highly paid admins would be great, except that university presidents at teaching colleges are the highest paid (reminder: branch campuses and liberal arts colleges never have Div 1 sports programs) and still make $50k-100K less than a green phd out of a top 10 CS program (200k is good money for a president of a branch campus or liberal arts college -- and remember, that's typically a person with 30+ years experience and a uniquely positive track record).

I worry about what will happen to the American teaching college (public or private) over the next 50 years if the current war on college -- which this article typifies -- wins out in the public mind.

If you doubt a single word of this feel free to reply to this comment. I'm extremely familiar with both markets. When I say that (even tenure-track) CS teaching requires a 200k paycut, I really do know what I'm taking about.

Tuition-dependent colleges and universities tend to look very different from their large endowment counterparts.

Mid-sized USA cities tend to have half decent transit around their urban core (often woefully under-served especially in cities that don't have either trains/subways or else dedicated right-of-way for buses, but none-the-less serviceable).

However, the thing that mid-sized cities almost uniformly lack in the US is any form of regional commuter rail.