blip54321

I think you're missing the big picture.

MIT's best research was 1960-2000. At the end of that period, the endowment was around 1/10th of what it is now. That's a very healthy level of funding. $2M per professor is enough that it doesn't need to worry about money, but that there aren't obscene levels flowing into odd pockets in odd ways.

Once an institution is overcapitalized like that, dynamics change. There's a ton of corruption in 2022-era MIT (and peer institutions), conflicts-of-interest, and the insane level of competition leads to a lot of unethical behavior. There's so much money swishing around that it's tempting to skim off the top. Combine that with the traditional MIT rule-breaking ethic, and you've got trouble.

I would love for MIT to spend 90% of its money, and be back to 2000-era endowment. I would even more love for MIT to expand it's faculty and student bodies 10-fold, and give more people an opportunity to be there.

What I hate is MIT now being brand-obsessed, money-focused, hyper-competitive, and valuing high-impact and high-profit research over high-integrity research.

MIT has an endowment of $20M per professor.

We spend plenty on research, just inequitably, and with bizarre pyramid schemes.

Culture is not the same as government.

This looks awesome.

I'd pay $50 to have this workflow fully integrated and working with my editor (emacs; I'd also take VSCode or something else that's not vim). apt-get install a few things, and it works.

That's conveying value-add, not suggesting someone should do it for that amount (which is to say I'd gladly pay, but I can't imagine it's worth doing for that amount, or that there's a reasonably way to find and collect $$$ from more people like me).

The Internet Archive is in a different class than EFF or ACLU. I would (and did) donate to the ACLU and EFF of decades ago. I wouldn't to the organizations of today.

There's a natural process by which organizations revert to the mean. Support the exceptional ones while they're exceptional.

My 5-minute version:

- Get the overall system structure in Python. Get the architectural design right, and the big-O stuff right.

- If there are bottlenecks, re-code those directly in C, cython, or similar. Or better yet, find libraries.

Python is great for expressing high-level operations and system design. It is also very easy to integrate with native code. I've never had much happiness in optimizing Python itself beyond that. Broadly speaking, code falls into three categories:

- Most code: Instant. Performance doesn't matter. Python

- Some code: Big-O(lifespan of the universe). Not worth building.

- Narrow slice of stuff in between: Don't do in Python, but use Python as the glue.

> Is there anything you do differently, relative to in person?

Online, everything needs to be done differently from in-person:

- Lecture is a waste of time. Pre-made stuff will be higher-quality than anything you can deliver, and save a lot of time

- Grading should be automated as much as possible. Immediate feedback is powerful for students, as is being able to move at their own pace, have targeted remediation, etc.

- Automating stuff leaves waaaaay more time for 1:1 interaction, reviewing student work, etc.

- Audio isn't the only way to interact. Students can use chat, embedded surveys, forums, etc. There are many ways for having interactive engagement impossible in person.

- You should make heavy use of peer teaching: Breakout rooms, structured peer review, etc.

Online during COVID19 crashed-and-burned since people took in-person and tried to run it over Zoom. Good online can be better than in-person.

Oh, and details matter. Everyone should have a headset. Mute is generally bad. You should have good whiteboarding tools. Everyone should have a pen tablet ($40) or copy stand ($100). Etc.

NONE of this happened at most schools during COVID.

AoPS Alcumus.

Beyond that, OCW is pretty good. Find homework assignments from good courses.

As a footnote, the linked-to article is mostly nonsense. If you want to learn math:

- Work on good problems

- Once you've struggled for a bit, go back and look up how to do them

Yes, this includes problems you don't have prerequisites for and have no chance of solving. You shouldn't start by watching videos or explainers until you've struggled. There's a pile of science behind this.

It's a little bit different.

Most of the ultra-wealthy live on debt. Here's the basic scheme:

- Let's say I own $10B in startup.com stock, but haven't sold it. My income tax burden is $0.

- If I sell the stock, and taxes are 40% in my country, I get $6B, and the government gets $4B.

- If I keep the stock, and borrow against it, I pay no taxes, but I do pay finance fees (e.g. interest). The finance fees work out to less than taxes.

The trick is to never realize gains or income. Debt lets you do that.

One more trick is that a lot of purchases are "investments." If I buy a mansion, yacht, and Picasso on debt I can use them. When I'm bored of them, I sell them, and pay off that debt. I've probably made money on the mansion and Picasso, and lost money on the yacht.

I've paid a little bit of interest and a little bit of tax, but much less than I would have otherwise.

(1) Societal impacts sometimes are

(2) People not actually wanting to do them sometimes is a good reason too (e.g. free market forces, addiction, peer pressure)

China after the opium wars is a good example of how this can play out.

Is it the role of society to stop people from taking risks they willingly accept? And at what point?

If I want to skydive? Kayak across the Atlantic? Run a fight club? Build a custom homestead? Smoke? Drink? Do drugs? Play Russian Roulette?

Where's the line? There's one somewhere. To me, digging tunnels feels like the sort of thing which we should allow. Some of the others, I'm okay banning.

It's better. I hate Office 365 usability with a passion, and I hate the bugs, but it's much better than periodically having a fire as Google burns your business to the ground.

Microsoft deals with businesses, and it works.

It's very hard for my explain, within my company, that "failure" isn't the opposite if "success."

... if you're not failing some of the time, you're not trying things ambitious enough ...

I do read carefully. The truth is in the middle. Vaccines don't prevent spread, as out establishment kept insisting to get everyone vaccinated. They do decrease spread.

Managed rationally, this means:

1) Vaccines aren't a reason to reduce other precautions, if you're worried about spread.

2) They are an important part of the solution. Being outdoors doesn't stop spread. High-filtration masks, used by untrained people, don't stop spread. Vaccines don't stop spread. Combine them? R0<<1, and we're all okay.

Most studies show brain damage.

If I wanted to be the central kingdom, with everyone sending me tribute, making everyone else dumber would probably accomplish that. And it'd be good payback for opium too.

I don't think this was a bioattack, but a more likely scenario is a leak of a bioweapon still being developed. I don't consider that even unlikely. If I assign that scenario, or one like it, even 10% odds, extreme precautions still make sense.

We're optimizing to short-term economic damage. I suspect if COVID rolls through a dozen times, and gives just a few percent of the population brain fog on each pass, the long-term economic damage will be astronomical.

Yeah, that argument makes no sense. The FSF is way smarter than that.

The nuance is a little bit more complex. Let's say I run an AGPL+CC project:

- I become a licensee of code under AGPL if I bring in AGPL code from an external project. Downstream can remove my clause.

- I don't become a licensee if I have an adequately strong CLA in place, and only accept contributions under the CLA.

- I become a licensee of the code if you make a PR to my project without a CLA. However, you're contributing code under AGPL+CC. Downstream cannot remove my clause.

That's my read on both the court ruling and the intent of the AGPL.

It'd be crazy for a court to rule that you can't modify a license. Once you have AGPL+CC, that's a new license (and it's not open source or compatible with most other sharealike licenses).

No, I think you're misreading the case. I think the intent of the clause, which the court reaffirmed, is:

- You develop didibusDB

- I enhance it. I add a clause. (note: as an enhancer, I'm the licensor)

- Alice grabs it from me. She can just remove the clause.

This clause avoids litigation for step 3. Otherwise, for Alice to do this, YOU would have to sue. This way, the power rests with Alice.

For the "original licensor" case, the license referred to here would have been AGPL+CC. I'm not sure of a legal argument to say I can't distribute something under a modified license. The only error was false advertising -- referring to it as an AGPL or open source or whatever other license, which the court found to be true.

It's complex. There are two sides to every equation. Buyer and seller. Sellers aren't necessarily sophisticated either. If I sell my home, and the roof is leaking and I don't know about it, I don't want to be liable down-the-road.

Peace-of-mind is worth a lot. Especially if the buyer is sophisticated, I might take a lower offer if it has the right waivers.

.... and for most of those, the exact same arguments apply.

Stanley Motta is an electronics store. Both men and women work there and buy electronics. AIFM is an energy company. Ditto. Spanx is the only one on your short list where a supermajority of stakeholders are women. Even with Spanx, one could argue that slimming undergarments are for couples, where having one male voice in the room could help make sensible decisions. Seems like for those oddball cases, it's mostly an upside to diversify representation, even if it's in the opposite direction.

I see a few problems:

- Boards mostly come from an old boys network. I have not seen board members who were exceptional individuals so much as well-connected individuals. Even "diversity hires" -- and they're called that -- seem to pick from a tiny number of individuals who managed to break into that network, rather than individuals from minority communities. It's not a meritocracy.

- Boards don't represent the customers or employees, and often do things which harm some groups. Having all voices in that room helps.

- And so on...

Those sorts of arguments align doing good, doing well, and reducing (legal) corruption.

If I wanted to improve the law, I'd try to come up with ways to have representation from African American inner city communities, poor Southern rural communities, immigrant communities, and a whole slew of other groups who are traditionally underrepresented. I'll emphasizes communities -- people who are aware of culture and day-to-day life.

"Good" laws also do less tokenizing, and are less likely to become obsolete:

- The argument that a woman is on a board because... she's a woman is pretty stigmatizing. The argument that diverse perspectives maximize shareholder value is not.

- If, in 2050, we have 50/50 representation, or a minority group moves into the majority, balanced laws self-correct.

No, they don't accomplish the same thing. One accomplishes something better than the other.

To your points, all-female organizations would not be criminalized:

1) First of all, these should be civil laws, not criminal laws. Having this be a criminal law would be insane.

2) Second of all, these are about corporate boards for publicly-traded corporations. All-female associations are permitted. Even all-female publicly-traded corporations would be permitted, although I'm unaware of any. Board members are independent individuals who provide oversight. Most aren't insiders. It's fine to have a female board member for an all-male organization, or vice-versa. To your nursing example, if there were a publicly-traded nursing corporation, it's even helpful since half of patients are males.

3) Third of all, LA Times is not all-female; I'm not sure where that comes from. LA County is a government, not a publicly-traded corporation, and the law doesn't apply.

The right idea which should be behind the law -- which is sensible -- is that corporate oversight should understand the impact on everyone, and should represent the workforce, customers, etc. It's sensible. I suspect it's good for shareholder value too.

This is what happens what idiots write laws.

Good law: Boards cannot be uni-gender.

Bad law: Each board must have at least one woman.

They both accomplish the same thing, but one discriminates on sex, and the other doesn't.

Another example are laws targeting blacks. There are historical wrongs to be righted. Laws should, for example, convey benefits to individuals whose ancestors were slaves, were denied rights under the GI bill, or otherwise were specifically unfairly harmed. That's different from a law looking at skin color directly. It might seem subtle -- it's the same communities -- but one of those does something for bad reasons (and is likely to be found illegal), and one of those does something for good reasons (and is likely to be fully legal).

It matters. I support the court ruling. I also support the intent of the law. The legislature ought to go back to the drawing board, and write a law which does the same thing the right way.

Yeah. Big part of this is cultural. US is "If you can't say anything nice, don't say anything at all."

Many other parts of the world value honor and honesty more. I'm from one of those parts. In this case, it's a prisoner's dilemma. Honesty doesn't benefit you, but if everyone's honest, it benefits everyone. Exit interview, it doesn't hurt too much either. It's a safe time to be honest.

Having been on the other side of the table, if your manager is professional -- and most managers are -- it's a numbers game. The odds of an exit interview having an impact are close to zero. It's one person's issue. That doesn't make it invalid -- if you have a 30 minute longer commute to the new office, hate the way we do performance evaluations, or want growth paths into product management, you've got a valid reason to leave.

If five people hate the new location, how we do performance reviews, or don't think they have good internal growth paths, I'll act on that.

I don't think that's a good conclusion. Most of Google's business is search and ad sense. HN consensus is that Google is doing fine. The HN consensus -- which I think is correct -- is not to rely on Google for any critical functionality which might put your business (or personal life) at risk.

- Adsense as an advertising option makes complete sense. Adsense as your sole source of customer growth is asking for trouble.

- Stay the !@#$% away from Google Workspace, if you're counting on business continuity. It's fine for low-stakes use (e.g. students in a school, where there is no long-term continuity).

- GCE is fine as an option if you have a generic dev-ops infrastructure where you just need cloud servers, and where you can move to AWS or Azure in a blink. Relying on Google Cloud proprietary services is asking for trouble.

- Android App Store is fine for a quick buck, or as a supplement to a core business. Think long and hard before building a business around it, though.

... and so on.

... high-level contacts into Google's upper tier sometimes can't even help.

The two problems are closely connected. If one were to audit the elite where I went, one would find many scandals like this, albeit more cleverly disguised and perhaps legal. Good questions to ask:

- How many state school faculty come from elites?

- Why do we structure education as an elitist competition for a brand stamp? Why should someone have an advantage in life by having "Yale" stamped on their CV?

Many countries have roughly equal funding to all school, and similar admissions standards. Universities are nice in that as a freshman, you can take a freshman course or a grad-level course. What's the upside of sorting kids by background and sending them to different universities?

All schools also strive to move up US News and be like Harvard/Yale. If Yale jumps off a bridge, the state schools WILL almost always follow. The impact of what happens at the elites is disproportional.

A lot. It'd be worth addressing. We see growing evidence of this.

Many are new too. Viral dynamics... changed with the rise of cities a few thousand years back.

My experience is that there are a lot of synergies to be leveraged. It's almost ridiculously easy to start a successful startup in a place like Nigeria or Ghana which either:

- Providing things Nigeria excels at to Western markets (e.g. the same types of call center models India started out with a few decades back).

- Launch services in Nigeria common in Western markets but not yet there.

- Integrate Western and African models

Those sorts of things tend to benefit everyone involved on both sides of the pond. Compared to, for example, making web frameworks, there is a lot more low-hanging fruit. Silicon Valley has an almost unlimited supply of people wanting to do startups for people like themselves. This leaves many other markets under-tapped.

In terms of fostering local entrepreneurs, that happens naturally. Look at China, Taiwan, Korea, or Japan. They started with this model, at which point locals noticed they could do the same better and cheaper themselves, while retaining more profit. India is about half-way there now. It's very hard to foster entrepreneurship by telling. It's very easy by modeling.

China, Korea, Taiwan, India, and Japan never turned into Western colonies, as you're predicting.

My question was a little bit deeper. I understand why we have elite schools.

Harvard's an easy one. Rich and powerful people look smart by association with smart people. Smart people gain access to powerful social networks.

My question was why this was a good idea for society. MIT was funded through federal grants, financial aid, and philanthropic donations. It's a 501(c)3. If it's bad for the world, we ought to have some mechanism to fix it.

We don't, and that's a deeper discussion about how we want to organize society. We ought to. In the limit case of a self-perpetuating endowment harming the world, there ought to be some way for that to be repurposed to help the world.