b9a2cab5

Sounds like you screwed up. Current private student loan rates are on the order of 3-5%. Refi's are less than that. And that's ignoring the newfangled stuff like ISAs where there aren't even any loans to begin with.

> I don't want people choosing art history over CS [because of price]

This reasoning right here is why people are stuck in $100k student loans with useless degrees. Sometimes it is more useful for society as a whole to have fewer art history majors and more engineering majors. The overwhelming majority of highly impactful social science research (if it can be called that with the reproducibility issues it has) comes from the top 10 institutions. Everyone else is just getting a degree in barista science.

If you follow Ray Dalio, he hypothesizes a lot of what we're seeing is due to decadence. Specifically, newer generations expect a high living standard but are not actually as rich as prior generations to support that living standard. At the same time, the population spends on decadent luxury goods instead of productive goods (which you can see in deindustrialization and the rapid expansion of low quality goods on Amazon).

If you look at the history of the Roman Empire you see a lot of similarities. Increasing living standards means people aren't willing to work as hard as their predecessors because they already have it pretty good. That leads to a loss of competitiveness, loss of competent political leadership (because why struggle in politics when you don't have to?), and increasing corruption. At the same time, people continue to purchase luxury goods to maintain their living standard rather than re-invest in productive assets that might decrease their living standard.

Instead they invest in scientific R&D in areas they care about (i.e. anti-aging, Alzheimer's research, long-shot AI, space, etc)

Arguably this is better for humanity than building another museum that's going to decay 50 years from now.

If you're going to make an argument for why we should tax companies like Amazon more, you need to also make an argument for why that money wouldn't otherwise be better used by Amazon compared to the government.

In my opinion anything larger than a medium sized city (think one of the many cities in SoCal) is basically guaranteed to use its funds poorly.

The post office is increasingly losing business to low cost shippers and high margin business to same-day delivery tech startups. Not only that, their trucks are older than some drivers and the contract to replace them was filled with government pork.

You're never going to get innovation from people who are just trying to work a job for a wage and pension. That's why Amazon is disrupting everyone with their logistics network.

Perhaps you should remove your political colored glasses and consider things objectively. A company can be both remarkable in its achievements and have committed labor abuses.

> doesn't pay taxes

Yes, that's what R&D investment is. They "make" billions but reinvest all of it into growth so they pay no tax. If you somehow split off AWS and taxed its profits separately Amazon would just increase its R&D spending on AWS and go back to paying $0 in tax.

This is a good policy, by the way. You want to encourage companies to make use of their money rather than returning it to shareholders. That creates more jobs and results in more technological and economic growth.

The most potent competitor to Starbucks (Luckin' Coffee) took a huge hit due to accounting scandals just recently. More generally Starbucks in China is considered a luxury brand (not what it is in the US).

Anything manufacturing related I can see being shifted overseas. US labor is too costly compared to international labor. I think Deere and co already see the writing on the wall and are moving to diversify with autonomous capabilities. Software is not so easy to copy or we would've seen an iOS clone out of China that actually works as well as iOS.

You make fat profits in normal years and potentially lose out catastrophically during Great Recesssion-like events. Lots of lenders straight up went bankrupt due to defaults.

Of course since the government made it so you can't discharge student loans in bankruptcy that market isn't functioning anymore.

You don't need to be in the top 10%, you just need to be likely to be able to repay your loans.

Government subsidizes home loans, student loans, and small business loans to the tune of hundreds of billions of dollars a year through the Fed's asset purchase program. The reason broke people can't get anything better than a 10%+ APR credit card is because they're so unlikely to repay the loan that the issuer has to charge that much interest to break even.

I actually took a look at my banking app and I'm offered "personal loans" up to $20k at under 6%, which is actually incredible considering this is an unsecured loan to a random person.

Good IT and engineering staff is a lot more expensive than even a fine for 10% of your annual revenue. Hiring a single good security engineer might cost 400k or more.

Good luck getting your insurance company that pays $100k to senior engineers to do that.

Being able to see exactly which columns are partitioned on so I know whether my query is going to be efficient or require full scan+materialization. I understand from your link Snowflake has clustering keys but it seems like this only "suggests" that Snowflake should partition on that key. Also with Spark I can look at the query plan while it's executing and figure out what part of my query is bottlenecking. Seems like Snowflake requires a query to complete for me to do that.

Intel has stopped disclosing a lot of details on their newer products, probably because they're no longer far and away the market leader. I think if AWS ever develops a 4-5 year lead over everyone else we'll see similar disclosures out of them. Facebook publishes a lot of info about Oculus asynchronous reprojection techniques and computer vision because they have a 2/3 marketshare in VR.

Unfortunately those claims don't translate to servers, because the IO die's power usage increased and perf/w isn't much better [1]. Do the math and you get around 14% gain in SPEC MT workloads.

[1]: https://www.anandtech.com/show/16778/amd-epyc-milan-review-p...

You can buy Ampere Arm CPUs that are based on the same core as AWS Graviton2. And in general mobile CPUs are a generation ahead of server architecturally.

On a per thread basis - yes. On a per dollar basis on AWS - no. Per instance basis - depends a lot.

I've had terrible experiences with Snowflake "automatic" optimizations. You have zero visibility or control over partitioning, join strategy, or anything else compared to Spark/Databricks.

You are free to tip until they make whatever you consider a sufficient wage. "We" should start with "you".

I hope you realize this is already the case. Property taxes are voted on locally and go to pay for local schools and there is vehement opposition towards redistributing funding for schools state wide. People move to certain cities where schools are known to be good (and no surprise, those cities also have lots of generous property tax-based and PTA-based funding for schools).

I've tried it and half the time it's like "god damn this is the future" and the other half it's just repeating garbage in a syntactically correct way. It does particularly poorly for unstructured tasks like writing new functionality or classes.

Last 3 categories can afford it if they get a job. Single payer proponents claim on the order of 10x cost savings compared to the status quo. Presumably insurance costs wouldn't exceed $200/mo under single payer. Amazon pays $15/h minimum nationally. That's a day and a half of work for health insurance for the month.

The first 2 categories need better help than healthcare insurance. I think it's plainly obvious after SF spending $50k/homeless person that more money isn't going to help them.

Medicine is an entirely different problem. Salaries for doctors in the UK and Europe are much lower in part because the AMA limits the number of residencies and thereby the supply of new doctors in the U.S. That's de facto a monopoly.

If you're a top 1% pediatrician you don't get paid much more than a median pediatrician. That's because it's pretty hard (impossible?) to be a 10x pediatrician. On the other hand, you can make the argument that 10x neurosurgeons or heart surgeons exist.

Socialized healthcare might be much more amenable if it were single-payer but without subsidies. You pay for insurance just like you do now, except there's only 1 provider and it's the government. The primary benefit most M4A people tout is that costs would be lower under a single payer system. This makes that happen without the welfare leeching.

I and I'm sure many others are opposed to subsidizing people who decided to play video games and smoke weed all day with our tax dollars. If you're an entrepreneur, you're not harmed by this, you can afford to pay for it. If you're a small business owner, this should save you money. If you're disabled or old, you already get subsidized and would continue to get subsidized. Everyone else - pay for the services you receive.

Plenty of places in the Midwest/South if you're a tradesman. Obviously a McDonald's worker or Starbucks barista is probably not going to be able to afford a mortgage.

> You may find that some of us on the left are trying to call out this bullshit too

You should take a look at this video by John Oliver [1]. I find it hilarious how 9 years and 2 presidents later it is still relevant. You refuse to admit that _you're the one spewing the bullshit_. Any thought that you could be wrong is just inconceivable to you.

[1]: https://www.cc.com/video/slfmox/the-daily-show-with-jon-stew...

It's ultimately a result of increasing returns to scale from technology. If the top 1% of every class gets you marginally better competitiveness in the market as a firm and you can feasibly capture 100% of the market (or a large majority) across billions of customers with a few thousand employees then you're going to pay for the top 1%.

On top of that, increasing adoption of technology means that more of the economy becomes driven by engineers. Computer science is one of the few fields where a top 1% individual can actually be 10x more productive than the median.

Combine these two things together - being able to instantaneously scale improvements across billions of customers for almost free because of the Internet, and technology's growth resulting in more jobs having a wider productivity gap between median and top 1% - and you get massive salaries for the top 1% and comparative peanuts for everyone else.

The exact same dynamic plays out in finance, Hollywood, YouTube content creation, and other winner takes all markets.

Asset prices are merely a function of discounted future cash flows. As interest rates are lower, the discount on future cash flows decreases. The world is also much more stable than it was in the 1950s-1980s - look at the state of Asia now and then. That means there is a lower risk discount on future cash flows.

If paper wealth takes a shit the people that will be impacted will be retirees and the wealthy, not your average joes.

It only happens when there is not robust competition (even 5-6 big players is not enough), which is the case in pretty much every non-commodity market today.

You may be doing fine, but objective measures of the average salary of humanities majors at schools like UCB say otherwise.

Most of the PMs I know have either business or STEM degrees.

If you aren't willing to put in the effort for a multiple choice (easy) exam, you are failing any STEM college degree (the only ones that give you a chance at making it out of the poorhouse), straight up. Compared to the SAT intro calculus at any top university is magnitudes harder.

If you admit the kids you're talking about under affirmative action they'd just be weeded out in intro courses and get relegated to some social sciences degree with negative return on tuition. That is objectively worse than rejecting them.