Stasis5001

Oops, that's what I get for mathing before coffee-- mixed up the formula for \sum (1+r)^n vs. \sum r^n

It may be "free money" as you frame it. But a cash stream that provides n dollars per year forever can be valued in today's dollars, assuming a discount rate of d, at n / (1-d). So it's reasonable to prefer cash now to revenue forever, at that exchange rate, depending on your corporate interests.

https://www.investopedia.com/terms/p/present-value-annuity.a...

Tons of comments already in the direction of 'eat less' and 'exercise more'. Of course. One thing I've found extremely helpful is to cook more! WFH is amazing for this. If you can find 15 minutes of meeting time throughout the day where you aren't an active speaker, you can chop vegetables or do dishes for lunch and dinner.

Back to the primary topic: cooking your own meals makes 'eating less' a matter of dialing in the portion sizes over time. I eat ~110g of pasta, or 0.8 cups of rice, or similar, and about 0.3-0.5 lb of protein, and about 1-1.5 vegetables per meal. Overall, I eat two meals like this, black coffee, a breakfast bar, and a snack, which puts me at 2000-2500 cal.

Now, on to a weird tip: meal planning! What a fun mix of intuition and ideas from operations research.

When you say tech industry pays more, I'm assuming the effective long-run hourly earnings of poker players is way less glamorous than it looks from an outsider seeing highlights? Doing a basic google search led me to https://pokerdb.thehendonmob.com/ranking/6737/, which seems to suggest that 81 players earned over 1mm USD in 2021. But presumably those players don't earn that YoY, and even if they did, plenty more software engineers earn > 1mm USD per annum, so this seems to support your conclusion. OTOH, I have no insight to the home games or whatever sources are not included in these rankings, so I'd love to hear a little more.

It's interesting to compare this to the physical-world analog. What if somebody shows up at the US-Canada border, start shooting rockets at Canada, and the US refuses to acknowledge this as a crime or extradite? What if the rockets are aimed at a critical piece of infrastructure near the border and can cause billions in damage? One could argue that if the US condones these attacks, they have effectively already declared war on Canada.

See Matt Levine's 10 laws of insider trading. In particular, "5. Don’t do it by planting bombs at a company and shorting its stock." Somebody nicely put a non-paywall link here: https://github.com/0xNF/lawsofinsidertrading.com

Buying puts in this fashion will generally be a remarkable, traceable event. There's a reason ransoms typically go through cryptocurrency.

I started looking at the map in the article and realized I had just booked a flight where there was an existing rail line ! So I checked out the Amtrak site, and what's 1h20m by plane is 14h40m by train -- and 8h by car. Maybe by getting a sleeper cabin I could have had an enjoyable trip by train, but as the trip scales things get dramatically worse.

Of course, maybe this is exactly what the future of transportation should look like: more localized travel on modes that can be powered by renewable sources or nuclear.

As an employee at a megacorp, sure, you can see a little bit about how things are going from the inside. That said, you probably spend less time looking at reports than professional analysts. Secondly, sure, maybe you can see how great your company is, but do you have a frame of reference to a million other companies and their projects and cultures? There's a strong bias towards thinking you know more than the market that I would be wary of.

As somebody with some ML background but no expertise in crypto, is the following ELI~20 summary correct?

We take an ML model trained on unencrypted data, use a 'homomorphic evaluation' technique (let's just leave that as magic here) to convert the model operation-by-operation to a model that runs on encrypted data, do a little more crypto magic, and we've solved the business problem described at the beginning of the article.

(In particular, if you train a model on encrypted data you get a really bad model, right?)

Any good references on the art part, or is it just an intuition you develop over time? In my experience, all the ML education will teach you a ton of theory and basics but none of the practical details you're referring to.

Sure, Wiles obviously understands the main thrust of his proof. But one could argue that Wiles' result depends on lots of other results, which in turn depend on other results, and so on through decades and decades of work, ultimately going back to the foundations of mathematics. Neither Wiles nor anybody else can claim to rigorously understand all of it. You can imagine this as a tree, with Wiles' work as root, and his dependencies as ancestors, and so on. An error at a lower level of the tree could, in theory, invalidate the root node.

I do agree with Buzzard that it's hard to be sure. I've definitely read papers where a critical argument isn't well written or what is written seems wrong. However, if there are low-level errors, I suspect that with some work things could be patched up.

Tipping is a form of price discrimination, which generally makes things more efficient and has been a staple of capitalism since forever. A business actively wants to extract more money from customers who can afford more, and still sell to customers who are more price sensitive. Coupons are the more normal way to do this, but tipping serves this exact purpose. This is in theory good for customers (except in monopolies, where the business extracts all the value), but in practice, sure, I agree tipping is obnoxious as a customer and possibly even worse for those whose pay involves tips.

It does change something. DoorDash's old policy, weird as it is, reduces variability of daily pay for the Dashers. It literally put a derivative structure onto the Dasher pay that reduces volatility -- the pay was cash plus a call option on the tip. Now they will get cash plus the tip. Obviously DoorDash will tune their base pay, but they can't meet the prior model's mean and variance at the same time. How Dashers balance those two is unclear, but generally people need a higher mean to accept higher variance.

The problem is that the old policy only works if the customers don't understand it, and maybe even seems generally unethical.

I think you may be missing the point -- the one being named and maybe shamed in this post is actually Airbnb.

A lot of academic papers actually aren't all that great, for a variety of reasons. Normally, you can use citations, journal, and author credentials to get a sense of whether a paper is even worth skimming. The only "paper" on the "journal" right now looks like it's just a watered-down, html-only version of https://arxiv.org/abs/1612.07828!

Seems like more of a PR stunt than anything useful, but who knows.

This comparison isn't so great because of the following:

    for product_id in unique_products:
        product_items = [x for x in dataset_python if x[0]==product_id ]

This is O(unique_products * observations), and it looks like O(unique_products) = O(observations). Thus we have a quadratic scan when a linear one would suffice. You'll get the best performance using whichever solution lets you code this to linear the fastest. E.g. pure python, make a dict from product_id to observations and iterate over that, or for pandas, use groupby.

For reference, generally people learn all this over four years of undergrad and probably the first few years of grad school (I have a BS in physics, and a PhD in a different field, so not 100% sure on the grad school work). That's 6 years of more or less full-time work, surrounded by excellent peers and mentors, where every week you read 2-4 chapters and do 10 problems per chapter. If you're motivated and talented, you'll breeze through the first few years of problems, but anecdotally, everybody hits a wall where the problems start to get really hard. I have no doubt you can replicate the undergrad education through self-study, and maybe even save money and time, but after that point, why not just go to grad school? You won't save much time doing it on your own, and you get mentorship, exposure to the research aspect (not exactly trivial to learn), credibility, and funding.

I'm pretty sure the number of high-quality researchers in theoretical physics, or any major field, who are totally self-taught is really quite small. This website feels like it's in part to dissuade amateurs from sending their "awesome result" to professionals, which the author mentions in the intro.

Sure, and what about skittish Groupon employees?

Right. Airlines overbook because they make more money, and they are bound by laws like "If the substitute transportation is scheduled to get you to your destination more than two hours later (four hours internationally), or if the airline does not make any substitute travel arrangements for you, the compensation doubles (400% of your one-way fare, $1350 maximum)."

Thus, if expected revenue from overbooking > expected payouts, overbooking is worth it.

HOWEVER, the legal compensation required scales really weirdly - $1350 maximum for anything over a two hour delay? Why would being delayed three hours be equivalent to being delayed something like 12-18 hours? I'm also not sure that maximum number has scaled with inflation.

Anyway, some airlines already do auction off this. All airlines ask for volunteers because the compensation they offer is always less than the legal requirement for involuntary bumping.

This is just a mess because United had already let the passengers on the plane, and weren't even offering that maximum, much less over it. If somebody here actually knows contract of carriage laws, does this sort of 'required compensation' thing hold after boarding happened?

This study is suspect. First, using self-identification as high-performance is already troublesome. Second, the author never defines what an HPE is, and the most concrete specification is from this article they backlink to:

  >For example, in a software or tech company they would be:
  > Software Engineers
  > Data Scientists
  > Visual Designers
  > Structural & 3-D Prototype Designers
  > Narrative Developers

This certainly doesn't agree with what almost anybody else uses to define high-performance employees.

Third, the question they use to resolve this is from their survey:

  >Are you motivated more by mastery of skill/craft, or motivated more by earning money?
  > Mastery of skill or craft
  > Money
  > Other

I don't think we can draw any conclusion from the fact that 76% of the people who answered this public google form chose 'mastery'.

When you get a PhD, you mainly only take classes in your department for the whole time, whereas in undergrad you still have classes in other fields. This usually extends beyond coworkers into friend circles too. This isn't math specific, but more generally to grad school.

Second, programmers at companies almost never work in isolation from other programmers, and in most open office environments are close enough they could touch another one from their desk. On the other hand, a startup with five to ten engineers may be hiring you as their first data scientist, and bigger companies may be putting you on an embedded team, with the nearest data scientist a hallway or floor away. And this isn't a big deal in terms of teamwork, but most data scientists don't have a PhD in mathematics, so if you find higher level mathematical ideas and notation to be the most efficient way for you to think about a problem, your colleagues may not. That's also not a math specific thing though -- an economist, statistician, mathematician, computer scientist, physicist, etc. are all going to have slightly different ways they think about things internally.

With pizza, your solution can't work. You have to pay the driver base pay, and they get tips on top. So the base pay has to come from the menu price + delivery fee, and if you get rid of the delivery fee, it has to come from the menu price. But if you raise that, you're distorting demand by making in-restaurant purchases more expensive.

Every article on salaries seems to really screw up two things. One is, salary is only a portion of total compensation. Which do they mean here? If you don't look at equity packages, my intuition is you're going to think SF sucks, since I think more companies there use stock incentives. I'm not sure anybody has good data on equity packages/total comp, and at the very least, it's not trivial even if you have that data, since for example you can't just treat an Uber RSU the same as a Google one. So comparing total comp, which is what matters, is a subtle issue with lots of methodological decisions to make that will alter the results, and when articles (including the original source at hired.com) don't even mention any of that, it's hard to take the results seriously.

Second is, you can't just scale up salaries by the COL adjustment when the comp is this high. If you do that, you'll think the following are equally good:

  - monthly income 10k, 3k on rent, 2k other expenses
  - monthly income 5k, 1.5k on rent, 1k on other

Not all expenses scale linearly with COL, and in particular, one of the most important imputed expenses, savings, doesn't scale that way, unless you plan to retire with the exact same lifestyle.

A philosophy class teaching fundamentals in logic - taught in a way similar to math - is probably more like math in the sense of the article than philosophy in the sense of the grandparent post. The grandparent probably meant something more like a class on ethics, which doesn't really have the same style of beauty and truth that math does. This is certainly my impression after taking a few philosophy classes and a ton of math ones, including a philosophical logic class cross-listed with math.

This isn't a flaw of philosophy. Lots of fields of human study are great. Math just has some things that make it unique. But just like defining what art is, it's hard to pin down exactly what it all is.

Math classes ramp up in a gradual way that develops critical thinking and intuition in a very small sandbox, where it's easier to appreciate the results. Thinking in, say, just the x-y plane makes it far easier to isolate relationships than philosophy, which tackles much bigger problems and gets you into controversial problems almost immediately, with no clear answers. The basic math framework is rigid and precise, whereas philosophy only gets that rigidity at much higher levels where, guess what, it starts to merge with the field of mathematical logic.

Of course, a one size approach won't fit everybody.

Pytz handles a certain class of funkiness in TZ changes, but not the one mentioned here. Consider, a user in footopia enters a datetime 3 months from now which is saved in UTC in your database. Then footopia changes their TZ definition. The only way you could get this right is if you also know when the datetime in the database was created. You would need to tell your datetime library the datetime and _when it was created_, and pytz doesn't do that (I don't know any that do, actually).

Sure, I forgot the self-join case. In my experience these are a minority of joins, but definitely. In postgres at least, a lot of self-joins can be solved with a DISTINCT ON instead.

Sub-selection with a FROM clause, yeah, those definitely need names. But those don't have perfectly good names that everybody already knows, whereas tables do. If you're maintaining somebody's code and they declare

  foo_bar = 10

I think it'd be pretty strange to ever consider writing

  fb = foo_bar

just because you find foo_bar too long to write or something.

Weird, I really dislike a lot of the suggestions. In particular, an example:

  SELECT first_name AS fn
    FROM staff AS s1
    JOIN students AS s2
      ON s2.mentor_id = s1.staff_num;

We already agreed that staff is a good name for a table, so why are we renaming it to s1? There are all sorts of subtle bugs that arise when s2.mentor_id = s1.staff_num is wrong, and the variable names provide no help here that we're doing things right.

How about:

  SELECT first_name
    FROM staff
    JOIN students
      ON student.staff_id = staff.id;

If you're reading some 50+ line SQL query with a bug, things like this are easy to read and be confident in. Honestly, I think AS should basically never be used for tables.

Likewise, the guide gives the example of naming join tables 'services' instead of 'car_mechanics'. But if I see a table 'car', a table 'mechanic', and a table 'car_mechanic', I instantly know how to join against this in any direction. The hard part of understanding a schema is how it fits together, not what it represents, and when you have hundreds of tables and need to remember that join table's name, it's really, really nice if it's trivial to derive.

Lastly, the guide's recommended indentation for FROM vs JOIN statements seems off to me. Consider:

  FROM riders AS r
         INNER JOIN bikes AS b
         ON r.bike_vin_num = b.vin_num
            AND b.engines > 2

Well, the thing is, the riders table and bikes table have the exact same priority in the query, but this spacing really emphasizes the riders table and it's not clear why.

In this instance, Uber is making public transportation _more_ accessible. They are crowding out people driving from home to a parking lot.

Anyway, the article mentions this is maybe ten to a hundred rides a day. This is either a PR thing or a prototype; time will tell.

This article just rambles, and I'm not really seeing what point it's trying to make.

Point one, I think: something about inequality because DSC has 190 employees. But it still takes the same order of magnitude number of people to make and sell a razor; they just don't all work at DSC. There are still employees at the Korean manufacturing center, and it's not like they would have gotten rich if they were owned by DSC. How many people were made millionaires by Gillette?

Point two: big companies should be afraid of disruption. If Gillette was really afraid of DSC, well, we know how much it costs to buy them, and Gillette was worth 57x more, 11 years ago. I think a more likely explanation is Gillette had almost a monopoly but didn't find a way to use price discrimination, and DSC picked up where they left off.