_dps

An acquaintance of mine was working on an accounting PhD on basically this subject (trying to quantify CEO contribution to revenues and profits, all else held equal). He said that after years studying it, he thought it's too noisy with too little data over too long a timeframe to do a good quantitative analysis.

He also pointed out there is a big "apples to oranges" problem even with the limited data available. In sports (but not business) you have lots of reliably tracked micro metrics in addition to final metrics (wins/revenue). The micro metrics are generated regularly (stats for each player for each game) and those metrics are portable across players (everyone is going for points, runs, blocks, whatever). In comparison, for CEOs you mainly have quarterly reporting of coarse financials (but nothing like employee morale, number of offices opened, etc.).

There is a level of this that is reasonable, which is that what is presented in the writing be factual and not sensationalized. I imagine The Dispatch believes they meet this bar (I have no idea, I just heard of them now).

But there is a level of this that is impossible, which is avoiding prioritizing some aspects and not others. Merely choosing what to put in a headline vs first paragraph vs 5th paragraph is making a subjective decision about what is most important.

Even more subjective is choice of words for complex phenomena that defy simple summary. For example, were the activities of last summer protests, or were they riots? Most people have an opinion on which word is a better fit, but there is no way to deny that both have a basis in objective fact. So, writers choose one or the other depending on their own biases and what they want to emphasize.

Finally, there is the "file drawer" source of bias, which is simply not covering some things that you don't think are important. You're not lying or misinforming anyone, but you are making a subjective decision about what people should be thinking about.

I personally think it's best if a publication is open about how it's going to be making these various subjective decisions. And I think it's dishonest for someone to claim their news reporting can be purely objective.

Is it actually less appealing? My understanding is that DBAs consider Oracle very good, just very (very) expensive. This also lines up with my experience tuning queries against Oracle vs Postgres backends. The folk wisdom seems, to me, to be Postgres is 99% as good in the common cases and 90-95% as good in the difficult cases.

Educational Information has already been democratized by MIT Open Courseware, Khan, and others. But Education itself is an interactive human activity carried out by an expert professional. Some of those professionals will be better at educating than others, and people will compete for access to those better educators. There is nothing about Education as a profession that suggests it can be democratized, because the fundamental operational units (educators) are not scalable the way copying texts or videos is.

For a counter-opinion re: "developers don't pay for things" from someone with enviable entrepreneurial success

https://stackingthebricks.com/developers-dont-pay-for-things...

As far as I know there is no good data on "long covid" that substantially controls for confounders like obesity and age. Obese people are likely to have seemingly "long-term" effects from any respiratory problems that actually disappear if they lose weight (see for example obese asthma sufferers who often have what appears like chronic bronchitis). If you are aware of anything that provides evidence that accounts for confounders I'd be very happy to change my mind.

Also, the same-product-again recommendation doesn't have to be a great one; it only has to beat the alternative it displaces. So I could find it plausible that taking a swing on you buying another of a product you already bought beats out whatever the 5th-best recommendation happens to be.

Given how widely this policy is criticized, Amazon must be very aware of it and they must continue to do it intentionally. I would trust, given the amount of money on the line, that they see evidence that tells them it's a reasonable choice.

If you can do it on the water even better, but obviously that's harder to arrange. The basic physiological benefits come from either IMO, and 95% of the time I've been rowing on a machine.

P.S. If I've misunderstood your question and you're asking what machine specifically, I use a Concept 2 I bought 8 years ago but only because that was what my gym was using and I figured might as well match it at home. Nowadays there is a huge variety of machines to choose from, and the basic physiological benefits are going to be the same for all of them.

Others mentioned lifting weights — +1 to that but let me add something I haven't seen mentioned yet: rowing.

Rowing is, in my opinion, the perfect anti-programming exercise.

  - you unfurl your body instead of hunching over, stretching it in a way counter to the usual strains programming creates 
  
  - you exercise your back, especially lower back, giving you support muscles for good posture

  - you also exercise your forearms, building strength to support your wrists

I've been rowing consistently for about 20 years now, and have never once had wrist or lower back pain, despite programming professionally.

It's not characteristic of all forecasting, only purely empirical forecasting.

Definitionally, the only way to reason about risk that doesn't appear in training data is non-empirical (e.g. a priori assumptions about distributions, or worst cases, or out-of-paradigm tools like refusing to provide predictions for highly non-central inputs).

DL is not any better (or worse) than any other purely empirical method at answering questions about fat-tail risk, and the only way to do better is to use non-empirical/a-priori tools. Obviously the tradeoff here is that your a priori assumptions can be wrong, and that too needs to be included in your risk model (see e.g. Robust Optimization / Robust Control).

It is absolutely untrue that DL is immune to fat-fail problems, and it is important that no one operate mission critical systems under this assumption.

The two fat tail questions one has to engage are:

- is it possible that a catastrophic input might be lurking in the wild that would not be present in a typical training set? Even with a 1M instance training set, a one-in-a-million situation will only appear (and affect your objective function) on average one time, and could very well not appear at all.

- can I bound how badly I will suffer if my system is allowed to operate in the wild on such an input?

DL gives no additional tools to engage these questions.

Texas has a limit of I believe a 10% increase in tax assessment per year, so this prevents the worst outcomes but only slows down the rest. If you think of property tax as crudely analogous to rent (from a budget perspective), most families would not be able to endure 10% rising rent every year for an extended period. Yet this future sure seems to be baked into the market here in Austin right now.

Oh hey, you made APE and redbean! Very cool projects, and something I always wondered if it was possible (I once tried a different strategy, failed, and then never tried again). I was very pleased to see someone had succeeded.

> I collect different C libraries as a personal stash of toolkits

Might you have a write-up or link for these, and if so might you be willing to share it?

I have been using Sean Barrett's libraries [0], as well as his curated list of other people's single-header libraries [1], and, like you, I am always on the lookout for new things to add to the collection :)

[0] https://github.com/nothings/stb

[1] https://github.com/nothings/single_file_libs

This is very interesting and helpful, thank you!

I'd be interested in reading suggestions for criticism of Maslow or "win win".

The few criticisms I've found of Maslow seem to be that it's not 100% universal across individuals and cultures. This is true, but I think not much of a criticism of a model that's meant to be a rough guideline to a very complex topic. If Maslow is only 80% valid across all of humanity I think that's a pretty decent model, but if there are deeper criticisms I'd be interested to learn.

Similarly I'm not aware of any criticism of the concept of "win win" which, in my understanding, is just a definition of certain kinds of mutually beneficial situations. So I don't even have a frame for how it could be wrong :) Any reading advice on this subject would also be appreciated.

This lines up with my unverified theory (I can't seem to find hard information about FPS boost online).

My theory is that it detects a list of certain compatible games and then "lies" to the game engine through syscalls for clocks or vsync or whatever the game uses to determine frame timings (this would trick a game locked at 30 to run at 60, for example).

If this is indeed how it works, it could well break other things that relied on those timing systems, like the scripted events you're describing.

If you have an interest in this kind of art, this GDC presentation by Mark Ferrari "8 Bit & 8 Bit-ish Graphics" gives lots of great detail about the art process, tools, and practices of the time (Ferrari worked on the EGA art for Monkey Island, among several others).

https://www.youtube.com/watch?v=aMcJ1Jvtef0

I have a fair amount of experience with this in a startup setting. Unfortunately most of it I picked up in a diffuse way online. If you want to kick around ideas feel free to email me (see my profile).

Just a personal anecdote, but it regularly makes a substantial (like 2x+) difference for me in tight linear algebra loops. It seems to be required if you want to coax clang to emit vfmad instructions for common C linear algebra idioms. Narrow use case, I know, but it can definitely make a big difference in some domains.

Aha! Good correction, thank you. I must have been remembering how long the discovery remained classified, not how long it predated the public discovery.

There's a very amusing epilogue in Steven Levy's "Crypto" [0] about the fact that either RSA or Diffie-Hellman (I've forgotten which) was in fact known to the intelligence services for 10+ years prior to its re-discovery among public researchers. These are people who are very good at keeping secrets.

[0] https://www.amazon.com/Crypto-Rebels-Government-Privacy-Digi...

re: cognitive vs physical development, I think at the level of star athlete most of the difference in performance is, for lack of a better word, mental

If you look at Michael Jordan, Tom Brady, Jeremy Lin, Tiger Woods ... they're very physically gifted, moreso than a typical pro player ... but it's not like they win by being drastically more physically gifted than their competition.

The thing that helps them to win may not be exactly conscious thought, but I think it's effectively a mental/behavioral advantage — certainly to me it seems like a better fit than attributing their success to their physical build (compared to competitors). Michael Phelps on the other hand ... well, some people are just born with the bodies to play certain sports :)

Cars my family bought in 1990 (i.e. 31 years ago) drove happily for 15-20 years. I don't think the "lasts way longer" claim stands up to scrutiny.

Another related aspect to consider is that it's really not very hard for the wealthy to create an inheritance-like legal vehicle even without special cases like trusts.

Should it be legal to contract with someone that you will give them $1B now, and they will provide each of your descendants at least $X/year (or guaranteed job, housing, etc.) for the next 50 years? If this kind of contract remains legal — and I don't see how you could make it illegal without stomping on a lot of contract principles — then the only thing we're really changing is the timing and quantity of taxation.

Allowing inheritance makes intergenerational economic development possible for far more than the ultra-rich (who could already do it anyway). That doesn't mean taxation policy shouldn't be considered, of course.

(Almost) entire game soundtrack by various artists, including a nice version of my favorite piece, Fisher's Hornpipe (note Yo Yo Ma also has a version)

https://open.spotify.com/playlist/0jHjEJUeGqKisKEagkh46N

The professor's quip is cute but myopic. The pieces of paper are claims on future production so the true cost is seen later.

The end result was

  - American Boomers got cars

  - Japanese boomers got USD
  
  - Japanese investors then spent much of the USD on things like US real estate investment vehicles (hardly Japan only, any country that collected large amounts of USD due to trade deficit)

  - American Boomers in prime markets saw tremendous returns on real estate
  
  - and American millennials 30 years later got priced out of houses

Yes I know foreign investment is not the only cause of rapid price growth in real estate. But it is a significant contributor.

It was (and remains) an investment in the sense that it allowed you to live in a nicer place farther from work, or to commute to a job you otherwise couldn't take.

But was it ever an investment as an appreciating asset per se? I'm not aware of such a situation but I'd be interested to learn otherwise.

Two simple reasons. First, "hello world" requires less explanation

  print "hello world"
  vs
  print ("hello world")

Second, printing a line to stdout is an extremely common and special activity (especially during debugging). There is almost no programmer-defined function you could write that has a function as fundamental as that of "print".

Larry Wall of Perl thought about these things in terms of Huffman Coding, i.e. the most common thing you might want to write should have the syntax elide any easily understood context.

In the case of "print", we know we're not talking about any random function here: it's the function for getting output to the terminal, one of the most fundamental operations in all of programming. So I think it's totally fine for it to have a special representation in the language.

Agree with everything you said here, but let me riff on this

> The "print" statement being a great example. Was it kind of weird? Sure.

I am reminded of the Emerson quip "A foolish consistency is the hobgoblin of little minds". The Py2 print statement is special/weird because it is a very important and special behavior, especially for newcomers. It's not randomly special, it's special because it models something that deserves special treatment.

Py3 is more consistent, but less humane. I use it over Py2 only because the market has moved on, but every time I type print with parentheses I curse Guido.