tlb

Turns out that $ profit is one of the least gameable metrics.

Not completely ungameable, but for a consumer product manufacturer in a free market it's the least gameable metric invented so far.

And memory. In the heyday of ATM (late 90s) a few megabytes was quite expensive for a set-top box, so you couldn't buffer many seconds of compressed video.

Also, the phone companies had a pathological aversion to understanding Moore's law, because it suggested they'd have to charge half as much for bandwidth every 18 months. Long distance rates had gone down more like 50%/decade, and even that was too fast.

It’s nontrivial why it’s mc^2 and not 1/2 mc^2, since kinetic energy generally is 1/2 mv^2

It’s lucky when you can get the sign right on the first-order effects. Predicting 2nd order effects accurately enough to price them is economic voodoo.

Comparing fighter pilots to helicopter pilots might say something about the effect of G forces, since the cognitive loads are both high.

However, the control group is “10 adults with no experience in flying … matched for age, gender, and educational level”. I feel like fighter pilots are pretty different from other people with the same amount of education.

I find "this wasn't the point of copyright", referring to the motivations of 18th century legislators, unpersuasive. They were making up rules that were good for the foreseeable future, but they didn't foresee everything and certainly not everyone being connected to a global data network.

Persuasive arguments should focus on what's good for the world today.

Added fructose seems bad for you. Specifically, a high ratio of fructose:glucose (which doesn't occur in nature) does something weird to your metabolism. Your gut uses some hack like detecting only glucose to trigger intermediate steps in the metabolism of both, and consumption of high fructose:glucose foods causes weight gain.

More: https://pmc.ncbi.nlm.nih.gov/articles/PMC4145298/

The relevant comparison is the energy needed to turn asteroids into food vs. the energy needed to turn exhaled CO2 and poop back into food. Astronauts live in closed ecosystems, so you don't need to bring in additional mass. Just close the cycle.

Most bread and yogurt in the average grocery store is pretty bad stuff, full of HFCS, hydrogenated oils, and hydrolyzed proteins.

But yes, I'd rather have a classification that clearly separates Coke and Cheez Doodles from actual foods. There are some multi-billion dollar lobbies to prevent that happening, though.

People only have to think about property pricing every several years, when they move. Electricity spot prices change by the minute. So the mental burden is quite different.

The list of examples isn’t very compelling. Extensions should be things that can’t be done in Javascript.

Your phone knows how often you normally use it. It could do something intelligent based on a usage pattern to restart when there’s an unusually long gap.

I can see this being too hard to explain compared to “reboots after 72 hours.”

Don't let recruiters drive your career. Apply for jobs at companies you want to work for, and ask friends in the business to refer you.

PTC (positive temperature coefficient) thermistors do that. https://en.wikipedia.org/wiki/Thermistor.

They’re used as auto-resetting fuses in many devices.

> control flow graph contained 132856 basic blocks

That is a stunningly large function. I've looked around the onnxruntime sources and can't find anything like it. The largest C file is under 6000 lines. Does anyone know what function it's referring to?

You could store all the data on the satellite, upload new code to process it differently, and download the resulting image. Then, the communications link just has to handle code (several MB up) and images (several MB down) instead of petabytes of data.

The launch mass of a petabyte of SSD is under 10 kg. I don't know if it would survive 17 years of space radiation though.

The sad thing is that none of it is very amusing. Current events twitter is more aggravating than amusing. We're aggravating ourselves to death.

If you have 2 periodic waves of amplitude a1 and a2 coming from different directions the peaks combine at every intersection, so there's always a grid of a1 + a2 double-peaks. If the wavelengths are around 100m, there's a double-peak every 10000 m^2 (1 hectare).

If you add a 3rd periodic wave from another direction, at any point in time some of these double-peaks will align with the peak of the 3rd wave to form a triple-peak. If you set the cutoff at a1 + a2 + 0.95a3, then about 10% of the double-peaks are triple-peaks (because cos(10% * pi) ~ 0.95).

Add a 4th orthogonal wave and about 1% are at least a1 + a2 + 0.95a3 + 0.95a4. So with 4 waves, we have a quad-peak every 100 hectares (a small farm).

See the monthly Who is Hiring threads and search for functional, Haskell, OCaml, etc. https://news.ycombinator.com/item?id=41425910 has a few hits.

Also, this exists: https://functional.works-hub.com/

I haven’t seen someone confess to this type of science fraud. They mostly deny, say nothing, or mumble “I don’t know how that could have happened.”

It would be a service to the world for someone to come clean and explain their actions and motivations. It has happened in other fields — Bernie Madoff gave a good account.

In your dequeue/circular buffer implementation, how is it able to grow the queue without locking?

The code seems to rely on atomics for head & tail, but grows the queue without any special provisions I can see.

https://github.com/mtmucha/coros/blob/ee30d3c1d0602c3071aa26...

So, treat us to some more imaginative speculation.

Fossil fuel plants are chemical -> heat -> mechanical -> electrical. The heat -> mechanical step is the inefficient one, sadly limited to ~40% efficiency even with very fancy machinery. Most other conversions can be above 90% efficient.

That should have been done from day 1. Now, however, there are 1000s of doctors who make fat stacks prescribing CPAP machines, and they'll fight tooth and nail to not lose their whole business.

I don't have a big problem with the money: it's a fraction of month's salary. The usual problem with commercial languages is the freedom to set things up how I want and spin up new environments quickly. Mathematica, for instance, is a significant pain to run on my laptop, desktop, and a cloud machine.

I think there's room to innovate in licensing. I would err on the side of making it easy to install at the cost of some piracy. A great language could 100x its market share by losing 1/2 to piracy, so 50x overall.

The ROI of security is too noisy at a single company where you (hopefully) get less than one incident per year. But across the industry there is enough data to estimate ROI of various procedures.

There's no way to design vehicles that can't be made unsafe by a mechanic making a mistake. Even for purely mechanical things like brakes, a mechanic can leave out a bolt or crack a structural member so that it breaks on the road. EVs require careful procedures, for instance to make sure that any of the high-voltage wires can't rub against anything and wear through the insulation.

Dealer mechanics make mistakes too, but at least they're trained on the specific car, and they can collect data to notice a pattern of things breaking after certain people worked on them.

In software, most people dislike most other people's work. I wonder how much that varies between fields, or if it's been studied systematically. It might be pleasant to work in a field where you'd enjoy most other people's work.

An editor feature I'd like is that as I'm typing an identifier, a hover popup shows me how many other instances appear in my codebase. It should be easy to build a map of identifier->count for instant lookups. I generally know if I want 0 (a new unique identifier), 1, a small number, or a large number. For a few pixels, this would prevent a lot of dumb mistakes and ambiguous names.

However, they're comparing against no gut bacteria at all, a condition never found in nature. Gut bacteria are necessary for mammals to digest some foods. "Eliminating all gut bacteria harms fetal development" could be another title.

It'd be more interesting to compare different plausible microbiomes.