Insects are a lot stronger than humans, relative to their weight. This may be simply because of the size difference or a combination of size and materials. If you double the size of a butterfly, its weight will increase by a factor of 8, but the surface of its wings will increase only by a factor of 4.

Using a visualization of raisins and grains of rice is better than using humans. A picture of the device attached on a butterfly might be even better.

I just put a bunch of raisins on my kitchen scale, and they average 200 mg each, so Jonathan's raisins must be about six times bigger than mine. I'll weigh my grains of rice later.

This has been discussed before: https://science.slashdot.org/s... There basically needs to be about the same area of cooling panels as solar panels, as can roughly be seen from this picture: https://en.wikipedia.org/wiki/...

Comparing TIOBE with Github octoverse, for the top 10 languages.

TIOBE: Python, C, C++, Java, C#, JavaScript, Visual Basic, Delphi/Object Pascal, Perl, SQL

Github octoverse: TypeScript, Python, JavaScript, Java, C#, PHP, Shell, C++, HCL, Go

Common languages: C#, C++, Java, JavaScript, Python

TIOBE only: C, Delphi/Object Pascal, Perl, SQL, Visual Basic

Github octoverse only: Go, HCL, PHP, Shell, TypeScript

Intepretation

SQL statements are usually embedded in another language, so github probably counts them in the enclosing language.

Perl and Visual Basic programmers don't use git, so github ranks them low. HCL files are probably deployed via git, so they appear in Octoverse

TIOBE ranks Go lower, because Go has better offline documentation (go doc) than other languages, so Go programmers don't search the web so much for API documentation.

TIOBE has a low rank for TypeScript, because TypeScript APIs are really JavaScript APIs, so people search for JavaScript

I don't know how to explain C's absence from the Octoverse top-10

Who uses Delphi/Object Pascal?

"More popular" means more used, not more liked. To determine "more liked", a human survey is needed, where humans answer a questionnaire. stackoverflow does such a survey.

Yes, of course. I must have responded to the comment title, missing the comment "dramatically".

The forecast is in the article we're discussing:

therefore the forecast is 240*12/9 = 320GW for the year, which is less than last year:

According to the article's numbers, China is adding ~ 350 GWp of solar per year, more than the rest of the world, so let's say the world is adding 700 GWp of solar each year, and another 100-200 GW of wind, so 1 TWp of solar + wind. Let's assume that solar+wind production averages out to 25% of their peak capacity. Then this 1TWp of added renewables per year averages out to 250 GW. The world consumes 23000 GW, of which about 70% (16000 GW) is from fossil fuels. At the current pace, it would then take 64 years to replace all fossil fuels with renewables, without taking into account increasing energy use or running out of space to put solar+wind, or energy storage. These are very rough estimates. The time frame seems comparable to the transition from coal steam boats to diesel engines. The first oil refinery was built in 1856. The Titanic still burned coal in 1912.

apt sources will need Rust to compile. That doesn't mean that apt binaries will need Rust packages to run.

I suppose they could if they were linked dynamically to Rust libraries, but I think Rust by default links Rust libraries statically.

Do you run a compiler on small embedded systems? I install precompiled firmware on mine.

I get the impression that they will require Rust for systems that compile Debian, not for all Debian systems.

Rust also has a "stronger approach to unit testing", as mentioned in the referenced email. The Rust "cargo test" command can compile and run example source code snippets that are embedded in comments, to make sure that such examples compile, and the assertions don't fail. Rust lets you link actual source code files in a comment, so the linked code appears in the generated documentation, inline. It has a more integrated approach to code, tests, and documentation than other languages.

So why is IKEA moving from Zigbee to Matter? What's the advantage of having yet another protocol? Why does it take so long to settle on one standard? Zigbee, Z-Wave, Threads/Matter, what else? To use each of these protocols, you need to buy a different hub. It's not enough to install new software, you need new hardware too. You can't just install a Zigbee or Threads app on your computer and use it over the existing WiFi or Bluetooth hardware, even though they all use about the same frequencies. You have to get a Zigbee dongle of sorts, or something else for Threads, etc...

WiFi and Bluetooth started being used at ~ 1999, and they are still used widely today. Compare this with IOT protocols.

It takes time and money to resist tyranny.

It takes time and money to setup your own mail server instead of handing over your personal communication to Big Tech.

Some people value their privacy and independence enough to put in the time and money it takes to resist.

On a positive note, there are now companies that sell devices with Tasmota preinstalled, instead of their proprietary OS that works only with their servers. There are companies that advertise their devices to be compatible with zigbee2mqtt. There are companies that sell routers based on OpenWRT.

There have been too many cases of companies bricking their devices because they stop supporting their proprietary OS.

It takes time but saves you money to run Linux instead of Windows.

It takes time but saves you money to install Linux on a PC that Microsoft stopped supporting.

It is more economical for them to produce liquid air non-stop. Many industries could theoretically consume electricity only when it is abundant (during the day), but then they under-utilize their facilities. Redundant data centers could be spread around the world so that each data center operates only when the sun shines nearby and it has abundant electricity. But then they would operate only about 20% of the time, and therefore require 5x higher capital costs (computers, cooling, buildings, land, etc).

The 2nd step is intriguing, but it is not obvious that it is economically viable. If they generate more liquid air than they need and use it to generate electricity, then they need extra compressors. What advantage would they have over a separate plant which uses those extra compressors to store electricity without selling liquid air?

My question was about the geodesic polyhedrons that approximate the sphere, not about the sphere.