GeertB

The article made it seem that the tags way on the order of half a gram, but that would be about the butterfly's weight. Even at 60 mg it's a significant additional burden for the butterfly, I'd expect. Even if you only take the heaviest Monarch butterflies that weigh in about 600 mg, it is 10% of their body weight. Like an 6 kg backpack for a human weighing 60 kg.

I've definitely been guilty of wording my multi-line doc comments in a way that would ensure they'd flow without awkward right margins. I would have a macro in vim that would call the fmt program to reflow the comment to 79 (later 98) characters, and then would iterate on word choice until the comments looked good. Not exactly right-justified, but no very ragged edge either.

Nvidia's been pretty terrible for open source / free software. No need to quote Linus Torvalds here. They want to control what runs on their hardware. They want to you write code against their proprietary drivers and APIs, not directly against the hardware (which these days of course also contains plenty of software, but still).

Don't expect things to go differently this time around. Nvidia wants control over the software stack. Acquiring HF fits in perfectly. The play is long term.

Signed integer addition is only associative when overflow is defined to wrap around like unsigned arithmetic. This condition is matched here, because only debug builds panic on overflow. However, it's a bit of a gray area that the article completely ignores.

I remember that EMACS meant Eight Megabytes And Constantly Swapping, a reference to the Emacs editor requiring a system with 8 MB of memory and then still swapping. It was considered an exaggeration. Even Windows 95 could run OK-ish on a system with 4 MB of memory, with 8 MB being plenty. We have gained some things in the meantime, and lost some too.

The program terminated just a few moves into the game when there were plenty of legal moves left. Also, it is surprisingly hard to play chess as a human when there is no checkered board.

I don't quite understand how multiply doesn't require addition as well to combine the various partial products.

The article makes it sound like the Tesla Semi is physically infeasible. Yet, it is in active use on a sufficient number of long-haul routes that ignoring this proof of existence undercuts some of the central points the post tries to make.

The combination of higher efficiency, regenerative breaking, and some regulatory wiggle-room such as slightly higher allowable gross-weight (2000 lbs in the US, and 2000 kgs in the EU), together with reduced maintenance cost and time significantly affect the economics of trucking.

As regulatory frameworks price in more externalities of internal combustion engines, such as the climate and health effects of their emissions, burning diesel will no longer make economical sense. All road transport will end up being battery-electric. The declining cost of owning and operating electric vehicles compared to internal combustion ones will reach this point even without regulatory changes, just at a slower pace.

monochrome

Can't help myself: whenever I see Windows 3.11, I have to compute the difference between it and Windows 3.1. So, I fire up the calculator from the Accessories folder, and compute 3.11 - 3.1. The result: 0.00. Try it! :-)

How did you handle games where Stockfish would castle or promote?

For these devices the microcontroller needs to be super cheap. Microcontrollers like the Puya PY32 Series (e.g., PY32C642, PY32F002/F030) can cost in the $0.02 - $0.05 range for the kind of many-million volumes applicable for disposable vapes. These are 32-bit ARM Cortex M0 MCUs, running at a 24 MHz clock or similar, some with 24 KB of ROM and maybe 3 KB of RAM!

To put into context: this is 3x the ROM/RAM of the ZX81 home computer of the early 1980s. The ARM M0 processor does full 32-bit multiplication in hardware, versus the Z80 that doesn't even offer an 8-bit multiply instruction. If we look at some BASIC code doing soft-float computation, as was most common at the time, the execution speed is about 3 orders of magnitude faster, while the cost of the processor is 2 - 3 orders of magnitudes less. What an amazing time we live in!

I wonder if this will improve latency and signal strength

The title here says 100M, but the actual presentation says 1M?

That one has a special place for me. I was a consultant then and invited to join the chorus even though I wasn't at the more professional level the other cast members were. We did a performance in London and one in St Louis. Still, it was a lot of fun, and how I got to meet my spouse 25-ish years ago.

I recently visited SF with the fam. Besides riding the old-fashioned cable cars up and down, to and from, our Fisherman's Wharf apartment, it was a foregone conclusion that we'd try out the most contrasting way of transport: a Waymo. So, we took one to the Golden Gate Bridge park, a 20 minute or so ride. Everything was smooth, clean, nice music etc. The driving was much calmer and frankly more comfortable than any Uber driver we've had. After a few minutes of nervously being in awe, everyone started to relax and enjoying the ride more. We took a few more Waymo's over the next few days, and the experience was consistent.

A few days later as we went to the airport, our erratic Uber driver made us all wish we could have taken a Waymo. Alas, Waymo doesn't yet go there, or on any highways at all, so we'll have to wait for a little longer. However, on our trip we saw both transportation from the past and from the future. There's no doubt in my mind that driverless taxis will not only arrive everywhere, but take over quicker than we may now think.

This game is so fun! I spent entirely too much time playing it just once last night, and got to the 1024 piece in the first run. I don't know how long it took me, but it was long. With 2048 my runs are far quicker.

A smart compiler would have used the test (x & (x - 1)) == 0 to verify that x (assumed non-zero) is a power of two. Don't know why Clang didn't do that here.

This article really confuses the problems that legacy car manufacturers have with switching to EVs, with EVs in general. EVs work, just look at countries farther ahead in the transition, or ask the 1.8 million people who bought a Tesla last year. The transition is in full swing.

Tesla's excellent supercharger network is already more than sufficient for roadtripping in the USA and Europe, and now is open to non-Tesla EVs. This piece of writing is just a poorly researched bit of trash.

That's quite the bandwidth-delay product at 3.3GB.

Actually, that's mostly false. If you can charge at home, you'll have a fully charged car in your driveway every morning. With the Tesla charging network, you really can charge no matter where you want to go. Typically after 4 hours or so of driving, I'm ready to take a bathroom break, stretch my legs and get a coffee. By the time I'm back the car is charged to 80% or so, good to go for another 250 miles or so.

It seems that the proposed system is an improvement over plants, but not necessarily over the Sabatier reaction driven by energy captured by solar cells. Is that right? How would these compare?

The new satellites will orbit the EU.

You probably work too many hours, and actually can more often say 'no' than you think. Delegate some of your responsibilities. Work smarter, not harder. Be transparent with your manager about your night classes and the need to change some schedules. If your manager is sensible, they'll take 80% of your known performance at our level of experience over uncertainty and a long ramp to reach the equivalent of your 80%. Sure, there is a chance they won't accept, but it might be a risk worth taking. Oh, and by going to school you surely show that you want to grow.

In many cases, software is like a tool. I provided a screwdriver: generally it's used for good, but sometimes it's used for assembling a bomb. By providing the screwdriver, I have no influence on how it's used. For me, it's important that most software I write is free software with few restrictions on its use. That way, I ensure that my screwdriver is available for all good purposed, even though I acknowledge that I cannot prevent it from being able to be used for evil. It's a tricky balance, but I trust that the good uses of new tools outweigh the evil. A war on screwdrivers won't prevent missiles targeting civilians.

In particular, if you're developing Free Software or Open Source Software, and a company wants to hire you, it is _expected_ that they're OK with you continuing to do so. If any contract says otherwise, explain that you need to keep doing what you're doing in order to stay at the level of expertise that you're at. If a future employer would not agree for you to keep doing that, you shouldn't sign.

When I started studying CS in 1990, people told me that "programming" was a dead end: 3rd generation languages wouldn't require programmers and anyone would be able to bend computers to their will without programming. I didn't heed their advice... We're farther from retiring the profession of Software Engineer than we were 30 years ago. But, is that a good thing?

Why not just keep the wire attached to the Yubi-key, but leave it electrically floating (high-impedance state) and have the board ground it whenever it needs to be pressed? No need for mechanical triggers...

Their goal is absolutely to drive down cost. Right now they're battery-constrained, so it would not make sense to sell them to third parties. If Tesla can only sell a given number of cars, it makes financial sense to sell them at a higher price. The extra revenue serves to build more factories and grow faster. As production keeps scaling, average selling prices will keep going down. If they'd attempted to start with a $25K car, they'd producing no cars today.