• 0 Posts
  • 19 Comments
Joined 2 years ago
cake
Cake day: June 28th, 2023

help-circle





  • Hard disagree.

    This week, I’m designing a circuit which would traditionally use relays, but I’m considering IGBTs instead. IGBTs weren’t designed for my industry, but they’re so cheap (thanks to quadcopters) that I can just overspec them and get the job done despite the lack of optimization.

    Grid scale energy storage was already being researched before the EV boom – remember when people stopped talking about vanadium-flow? EV batteries undercut stationary-optimized batteries in $/kWh because EVs are lucrative enough to drive the research that much harder. Without the EV industry as the incubator for competing battery tech, stationary storage would still cost what it did in 2010.


  • In the USA, out of every economic sector, transportation creates the most GHG emissions [EPA1], and the majority of that is from passenger vehicles [EPA2]. Significant portions of the industrial sector’s emissions come from refining automotive fuel [EPA3]. US total GHG emissions are down around 20% from their peak in 2005, but almost all of that has come from the electrical power sector [CBO1][CBO2]. Vehicular pollution has dramatic direct health impact on top of GHG emissions [HSPH].

    Transport emissions are the long pole in the tent for the US. Solutions to that will be the focal point of US climate strategy for the next decade. Barring the demolition of the majority of US housing to re-establish walkability, our two best solutions are EVs and public transit.

    EVs cut lifecycle emissions by about 55-60%. [UCS][ANL][MIT][ICCT][BNEF][CB][MIT][IEA]

    Public transit cuts lifecycle emissions by… about 55-60%. [IEA][AFDC][USDOT]

    Neither is a magic bullet. Both get their asses kicked by bicyles (and to a lesser degree, microcars). Both get better with increased passengers per vehicle. Both can be fueled with renewable energy for additional reduction. Both can be manufactured with renewable energy for additional reduction. Both take surprisingly equivalent amounts of raw resources and energy. EVs need batteries that are carbon-intensive under current practices, but rail needs large quantities of steel which is equally carbon-intensive under current practices.

    There are a ton of factors I can barely touch on here, so here’s a rapid-fire overview. Public transit offers unique advantages from an urbanist perspective and the liveability of cities [ST], but that’s objectively different from sustainability. The US has such low average ridership/occupancy that our busses have more emissions per passenger mile than our cars [AFDC1][AFDC2], and that was before the pandemic – it’s even worse now [NCBI]. Low ridership can be partly attributed to the incompatibility of American suburbs with public transit – which could be a major roadblock because 2/3rds of Americans own detatched homes [FRED], representing $52t [PRN] in middle-class wealth that they will likely defend with voting power. Climate solutions will need to maneuver around this voting bloc. I personally think individual EVs and intercity rail are complementary technologies – the more cheap (short-ranged) EVs are out there, the more people will lean on public transit for long trips. Heavy rail gets way better efficiency per vehicle mile than light rail or commuter rail and I have no clue why [APTA][ORNL], but I’m not as impressed by light rail as I expected to be. Since public transit and personal transport leverage different raw resources and face different challenges to adoption, we will achieve the most rapid decarbonization if we do both at the same time.

    TL;DR

    This is a huge, huge question, and anything short of a dissertation would fail to answer it objectively. My best answer is that the most effective solutions to climate change are diverse, engaging multiple technologies in parallel. EVs are a piece of the puzzle, but not a one-size-fits-all solution.



  • Fashion accessories. For most fashion (not workwear), the expensive stuff is made from the same material and in the same factories as the cheap stuff, they just market it harder.

    Body wash. It’s watered-down soap. Just buy a bar of soap.

    Amazon Prime. Amazon used to be space-age Sears. Now it’s just Aliexpress. Fake reviews and bribery are rampant, dangerously nonfunctional products get top recommendations, used and broken products get resold as new while untouched returns get thrown into landfills, Amazon Basics violates IP, and they’re putting ads in Prime Video now.

    Microwaves and space heaters. The boxes may try to convince you otherwise, but the amount of heat these devices can deliver is bottlenecked by the power outlet. Every 1100W microwave is just as effective as the others. If you’re paying more, it’s for looks and for features you’ll never use like popcorn mode.

    Electronics, for most people. Most people won’t get more use out of a new $1500 phone than a last-gen model from the same manufacturer for $500. Do you really want a $200 smart coffee maker, or a $20 dumb coffee maker with a $10 plug-in timer?

    Software. Obligatory FOSS plug. I don’t blame people for sticking to what’s familiar, but if you have the time and energy to spare tinkering, most software out there has a good free or open-source equivalent these days. At least for personal use. In my use case, LibreOffice beats Microsoft Word, Photopea beats Photoshop, and Google Sheets beats Excel.



  • You’re not alone there. America’s fragile social safety nets and minimal workers’ rights mean we’re all one bad day away from being in your position.

    • Chemo can cost $50k, and FMLA exists (somehow) but only covers 3 months. A cancer diagnosis could easily wipe out six figures of savings.
    • Most employers require prompt attendance, so a car crash that leaves you dependent on America’s crumbling public transit systems will likely lead to losing your job.
    • The GOP is trying to kill off Social Security, so if you ever want to retire, you might have to save twice as much as what’s traditionally advised.
    • With climate change being what it is, losing your house to a hurricane or wildfire is a much bigger threat than it used to be.
    • Economic swings, downsizing, and “k-shaped recoveries” can make jobs vanish or become unsustainable without warning.

    Even for the top 10%, saving for these contingencies typically precludes a wealthy lifestyle, particularly for younger people with fewer savings. There’s now an entire demographic of (mostly millenial) Americans with traditionally high-paying jobs who still live at or below middle-class aspirations due to saving: HENRY, or High Earner, Not Rich Yet.


  • This.

    Last month, I installed Mint, which is my first ever Linux install. I chose it because people said it would be the most hassle-free.

    The bugs currently plaguing me include:

    • Steam’s UI scaling is off, to the extent that I practically need a magnifying glass to read it.
    • Bluetooth has now decided that it no longer wants to automatically connect to my speaker.
    • Discord won’t share audio during screen sharing anymore.

    But the big one, the one that made me stop and think, was the keyboard. Right out of the box, my function keys (brightness, airplane mode, etc) would not work. This turned out to be because the laptop was not recognizing its keyboard as a libinput device, but treating it as a HID sensor hub instead. To fix it, I had to:

    • Find similar problems on the forums and recognize which were applicable to my case.
    • Learn what the terminal was and how to copy code into it.
    • Learn that the terminal can be opened from different folders, which alters the meaning of the commands.
    • Learn the file system, including making how to make hidden files visible.
    • Figure out that a bunch of steps in the forum were just creating a text file, and that any text editor would do.
    • Figure out there were typos and missing steps in the forum solutions.
    • Learn what a kernel is, figure out mine was out of date, and update it.
    • Do it all over again a month later when for some reason my function keys stopped working again.

    For me, this was not a big deal. It did take me two evenings to solve, but that’s mostly because I’m lazy. But for someone with low technical literacy (such as my mom, who barely grasps the concept of ad blockers in Google Chrome), every one of these bullet points would be a monumental accomplishment.

    The FOSS crowd can be a bit insular, and they seem to regularly forget that about 95% of the people out there have such low technical literacy that they struggle to do anything more involved than turn on a lightbulb.



  • I’m not 100% convinced by some of the terrestrial applications for H2, on the economics side.

    In my opinion, the aviation industry can handle the cost increases inherent to greener fuel. People fly because it’s fast, not because it’s cheap. As long as the planes are still fast, there’s still a market.

    By contrast, people ride the bus because it’s cheap. According to Tokyo, H2 busses cost 2.6x as much to operate as diesel. According to Montpellier, H2 busses cost 6.3x as much to run as battery-electric busses (that’s including amortization). So while the tech seems like a great fit, the commercial case is weak.

    Shipping with semis is a toss-up. H2 can transport more cargo a longer distance than batteries, and I think some people will pay the premium for next-day shipping. But personally… I’d get the cheap-but-slow shipping 90% of the time.


  • Hydrogen works pretty well for aviation, though there are three main challenges they’re still working on: size, materals, and fuel source.

    Hydrogen is nice and lightweight, but the tanks and plumbing take up a lot of space, which cuts into cargo volume, basically limiting the range if you want to take passengers with you.

    The second issue is that fuel cells currently require quite a lot of platinum, and the PEM electrolysis also requires a lot of PGMs and rare metals like Iridium. The material scientists are working on this, and I figure if they can take the cobalt out of batteries, they can take the platinum out of fuel cells.

    The question that comes up the most when talking about hydrogen is where the hydrogen itself comes from. Right now, it’s mostly made by steam methane reformation or similar fossil fuel processing, which is nearly as bad for the environment as burning the fossil fuel directly. But there are promising advances in renewable electrolysis (such as taking advantage of peak solar for “free” electricity) which are closing the gap between SMR and renewable H2. It’ll never be as cheap as jet fuel, but it’s at least economically feasible.



  • It’s absolutely true that the training process requires downloading and storing images

    This is the process I was referring to when I said it makes copies. We’re on the same page there.

    I don’t know what the solution to the problem is, and I doubt I’m the right person to propose one. I don’t think copyright law applies here, but I’m certainly not arguing that copyright should be expanded to include the statistical matrices used in LLMs and DPMs. I suppose plagiarism law might apply for copying a specific style, but that’s not the argument I’m trying to make, either.

    The argument I’m trying to make is that while it might be true that artificial minds should have the same rights as human minds, the LLMs and DPMs of today absolutely aren’t artificial minds. Allowing them to run amok as if they were is not just unfair to living artists… it could deal irreparable damage to our culture because those LLMs and DPMs of today cannot take up the mantle of the artists they hedge out or pass down their knowledge to the next generation.


  • It doesn’t change anything you said about copyright law, but current-gen AI is absolutely not “a virtual brain” that creates “art in the same rough and inexact way that we humans do it.” What you are describing is called Artificial General Intelligence, and it simply does not exist yet.

    Today’s large language models (like ChatGPT) and diffusion models (like Stable Diffusion) are statistics machines. They copy down a huge amount of example material, process it, and use it to calculate the most statistically probable next word (or pixel), with a little noise thrown in so they don’t make the same thing twice. This is why ChatGPT is so bad at math and Stable Diffusion is so bad at counting fingers – they are not making any rational decisions about what they spit out. They’re not striving to make the correct answer. They’re just producing the most statistically average output given the input.

    Current-gen AI isn’t just viewing art, it’s storing a digital copy of it on a hard drive. It doesn’t create, it interpolates. In order to imitate a person’t style, it must make a copy of that person’s work; describing the style in words is insufficient. If human artists (and by extension, art teachers) lose their jobs, AI training sets stagnate, and everything they produce becomes repetitive and derivative.

    None of this matters to copyright law, but it matters to how we as a society respond. We do not want art itself to become a lost art.