• 0 Posts
  • 25 Comments
Joined 1 year ago
cake
Cake day: July 1st, 2023

help-circle
  • No.

    MW is the maximum capacity not the average.

    A nuclear reactor runs at close to its maximum output pretty much 24/7/365.

    A solar farm only operates during the day, and even then it only operates at maximum output in the middle of a clear sunny day.

    The overall average output of a nuclear plant is typically around 90% of its capacity.

    The overall average output of solar farm is 20-25%.

    This massive farm will still only output a bit more electricity than what a single nuclear reactor outputs.

    A nuclear power station typically has more than one reactor, so compared to a typical nuclear power station this isn’t even close to the average nuclear plant.

    Though it does beat a few of the smallest nuclear plants that only have a single reactor.

    Nuclear outputs a fuck-ton of electricity for its size.


  • Sandford Fleming (the guy who invented time zones) actually made it easier.

    Before timezones, every town had their own clock that defined the time for their town and was loosely set such that “noon is when the sun is at its highest point in the sky.” Which couldn’t be measured all that accurately.

    If it wasn’t for Fleming, we’d be dealing with every city or town having a separate time zone.


  • Yes you’re correct. I will qualify my previous statement as hydrogen powered road vehicles don’t make sense for now.

    The problem at the moment is that electricity generation is not carbon free and in most countries not even close.

    Unfortunately the transition to a carbon free electric grid is being significantly retarded by policymakers that are, as you say, myopic. As a result it will be at least two more decades before hydrogen makes sense.

    The carbon footprint of lithium battery manufacturing, is small compared to the carbon footprint of electricity generation. Until that changes significantly lithium batteries will continue to be a better choice than hydrogen fuel cell.

    Hydrogen may make sense in a future where we’ve eliminated all fossil fuel electricity generation and there’s an abundance of carbon free electricity that can be used to create green hydrogen as a form of energy storage. Though by the time that point comes, we may have developed battery technology or some other energy storage technology that doesn’t carry the same carbon footprint that lithium ion does today.


  • Hydrogen doesn’t make sense and never did as a strategy to reduce greenhouse gas emissions in vehicles.

    Most hydrogen is made from fossil fuels, and has a lot of emissions during manufacturing. But even green hydrogen, which is made by using carbon free generated electricity to split water into hydrogen and oxygen doesn’t make sense.

    If you’ve build new renewable power it’s more efficient to use it to charge batteries than to use it to generate hydrogen.

    There might be a case for compressed hydrogen, In vehicles where batteries are too heavy like aircraft.

    But for road vehicles, batteries are more effective at reducing emission.

    If you’re building any new renewable power, you’ll reduce more emissions by using it to displace coal power, the to generate green hydrogen.

    Some day when we’ve eliminated fossil fuel based electricity generation, Green hydrogen might start to make sense. But anybody trying to do it right now is not being as helpful as they could be.



  • I think I have you slightly beat… mine was an Apple II+, circa late 1981, with a disk drive, and a monochrome green screen monitor.

    First cell phone was around 1997. Though I honestly don’t remember what it was. I recall having a Nokia model from before they made that indestructible model in all the memes, as well as a Kyocera one that I could connect to a laptop and have wireless dial up internet at some abysmal speed like 20 kbps. (0.02 mbps). I had at least two more phones, including a Treo 650 “smartphone” before getting my first iPhone, a 3G. I’m on my sixth iPhone now.



  • It’s called home realm discovery. It’s common in business apps though it’s usually used with email & password logins not username & password logins.

    It’s done that way to support federated logins. Larger companies will often used a single sign on solution like Okta or Azure AD. Once the user’s email address is entered it checks the domain against a list of sign on providers for each domain and redirects the user to their company’s federated login if it finds it there instead of prompting for a password.

    This has several benefits:

    1. The user doesn’t have mutiple passwords to remember for different apps. Which is know to result in users either reusing passwords or writing down passwords somewhere.

    2. When an employee quits or is terminated the company only needs to disable their account in their company directory and not go into potential dozens of separate web apps to disable accounts.

    3. The software vendor never receives the password, if the vendor’s system is compromised they don’t even have password hashes to leak. (Let alone plain text or reversibly encrypted passwords)

    Websites that work that way are (usually) doing it right. If that doesn’t work with your password manager, you should (probably) blame the password manager not the website.




  • Lived in a house that had a heat pump with resistive electric heat as a backup in Canada. Never noticed a significant difference between that and other houses I’ve lived in that had natural gas furnaces.

    Aux heat would kick whenever it was below about -5°C. That house would be about 20 years old now and had decent insulation for the location and age. It never really felt like the furnace struggled to keep the house warm, or was running all the time.

    Cost wise it didn’t seem significantly better or worse than natural gas. It was definitely using more juice in the winter when there was a cold snap, but it wasn’t crazy amounts. The electric bill was actually highest in the summer when the heat pump was cooling.


  • I’m not intending to defend Musk or Tesla here, but this study is literally just insurance incidents by brand and makes no distinction between Teslas on autopilot and under human control.

    Teslas tend to attract a certain kind of driver that likes their performance characteristics who are not typically known for being the safest drivers.

    There’s no doubt that a lot of Tesla drivers abuse the autopilot capabilities, and the Elon Musk hype machine is at least partially to blame for that, probably more.

    But this isn’t evidence one way or the other about the safety of Tesla’s FSD.



  • Both China and Russia have built operational SMRs. (Not to mention the fact that the nuclear reactors we’ve had for decades in military submarines and ship are SMRs). They exist.

    We don’t have enough data about the economics or SMRs to say for sure, most (but not all) economic models put LCOE for SMRs at half the cost of traditional PWR nuclear reactors.

    It’s hard to judge from the current smr projects what the costs will be. The largest cost in building nuclear power is all the regulatory oversight. Every PWR plant is different and needs to go through the entire process from scratch. But once we have some successful and proven SMR designs. They can be mass produced from the same vetted and approved designs without needing to go through the massively expensive design process again.

    SMRs are simpler too. Which makes them cheaper. They don’t need as many layers of redundant safety systems like traditional reactors do. Even in the worst case scenario, an SMR can meltdown and a person living next door would be perfectly safe.

    All of that adds up to the a lot of potential cost savings if we mass produce them.

    If we can build enough solar or other renewable power to replace fossil fuels without nuclear, great.

    But most people have no idea just how much it’s going to take. We need to not only replace all the fossil fuels on the grid today. Plus have extra capacity to charge storage for use when its night and cover the added demand of all the electric cars, trucks, furnaces, everything else that needs to become electric.

    We need to be building nuclear too. We can’t build enough solar and wind fast enough.


  • I will absolutely give you that transitioning an established mature product to the subscription model is usually a terrible idea. Plenty of examples of that going horribly wrong.

    As for subscriptions being a “blatant money grab” that definitely happens sometimes… notably when there’s a mature product with a dominating market share. The company already captured most of the market share, so they can’t get much more revenue from new customers, existing customers are satisfied with the version they have so they’re not buying any updates. Sales go down and someone comes along say just make it a subscription and keep milking the cash cow forever…. Yep, I admit it, that totally happens. The enshitification ensues.

    But none of that’s the fault of the subscription model per se.

    The same subscription model that becomes the incumbent’s downfall, is what creates a market opportunity for a new competitor.

    A new competitor can coming in with a new product that was built with a subscription model from the start. The competitors product is cheap to try for a month, cheap to switch to with no big upfront costs. The newcomers can generally react much faster to customers needs than the incumbent. (Not because of the model, they can because they’re smaller)

    Established software companies doing blatant money grabs happen all the time. Hell most of us are here using Lemmy because Spez attempted a blatant money grab on Reddit. Had nothing to do with the model.

    Subscription model gets a lot of hate because greedy companies tried to use it as a blatant money grab exactly as you described. But it doesn’t have to be that way.

    Subscription models make it easier for newcomers entering a space, which is good for consumers. It’s more compatible with agile development methodologies because you don’t need wait until you’ve bundled enough features together to market it as a new version worth upgrading to. It’s in your best interest to ship new features immediately as they’re developed.

    It’s totally fair of you don’t like the model.

    But the model itself isn’t the problem.

    Shitty companies being greedy will always happen.




  • Wow… lots of people in here bashing the subscription model, but let me point out it’s maybe not as bad as you think…

    If you sell a product under a perpetual license model (I.e the one-time purchase model). Once you’ve sold the product, the manufacturer has almost no incentive to offering any support or updates to the product. At best it’s a marketing ploy, you offer support only to get word of mouth advertising of your product which is generally a losing proposition.

    Since there’s little incentive to improve the experience for existing customers. Your main income comes from if you can increase your market share which generally means making products bloated often leading to a worse experience for everyone.

    If the customer wants support, you need to sell them a support contract. If they want updates you have to make a new version and hope the customer sees enough additional value to be worth upgrading. Either way we’re back to a subscription model with more steps, more risk, and less upside than market expansion so it takes a backseat.

    If you want to make a great product without some variation on a subscription. You need to invest heavily upfront in development (which most companies don’t have the capital to do, and investors generally won’t invest in unproven software)

    From a product perspective, you don’t know if you’ve hit the mark until people start using your product. The first versions of anything but the most trivial of products is usually terrible, because no matter how good you are, half to three quarters of the ideas you build are going to be crap and not going to be what the customers need.

    Perpetual licensing works for a small single purpose application with no expectation of support or updates.

    It works for applications with broad market needs like office software.

    For most niche applications, subscription models offer a better experience for both the customer and the manufacturer.

    The customer isn’t facing a large transition cost to switch to a competitor’s product like they would if they had to buy a perpetual license of it, so you have a lot more incentive to support and improve your product. You also don’t see significant revenue if the customer that drops your service a couple months in… even more reason to focus on improving the product for existing customers.

    People ought hate the idea of paying small reoccurring fees for software instead of a few big upfront costs. But from a business model perspective, businesses are way more incentivized to focus on making their products better for you under that model.



  • The exhaust from a typical ICE wouldn’t have enough pressure to inflate a tire, so you’d need a compressor. Of course if you had a compressor you’d just use clean air.

    If for some reason you used a compressor to compress exhaust gases to fill a tire, it would mostly be the same as filling with air at first.

    Exhaust gas is mostly a mix of carbon dioxide and and water vapour, with small amounts of oil residue, and other organic compounds. The water vapour will condense as it cools likely leaving some liquid water in the tire, which won’t cause immediate issues but will cause vibrations which will accelerate wear not just on the tire but possibly the entire suspension.

    The organic compounds will cause the rubber to break down over time and the tire will wear out sooner.