JoeCoT
@JoeCoT@kbin.social
- Comment on If we can use hydrogen to power electric motors, why can’t we use water to run a car? 9 months ago:
Correct. Green hydrogen is expensive and energy intensive, and is not as cost effective as getting it from natural gas. So currently most hydrogen comes from natural gas.
But, unless we find ways to make batteries without rare earth metals, we will be better suited to moving towards fuel cell, once we have the excess electricity from renewables needed to split hydrogen from water. For now, batteries are the better option.
- Comment on If we can use hydrogen to power electric motors, why can’t we use water to run a car? 9 months ago:
Correct. Splitting hydrogen from water is quite energy intensive. Burning hydrogen into oxygen to make water releases energy, but not as much energy as it takes to split the hydrogen off in the first place. The reason to use hydrogen fuel cells is that the extra energy needed to generate the hydrogen is still far better than the carbon output and costly materials needed for making and charging a battery. Batteries need rare earth metals, and they lose their charging ability over time. Splitting water into hydrogen creates "potential energy" from the later creation of water again, making it a useful, clean way to store electricity.
Same as the plans for using cranes stacking concrete bricks to store electricity. It takes more electric to stack them than is produced by unstacking them. But it's a clean way to store potential energy, and far more efficient and sustainable than a battery.
- Comment on Dude, where’s my self-driving car? The many, many missed deadlines for a fully autonomous vehicular future. 9 months ago:
Maybe instead of trying to train an AI powered car to deal with the insane chaos that is the road system, what if we designed something to remove that chaos? Maybe like a path that's just for these self driving cars. There's a network of paths to get you to your final destination.
But if we did that, there'd still be our current problems of running out of fuel, or battery power. Which could be solved by electrifying those paths.
But it'd be very difficult to have each of those individual cars switch between paths. Maybe it would be easier if instead of the cars switching paths, the people switched paths. Maybe we just make really long cars, and numerous people can get in them, and then switch cars as needed. People would need to know where to switch between these long cars. So we'd want to set schedules of when they're running to where, and then have an app or something that just told you where to get on and off.
And if they're really long, maybe we could kickstart this before we have self-driving abilities anyway. We could just have one person in the front driving it.
And maybe to reduce the need for rubber, instead of regular wheels on a road, they could just be metal wheels on metal tracks.
Just throwing some ideas out there.
- Comment on Windows 11 24H2 to enforce hardware requirement - gHacks Tech News 9 months ago:
sure, but not having POPCNT means way older than not having TPM
- Comment on Beeper vs Apple battle intensifies: Lawmakers demand DOJ investigation - Android Authority 11 months ago:
Because their practices are anti-competitive. School kids are getting bullied for using Android phones because they're "green texters" in iMessage. But most importantly iMessage's connection with SMS causes all interaction to be very low quality images and videos. And when people complain to Tim Apple about the experience, his only response is "Get your grandma an iPhone". Our only saving grace is that the EU is requiring Apple to support RCS, which should solve these issues, except they'll probably find some new way to be anti-competitive about it.
- Comment on Everyone brings something 11 months ago:
Thanks for noticing my fear of abandonment. It's from being abandoned so many times.
- Comment on 'We don't have teachers' | This Austin private school lets AI teach core subjects 1 year ago:
So the premise of the Dune series is the Butlerian Jihad, where humans destroyed all "thinking machines" and declared that no machine would ever be made in the likeness of a human mind again. That's why everything's analogue, humans that can do computing in their head, etc.
But unlike what one might think, they didn't destroy thinking machines because AI robots had taken over (though his son Brian Herbert missed that memo). They destroyed thinking machines because, after humans had created AI, they were happy to offload any and all responsibilities and decisions. Humans turned to AI to make any decision, and at a certain point AI ran the galaxy, not because it had taken over, but because humans couldn't be bothered. They stopped learning, they stopped innovating, they stopped doing the things core to being humans.
So as I watch humans hand over more and more tasks and control to AI, apparently including teaching their children, I expect we're heading to the same crossroads at some point.
- Comment on What Cloud storage solution do you swear to? 1 year ago:
- Large files I don't care if I lose (perhaps videos of popular things): NAS. Hard drives are cheap, not worry about losing it, I can download it again if needed
- Storage with frequent access and security compliance: Wasabi. $6.99 per TB per month, free egress. Compatible with S3. SOC2 and PCI compliance. I use this for work as a backup to S3 for website images.
- Files I need to store cheaply, redundantly, and access often: Backblaze B2. $6 per TB per month for storage. You can download 3x the amount of storage you have per month for free, or connect Backblaze to a CDN partner like Cloudflare for free egress through them. It's also AWS S3 compatible, so you can just the AWS SDK/CLI or tools that work with AWS S3. I use this for hosting image files for my Mastodon server. Note that Backblaze B2 also has SOC2 compliance and US region available now, so it should be as secure as Wasabi at slightly lower cost if you don't have a ton of egress.
- Cheap long term backup storage: AWS S3 Glacier. $0.0036 per GB per month (so $3.6 per TB). Upload your files to S3, and add a lifecycle rule to migrate them to glacier. Glacier is cold storage, extremely cheap and great for a redundant backup. I use this for backing up photos and other files I'm going to want to store forever.
For anything I'm hosting, multiple backups. Home NAS is usually the first backup, followed by cloud storage. So if I need something now, I can get it from my NAS. If there's a problem with my NAS, I can get it from cloud (though with a delay for Glacier)
- Comment on What would it take for you to move away from Github? 1 year ago:
It's hard to overstate the psychology behind the github profile. As a developer, your github profile shows that you're actively developing, whether it's for open source projects or for work projects. My previously company used a private gitlab install, which meant only my open source work showed up on github. My current company uses github, which means my profile shows green all the time.
We're a small company, but the github costs are a drop in the bucket. As others have said, it'd take something truly federated, or a crazy price jump from Github, for me to consider moving. It's free for my open source projects, it's a small amount for my company, and I have a public profile I can point to whenever I'm discussing my development.
- Comment on Introducing ONCE, a new line of software products from 37signals: Pay one time, own forever. 1 year ago:
Right. I have boxes full of software I bought once, and I have the license to use it forever. But it's for Windows XP or older. I'd need emulators or WINE to run it now, and it's not really worth it. For some of it I even paid for a "lifetime" of updates, but that stops working out when they stop updating it. I apparently live a lot longer than 90s and 2000s software companies. Just let me pay for major versions again with a guarantee of updates for X years, and price it according to those expectations.
37Signals is the company that made Basecamp, so presumably they are writing web software that would often be SaaS and letting you host it. So it's great that you'll be able to get it for one time purchase. But it definitely needs updates, as libraries change versions, new security flaws are uncovered, obviously for bugs, etc. Buying web application software is only as useful as the length of the updates included. Them providing the source is better, but since that's not open source exactly a community couldn't really work together to continue updates themselves.
- Comment on Please don't repeat the same mistake as Linux 1 year ago:
Also because the mainstream manufacturers don't want to have to support Linux.
There is less hardware support for Linux than Windows on laptops -- largely because very cheaply made components just have their firmware loaded into them by the OS when it starts, and since they're largely proprietary firmware they conflict with open source licenses.
Linux laptops are just flat out more expensive to make, because you have to use more expensive components that don't do that, confirm compatibility, and have everything setup before you ship it. Also manufacturers don't preinstall bloatware because they feel like it. It's because they get paid. The kickbacks for preinstalling bloatwave well exceeds the cost of the Windows license.
So preinstalling Linux is more expensive component wise, support wise, and bloatware wise. There's little reason for companies to do it, unless they're trying to court software developers. Dell and Lenovo and others court software developers quite well. But there's little incentive for them to try to increase Linux's market share.