I'd never go x86 again after owning an m1. I'd have replaced an x86 laptop 2 or 3 times by now (2020 m1). My last $3500 dell xps 13z before buying the m1 was absolutely horrible.
Very bad battery life. Maybe 8hrs of reading with wifi off, BT off and only low screen backlight. I need to charge it daily or every other day at best.
Physical buttons are comically bad and often don't register presses.
Built-in book reading app is slow, lacks some critical features and hangs in many epub books. I replaced it with KOreader which finally made the device usable.
It's a great screen size though (9" would be even better).
I got about the same - then I got a power-cover and it lasts far longer, it was a worthwhile investment. Although after about 2-years, the pogo-pins are wearing out and sometimes it doesn't connect/charge.
The physical buttons were also great - for the first year, then they started wearing out.
> Very bad battery life. Maybe 8hrs of reading with wifi off, BT off and only low screen backlight. I need to charge it daily or every other day at best.
This seems like a wildly different experience than I have with it. I don't read 8 hours per day, but typically 1-2 hours per day, and charge it every 2-3 weeks.
Yeah, I mean I'd be interested... but you've got wifi/BT off, so I don't even know what would be taking up much battery otherwise.
My light levels are probably 50/50 6% (at night in bed) and ~33% (other times). I did briefly buy a Kobo remote and found that having bluetooth on for that was very bad for my battery life.
I can't and don't want to maintain all software I use. I need to be picky.
There's a big difference between putting money or time into a community of builders and learning in the process (myself or growing others) and maintaining a vibe coded project. That's why I want to know right away which projects are vibe coded so that I don't put any effort into them.
You're free to apply different criteria to software you use, but we're talking about having information to make an informed decision here.
I am actually hoping this is a future where it does become possible. At least on top of what is generally more-than-good-enough (Linux kernel, base GNU system with the compiler toolchain, basic desktop environment...).
So I'd like to have fully reproducible builts for a dozen or two apps I use seemlessly integrated into my laptop and phone update mechanisms. And yes, my smart TV, router, ereader and note-taking apps...
However, infinite forking also means harder to benefit from others' improvements.
I'd definitely be interested in moving towards such setup if I could reliably and easily accomplish it. There are already people who vibe code all the tools they need.
Personally, so far I got best results by carefully selecting a small number of apps and devices with good track record and community around them (think Emacs that's been around for decades, but also KOReader, Linux, AntennaPod on Android and many others) and supporting those in various ways. I might vibe code an extension or config for the apps, but then I still understand what it does and why.
AI is yet to change that. I use LLMs daily at work, but software maintenance remains problematic. Producing stable software in unstable times is actually what pays my bills.
Edit: And yes - if you fork, upstreaming is a problem. It's unfair for original software maintainer to need to review, test and integrate vibe coded code.
> They claim first tap to pay will happen later this year.
Big if true. I guess the main problem is, would the banks from all over the world join in?
Fidesmo (https://fidesmo.com/consumer/fidesmo-pay/) has managed to sidestep this by integrating with Curve (https://www.curve.com/), which issues their own card and then charges your bank’s card from their end when you pay with theirs (tokenized and emulated by Fidesmo).
(Fidesmo also integrates with a whole bunch of banks directly, though mainly EU.)
Unfortunately NFC payments via bank app is not a thing in the US to my knowledge, and I've not heard of any plans to start doing it, but good for Europe.
I've actually been thinking it'd be nice to have something like GitHub, but only for hand-written code. If anything, for now I'd like to learn primarily from and perhaps contribute to projects that have human involved at all stages.
But I don't think it will be easy to detect AI written code especially with frequent releases of new models.
It does and it's unfortunate. Perhaps Codeberg might try to block AI crawlers access to the entire website. I know it's all jazz nowadays, but I'm not in favor of using other's people work to train AI without explicit consent.
> Our lightest ultraportable, the Lemur Pro, now comes in a 14" and a brand new 16" option with 2K display, weighing just 2.2 lbs and 2.96 lbs. Enjoy multi-day battery life, the latest Intel® Core™ Ultra processors, and Linux engineered from the ground up.
That's what I'd probably do, but I'm a software engineer and devops person that also likes to tinker, so I like to have a lot of packages available. Fedora with its 80k packages (~30k apps) has been a blessing.
In comparison, Arch official repos only have 15k packages (~10k apps). There are ways to plug the gap (such as compile missing packages, add Nix package manager), but it's even better if you don't have to.
Yup. Rust tries to make it hard to do systems programming poorly. Zig tries to make it easy to do systems programming well. Rust pushes you into certain (good) practices. Zig makes it exceptionally easy to understand what's happening, making good practices the obvious choice.
How crazy is it to have a language where combining two foundation traits is such a huge footgun that the linter has to warn you (but somehow you can't ban it within the language)
I wonder if true transaction cost even reaches the 0.2-0.3% or could be further lowered.