Great work. I believe used M1/M2 machines will be favored by young developers as their personal fun laptop in a few years, like the Thinkpad T420 used to be. For different reasons, of course.
Do the M4 and M5 GPUs also change a lot from the M3? I hope it's not too much work to get those going once M3 is usable.
> I believe used M1/M2 machines will be favored by young developers as their personal fun laptop in a few years
I doubt it. For one, the SSDs have limited lifespans, and are soldered on the mainboard. They'll be fine enough for the planned life of the laptop, but eventually secondary market laptops will start seeing waves of failures, at which point people learn that purchasing one is a gamble.
The entire Apple silicon lineup is designed for limited lifespan.
It’s really a shame. May last “favorite” MacBook was from 2013 where everything was upgradable. I bought the fastest Core processor with the lowest everything else and upgraded to 16GB of RAM, SSD (granted at SATA speeds) and a second data drive in the optical drive bay. What luxury!
SSD can be resoldered and that service is actually becoming popular and inexpensive. It's not just MacBooks, nearly all laptops have SSD and RAM soldered. This will become a totally normal thing in a few years from now.
Soldered storage is extremely uncommon for laptops not from Apple. You pretty much only find it in very low-end Chromebook type hardware that's using eMMC for cost reasons, and a small fraction of more expensive Qualcomm-based laptops that use UFS for no good reason. All mainstream PC laptops use M.2 NVMe storage.
Neither the ram nor the ssd is on chip. The ram is on package, the ssd is on board.
On chip means literally on top of the silicon, like how AMD X3D cpus mount the SRAM chip. On modern Apple devices the ram is mounted on the organic package substrate. The difference is significant, and it's shitty that Apple outright lied about it.
I think that particular definition of "on chip" is entirely your invention. I've usually seen it broadly used for anything on-package, whether it's on-die or on a separate die within the same package.
"On chip" definitely does not have much if any history of referring specifically to stacked dies with TSVs, because that has been a very niche packaging technique until recently, and "on chip" is a much more broadly used term.
S**, I haven't felt much urge to upgrade from my 16GB M1 Air and I even use it to play some Windows games under Crossover. Quite possibly the best laptop I've ever owned.
Public information seems to describe the M4 GPU as mostly a performance-oriented refresh of the one from M3. M5 has brought bigger changes, not least neural/tensor accelerators on chip.
I just like the build quality and they are reaching the 200€ threshold on the used market. I bought one with 16GB RAM and a small black strip one the side of the screen (don’t bother me) for 230€ last week
>they are reaching the 200€ threshold on the used market.
Where?! I just cheeked the used market in Austria and 2020 M1s go for at least 350 for the 8GB RAM models and 450 for the 16GB model. Your 230 for the 16gb one fells more like a rare exception but not them norm everywhere.
Damn that's lucky. I checked facebook marketplace in Austria and prices are double that of what you're showing, even on Intel macbooks, there's no M1 macs for 200 Euros, only 400 Euros and up. Same on Vinted. No 200 Euro M1s, only at 2x the price.
The ones that I saw similarly low to yours are obvious scams from scam profiles all repeating the same message in the ad.
So maybe the ones you saw are scams as well. Otherwise hungary seems to be a lucky exception for some odd reason. Maybe because people have less disposable income, IDK?
Anyway, I wouldn't spend 400 Euros on a used mac with no warranty. The point of buying an old ThinkPad for cheap was that if something broke on it you could easily swap that part yourself for cheap because it was easily repairable and the used market was flooded with spare parts. But if your used macbook dies out of warranty, then you're shit out of luck, you can't fix anything, it's 400 Euros wasted.
I bought one already so I know it's not a scam. Scams usually communicate badly and they don't want to meet you in a public space (like a McDonald's with free wifi)
Obviously ymmv
>Anyway, I wouldn't spend 400 Euros on a used mac with no warranty.
This I agree with. I still prefer Thinkpads too but these M1s are also pretty good in almost every sense except for repairability
Apparently there's changes to boot that are more or less understood, but require some heavy work to handle.
Basically starting with M4 you have a choice between starting with Apple's page table monitor already running in their guarded mode extension, or all apple extensions disabled on the CPU cores.
> There's an underlying assumption that server-side code is inherently good, performant, and well crafted.
I didn't read it that way. I believe the underlying assumption is that the server-side code won't run in a power-constrained computer, thus having more performance headroom.
Haxe has a really elegant solution to this in the form of Abstracts[0][1]. I wonder why this particular feature never became popular in other languages, at least to my knowledge.
Yup, the way you interact with MongoDB collections in Modelence is via Store, which has a Zod-compatible schema, enforced at build-time and pre-deployment, instead of runtime (since at that point it's too late).
How does your framework compares to Meteor.js? I see similarities in the problems being solved, and the tech stack being used. Do you have examples of the idiomatic way of client/server communication in Modelence?
I think the line between the framework and the AI code generation tool is blurry.
We've been one of the very early Meteor users, since 2013 (our previous startup is featured on their landing page). After about 10 years of scaling on Meteor & Galaxy, we ended up moving Meteor into our own custom AWS cloud because of lack of observability.
As for the framework, we always wanted to have things like built-in config management, cron jobs, and better live data support (pub/sub was too rigid) - Meteor was actually a huge inspiration in creating Modelence.
Alternatively, work on developing protocols for game launchers instead. Get the Heroic Launcher devs and devs from other launchers to work on a common interface.
It's intended for the other direction of other launchers (or third party integrations with other launchers) feeding data to GOG Galaxy, but it's still one of the more interesting attempts in the wild of a launcher trying to be a little bit more than just a walled garden.
I don't know if in an Official Linux port of Galaxy if they'll try to find more ways to integrate beyond what they've already done with their Python API and how much they would be willing work with other launchers, especially Heroic, but of the big game stores, GOG seems one of the few that actually wants to try. Maybe they will. It would be nice to see. It's interesting seeing so many comments assume the worst of them, as someone who has played around with that Python API a little bit. (I was toying with a third-party Itch.io integration. Didn't get very far, but it was neat what seemed possible.)
I wouldn't say you need launchers necessarily, but installers/configurators maybe. Getting the directory structure and the right WINE or Proton dependencies is a bit involved sometimes. Especially when what you have are really OLD DOS or Windows installer files.
Their website is indeed out of date. Reminds me of Haxe in that aspect. The language itself is receiving significant development, but the website looks abandoned, and no new blog posts have been posted in a while.
Do the M4 and M5 GPUs also change a lot from the M3? I hope it's not too much work to get those going once M3 is usable.
reply