Joined 33w ago. Seen 3d ago.
i often either feel im not programming enough, or, not reading enough. have yet to find a middle ground i like
work / life balance is a sham.
It's possible to have private messages on Subreply. They can be encrypted using the password hash. Changing the password will fail to decrypt old messages.
Why not destruct a message after logout ? Passwords are not control mechanisms. Sometimes you have to change your password for security purposes. An attack on an account could result in the loss of the important information. Better to just delete everything in a clear manner.
It's going to happen. I just registered the remote roast club! Stay tuned.
do we need a ticket? they are all unavailable
> Many times apple will replace your computer
> Many times Dell will replace your computer FTFY
I ssh to the server, open a tmux session with vim and other CLI tools and get coding, the only apps i use on my laptop are the browser and the terminal. As for editing, of all the people i know, most of them do their editing either on a custom built desktop PC or a desktop Mac, and even those who edit on a laptop they all use external monitors and external peripherals, so it's more of a desktop thing than a laptop, but yeah, why not. btw, i'm out, can't stand the char limit.
Yeah you're basically not the target market for an M1 mac
A long lasting battery is doable with any modern ARM chip, with the advancement made in the recent years, batteries and CPU's are becoming more efficient, apple is ahead but others will follow, ARM is the future for small devices. Can i ask you why you want it faster? I mean, the best way to solve such an issue right now is to have a server somewhere (or build one) and delegate all power hungry stuff to it and worry less about having a faster laptop and a lasting battery!
A lot of processing today is handled locally, what server based applications you are using everyday? These are consumer & prosumer devices - think offloading 100GB+ of video data and editing on the go being one of the more "intense" activities. "A long lasting battery is doable with any modern ARM chip" - except macOS is the only "full feature" OS available on an ARM chip so until someone else steps up to the plate, we have CPU packaged ram.
Software doesn't have the same constraints as hardware, putting them into the same box is a big stretch, software was made to remove hardware constraints in the first place, all software is modular, either in its architecture (the OS, the Web...) or in it's composition (OOP and FP paradigms). As for hardware, whether it is a preference or not, we shouldn't sacrifice the right to repair for 2x times the speed, it is a bad compromise, welled gardens are hard to escape from.
Right to repair is a different conversation IMO. All I said was I don't mind having the ram, gpu, and cpu in a single chip. If we had it your way, every computer would be the side of an ITX desktop because every chip would be on a separate PCB. I'm happy to have those computers available too - but when it comes to my laptop, I want it fast, lightweight, and long lasting. Find me a i7 windows PC with 20hr battery life and all ears.
Being modular actually takes the complexity away and puts it into different small parts, which makes it easy to understand, fix and develop, the idea that something modular is more expensive isn't true at all, PCs are way modular than macs for the third of the price, heat isn't really an issue either, you can cool down each part individually and have an overall better air flow. Also, being upgradable and repairable is much more better for the environment.
No software / hardware developer will agree with your points. If you don't care about the size of your device, are willing to utilize slow changing standards, and want to pick-and-choose parts - modular is good for you. But if you want a tightly coupled, focused product, modular is not going to fit the bill. It will be larger, there is waste in the manufacturing of boards that plug into other boards, and the heat can't be carefully managed for every different modular part.
If we can make components so fast and so thin, why don't we take the effort to make them modular? That's actually how the future is supposed to be, what's the point of progress if we all end up with thin and fast walled gardens that we don't even comprehend? I am just hoping that other chip makers won't go down that path and make wise choices for a better future and not just a faster and a thinner one.
modular in general = more complex, more expensive, larger, more heat. The future is not modular, the future is smaller, faster, cheaper, recyclable. Modular has nothing to do with it, it's an option. You can have both. Modular is only less expensive for incremental updates, it's actually more expensive in the long term.
That's not just about upgrading, you can't replace it either if it goes bad. I know that in Linux you can mark some of its parts as *bad* so that kernel won't use them, I don't know if that's possible with MacOS.
There are about 5,000 parts in a computer that you can't replace when they go bad without soldering / reworking. Many times apple will replace your computer for free even when it's out of warranty as long as you didn't damage it. Replaceable ram is mostly about upgrading, but apple computers just aren't upgradable anymore. It's a different approach, they want faster ram and thinner devices (and I do too). I have to swap the entire machine to upgrade - I'm cool with it.
Am i the only one who's not excited about this M1 chip? Ram is inside the chip, everything is hard soldered, you can't replace any component. I know it has been this way with macs lately but this time it's even worse. It seems to me that the sole purpose of those new chips is to make them more like iphones and ipads so that the only way to upgrade is to keep buying the newer models.
I really can't think of the last time that I upgraded ram in a computer that didn't involve upgrading the mobo or replacing the entire thing. If the M1 leads to better performance/watt i'm all for it. Apple is positioned to do everything that no other company could do because of "compatibility".
I don't think any of these can or will really be enforced (how would you check or even prove that other companies use dropout in their proprietory code?), but it's just ridiculous to me how those companies could even get those patents.
You can pretty much get a patent for any method as long as it's novel. like you said, you can't always enforce it.
I think I'm done with smartphones. Were I able to script stuff on them they'd be useful but I don't need an ad dongle stuck in my pocker. Soon to be back to Nokia 105. Hopefully.
I don't use apps unless they are paid. Ads on social are inescapable but you don't have to use those either. I tried the light phone (2) and enjoyed the first day or 2 with it, but honestly a phone without GPS or the internet is kinda useless. Everyone communicates with links, images, videos, articles, locations now more than ever. I don't really want to talk on the phone anyways. If anything I want a mobile device that can't make phone calls, not the other way around.
Yes sure, here are the numbers: US9406017B2, US9715642B2 and US7747070B2 in respective order. :)
Thanks, it seems they own at least one of the well known techniques. Interesting that everyone seems to know it. Makes you wonder if companies are spending a lot of time working around the patent since google hasn't committed it being open.
What's your coffee making technique?
espresso machine: 20g double shot with steamed milk. Smooth and creamy please, no bubbles - keep the steam wand just below the surface, too deep and you're just warming it, too high and it will "roar" loudly and bubble up quickly. Pour over: 25g grounds with a small hole dug out of the center, add 75g water, wait 30 seconds, add 250g water. Swirl it around toward the end of the brew so you're left with a nice even bed of dry grinds.
I just found out that not only does Google hold the patent on dropout in neural networks, but also on processing images using deep neural networks and Microsoft even holds the patent to train ConvNets on GPUs. What the hell is wrong with the patent office?!
Do they hold "a" patent on those? Is it possible that it's for a very specific use case? Do you have the patent numbers? Would be curious to see.