Except 95% of companies have no need of ultra scalable super cloud.
If you are a very big SaaS company that is not Google or Apple, you are probably serving hundreds of thousands, maybe millions of unique users. AWS may be convenient, but you don't /need/ it, you can build an infrastructure that will handle such workload with any of the big european providers.
You'll just lose in comfort what you'll gain in data sovereignty and infrastructure costs.
I worked for a 7M€ MRR company that had maybe a million of users who used the software every day. The thing ran on a dozen of OVH servers, including multi-site redundancy.
Exactly. AWS proposition was much more alluring where compute was more expensive and it required yearly estimations and updates.
In times when one physical server can have 32, 64 or even 96 cores... you pack your own little datacenter right there and it's pretty cheap to simply overkill it, have one or two servers for redundancy and bye.
So many businesses will happily run from 4 core 10usd VPS (that would have been beefy server 20 years ago).
Just for the joke, I own the og Xbox One and it’s the only console I hated from day one.
I clearly remember plugging it to my TV with excitement and being greeted with gigabytes of mandatory updates. And then I discovered that you weren’t able to play the game from the disk and that you need to install it on the fucking hard drive !! And then I discovered that the disc reader was actually slower than my fiber connection which means it was faster to play a game from the online store than installing it from a real disk.
I think I had to wait for at least a full hour just to play my first game.
And on top of that the performance was actually not that good. 30fps everywhere, it was worse than the Nintendo games on Wii / GameCube which usually ran at 50/60fps.
I still own this shit but I never liked it. At least it was useful some month ago when I had to update my Xbox controller firmware (but since I didn’t power it on for years , I also had to wait for updates :) ).
Look at my list here [1] but I think it’s coming back. Sure the big studios are all collapsing from everywhere and extracting value from everywhere like any shitty corporations. Nintendo feels like they are surviving a little more but even them are more and more doing corporate shit.
But what I see is also happening in parallel, is that the people nostalgic from the 90s era of video games are now 30 to 40, are now senior programmers and they are determined to create another batch of truly good games.
Sure the biggest studios have the biggest marketing budget but when you read a little about them, they are just all slowly dying. Most news about big old studios are about firing thousands of people, being bought by other corporations who will also fire people.
Sure, Expedition 33 feels like an outlier, but it’s just a game from ex Ubisoft employees. Ubisoft which is sadly also slowly dying.
I think Nintendo and Sony were almost the pioneers of "corporate shit": yes, Nintendo has a bit different style of gameplay they target, but their business practices have been corporate-protectionism for decades.
That's not entirely untrue. Triple A is the current day shovelware. It's just that the shovel is made of gold and expensive.
I find my enjoyment in select retro games and indies nowadays. When I find a game I really like that is not an indie, it is typically something that is explicitly not AAA (such as Octopath Traveler).
Hell, one of my all-time favorites is a indie I olayed a couple of years ago - Ender Lilies. It became the best Metroidvania ever for me, when I thought nothing would ever dethrone Castlevania Aria of Sorrow.
So yeah. If gaming has a future for me, it is with indies.
I really didn’t expect to get a new favourite game of all time in my 30s, surely the nostalgia factor was too strong, but Outer Wilds was exactly that for me.
1) Anyone who says Celeste’s music is better than Super Mario Bros’ is a liar, and I don’t even like Nintendo games.
2) Let’s look at some of those release dates, shall we? 2019, 2013, 2020, 2017, 2017, 2021, 2022, 2023, 2016, 2013, 2013, 2008(!), 2019, 2021, and 2016.
That’s a period of 15 years. For an American, the NES released in 1985 and the PS2 released in 2000, also a period of 15 years. The fact that your “games of today” list is kind-of competing with four console generations itself is an indication that quality isn’t higher now, even with a considerably higher volume of releases.
Also only two of those games came out in the last 5 years, so things really aren’t looking great for modern games.
Except most of those games aren't "retro" because, unlike real retro games, they are mostly still updated and they work on any recent computer/console.
So for me, even for the oldest ones, they are still part of the same "era". What I mean by that is that if you buy any of the items in this list in 2026, it will not feel like it's an "old" game.
That's not what GP said, what they said is one doesn't feel _older_ than the other, not that they don't feel different at all.
That aside, is innovation (technological or otherwise) the goal of video games? If past a certain year the best games of every don't seem to be any better than any other, but just different kinds of good, to me that's not stagnation, but rather that the designers collectively figured out how fun games that are not constrained by the hardware. There's a reason, for example, that past a certain point lives systems just about went away, or autosaving nearly completely replaced manual saving.
It's interesting idea but a lot of mail "contact" is either rare but still important, or one off thing that you read and/or remove/spam.
Funnily enough in old school mail client you could trivally make directory per contact via sieve filter but mainstream mail clients don't really want to give users much power.
But that's what sorted by date means, right? When you get a new mail, it goes at the top of the mailbox, and after new ones arrive, it goes down.
My mailbox sorted by date is a total mess. Having everything grouped by sender email would automatically make it tidy.
That works for single-sender mails, but most of my work mails have almost a different sent of contacts per topic. Grouping mails by subject (topic) makes this more manageable.
In all mail clients I've used, you're 1 or 2 clicks away from seeing your unread messages only, which greatly helps with filtering what's important to read soon.
For me, the email interface works great. I treat it as a self-populating to-do list. The 'one thread per contact' wouldn't work because I organize by project/topic and often have multiple of these with the same contacts.
IM threads have a place of course, but regular email works for me and is actually an important structural element of my work flow.
Well. I own an iPhone, a Macbook, Airpods, Apple Watch. I'm in the Apple ecosystem since the last 16 years.
Unfortunately, due to their behavior in the latest years, I'm not going to buy anything Apple anymore.
Fortunately for me, I prefer Linux to MacOS so I never have been totally tied in the Apple ecosystem and I know how to leave the boat without a lot of hassle.
I'm really saddened because they know how to make great products when they want to. It's just infuriating that everything that is shitty in their products is never due to randomness or bugs or whatever, but ALWAYS because they decided to fuck you.
Though i must admit that for maybe more complex applications, I feel like aspnetcore also fulfills this definition. I feel like it’s easier to create something more complex with aspnetcore while still keeping the code boring and opinionated.
I feel like Django, for bigger apps, fall apart in the "opinionated" side. For "simple" websites, you can’t go wrong, but for anything really big, basically everyone invents their own project structure.
But don’t get me wrong, I still love Django for what it is and it’s my first love in web frameworks anyway.
And I’d go further and say that the Django documentation is so awesome, that 15 years ago, it was where I learnt how websites/http/etc… really worked.
Yep but with AirPods, you are listening music or watching a video, on your Mac, your iPhone rings (on your AirPods), you accept the call, and now the video is paused on your Mac and your AirPods are already connected to your iPhone.
Any time any of the registered devices needs to emit sound, the AirPods instantly switch to this device (and both devices will show an unobtrusive notification to reverse the auto switch).
And it works every. single. time.
Apple can't make Airdrop work reliably after decades but somehow, they are able to magically and instantly transfer bluetooth audio from a device to another device.
Though, if you use your airpods with anything non apple, it will juste work like a classical bluetooth device, with manual pairing and no magic switching.
That is a great point. Airdrop on my iphone currently has this weird bug where if I try and airdrop directly to a target (eg my laptop) it doesn't work, but if I go into airdrop and select the exact same target, works fine. This is even weirder because it's followed me between phones (I restored from icloud backup). Yet, yeah, my airpods are fine at switching.
If you are a very big SaaS company that is not Google or Apple, you are probably serving hundreds of thousands, maybe millions of unique users. AWS may be convenient, but you don't /need/ it, you can build an infrastructure that will handle such workload with any of the big european providers.
You'll just lose in comfort what you'll gain in data sovereignty and infrastructure costs.
I worked for a 7M€ MRR company that had maybe a million of users who used the software every day. The thing ran on a dozen of OVH servers, including multi-site redundancy.
reply