It's a little under 2 hours until the AMP 3 Premiere! See the event above, head to cubecoders.com/AMP3 , or get the notification from YouTube directly at youtube.com/watch?v=hH3-CwWl…
For anyone wanting to take part in the beta, now is your absolute last opportunity to be in with a chance to be included - make sure your Discord account is linked and you've enabled Product Notifications in your account page.
After the conclusion of the premiere there be a live stream immediately following it where we'll show you a bit more of what's new, answer some of your questions about AMP 3s functionality, and what's coming next. You'll want to make sure you're ready a minute or two early if possible to make sure you don't miss anything.
See you soon!
We're all ready for tomorrow. If you want to participate in the AMP 3 Beta, make sure again that you've linked your CubeCoders and Discord accounts, and enabled Product Notifications in your email settings. Otherwise, we'll see you at cubecoders.com/AMP3 or youtu.be/hH3-CwWlG6U tomorrow at 8PM UTC.
1
13
669
With memory prices for server RAM hitting nearly $70/GB, one of the things we really looked at carefully for AMP 3 was memory usage, because occupying even a single gigabyte that you don't have to literally costs money.
The result is a very efficient system that leaves more memory on the table for running the actual applications that matter to you.
19
708
AMP 2.8.0.10 has been released! Featuring significantly reduced RAM usage, sleep mode improvements, better performance in applications running in Wine/Proton and better LDAP handling. Check out the full release notes: discourse.cubecoders.com/t/a…
1
17
627
We know a lot of our users are relatively new to Linux, so we made sure that in the new GetAMP installer for AMP 3 - the first time you hit CTRL+C while trying to copy the URL to finish setup, it doesn't kill the script - the copy happens instead. Even if you do accidentally kill it, the URL is left in the terminal afterwards :)
14
745
We're just 10 days and 10 hours away from the premiere and the first stage of the AMP 3 Beta! We've updated the announcement with information on eligibility and what you can expect from the beta: discourse.cubecoders.com/t/a…
Headline details for those who want to try out the beta to be eligible for the Batch 0 keys:
You need to have owned AMP Advanced or Enterprise before the 1st of July this year (inclusive of that date)
Your Discord account needs to be linked so your exclusive role can be assigned.
Your email preferences in your account needs to have "Product Notifications" turned on so we can actually send you the invitation.
Batch 0 will be a pool of 250 keys - but if the live premiere exceeds 1000 views during the premiere itself, we're going to double that to 500 keys. So make sure you're there for it and register your interest in the event just above so you get notified!
We've spent the last week testing and doing validation on a variety of systems and setups, checking things like automatic updates, rollback handling, compatibility and general behaviour. We gained access to Mythos to do a (very expensive!) security sweep of the entire stack. Windows users in particular benefit from this because we've got some most excellent functionality come out of that to provide isolation in the absence of containers.
Getting AMP 3 built and into the state where it's ready to show off has been an absolutely labour of love over these last few years - but it's going to be worth it. As for Batch 1 onwards, the speed of that basically depends on what kind of issues come up during the beta. If everything goes as planned we'll be expanding the number of beta slots fairly quickly.
See you all November 6th! <3
One of the rules in AMP 3s development is that *every* failure condition must have some kind of human readable explanation - no generic "Something went wrong" situations. It's a lot of work authoring messages for every condition but it means that dealing with issues when they do come up will be a lot easier with only a single message per cause.
2
24
796
We've recently updated our AI disclosure for AMP 3 development to include the use of Mythos for security testing.
You don't need us to tell you it has been a bloodbath this week with LLMs discovering or producing PoCs for all kinds of exploits, and AMP of course is installed on a large number of fairly powerful systems which makes it a really attractive target. So we're using the same kind of tooling to make sure not just AMP but the game servers running inside it don't become a vector.
We're committed to aggressive proactive testing before builds reach users to help keep you guys safe, rather than reacting to things after they happen.
On container-less systems, AMP 3 still isolates each game server process within its own user. So even if one was compromised, it wouldn't be able to affect anything else on the system. This applies on both Windows and Linux hosts.
We don't use PHP for any products we ship or systems we use. If you're using any legacy panels that still rely on PHP then AMP 3 is about to offer you an excellent upgrade path soon :)
19
803
Send us your favourite "AMP 3 can/will..." memes / image macros! We'll pick a handful of our favourites and make sure they get access to the beta!
Are high RAM prices affecting your game servers?
72%Definitely
12%Somewhat
16%Not at all
64 votes • Final results 2
2
495
Make sure to check out the Live AMP 3 Premiere on November 6th! discord.gg/cubecoders?event=…
Anything that can create standard Docker contains on your system has the equivalent of root access to your entire system and everything in it. This is part of why AMP uses either rootless Docker or Podman (which is always rootless)
If you're still using a panel that creates normal containers, AMP 3 has the ability to import from a variety of other server panels so you can level up your security.
P.s. It's also a good reason to make sure you're not using something vibe-coded that has de-facto root access to your system whose developer doesn't understand how it works ;)
One of the things we will never do is shoehorn unwanted AI features into our software just for the sake of it or push it onto users who don't want it.
If we ever do find a use for AI that actually adds value in some way to our software, it will be disabled by default, explicitly opt-in, and you will never be pestered or annoyed into turning it on.
If a feature can be implemented *without* using AI to do it, that will remain the preferred approach.

