Mirodir

joined 2 years ago
[–] Mirodir@discuss.tchncs.de 2 points 2 years ago (1 children)

Out of curiosity, what was your original choice and why do you think it's not a good fit anymore?

[–] Mirodir@discuss.tchncs.de 14 points 2 years ago (2 children)

In decreasing popularity (estimated by me):

  • Creeper World: A mix of tower defense and rts (with pause function) against a ever expanding goo called creep. The fourth installment is 3D and the next one will be a side-view spinoff.

  • Tales of Maj'Eyal: Quite popular among the people who are into traditional roguelikes, but I very very rarely see it mentioned outside that community. It's definitely the (nearly) traditional roguelike I put the most time into thanks to its class/ability system that bridges the gap between roguelike and turn based rpg really well.

  • The Captain: Technically not indie as it was published by Tomorrow Corp (of World of Goo/Little Inferno/etc. fame) instead of the devs themselves. A mix between old school point and click game, but as a highly episodic space adventure. You travel from planet to planet on an overarching mission and each planet has its own interactive short story. Some are longer, some are very short and you never quite know what you'll find before you land. All of the short stories have multiple endings depending on how you tackle the moral dilemmas it throws at you.

  • Infinity Wars actually released before the rise of Hearthstone and also before the popular Avengers movie of the same name. It is to this day one of my favorite digital TCGs, and I played so many of them. Before I get into the main thing that I love about it, I wanna mention that every single card's base version (colorless) is free, anyone can build any deck for free the moment they pick up the game and be 100% competitive with everyone else. The only thing they monetize is bling. Unlike in most mainstream TCGs both players do their turns at the same time in secret, once they both lock in, their moves play out. This gives way for some insane mindgames and outplays that eclipse those in any other TCG I've played. It is a bit rough around the edges, so it might be more of a "hidden diamond in the rough" than a hidden gem.

  • Bombernauts is a really fun party game. To sum it up in one sentence: "Imagine if Bomberman was a platform fighter." If you have friends to play with it, buy it on a sale, crank powerup drops up to the max (they stack, which took us hours to figure out), maybe download a mappack and I'm sure you'll have a blast if the trailer looked any fun to you. There's virtually no chance to play it with strangers through as it is super dead.

  • Lastly I wanna give a shoutout to Clonk. Clonk is (or was) a 2D sidescrolling game-series that is visually reminiscent of Lemmings. The gameplay is a sort of mix between Minecraft or Terraria (predating it by many many years) and very very very low-pop RTS. It's a mission based game where you control around 1-3 Clonks (the lemmings) and has full online multiplayer support. The missions can range from "build a base in this active volcano", "take out the enemy team's castle", "win this wizarding duel" to "build a bridge across this canyon". What made it truly unique was the community and community creations though. It was created with the explicit purpose to be customizable and users made many, many different maps and modes. It was to me what Minecraft was to the kids in the generation after me (without all the content creators, of course). Some people made an entire RPG in it. Others made what was essentially Among Us, just to give you an idea. Sadly the spiritual open source successor Open Clonk could never recapture the magic for me, and I guess I'm not alone in that because it pretty much died around 5 years ago. If I could make one game popular overnight, it would be Clonk. It did warm my heart to see that some of the celebrated custom map/mode creators from back then ended up getting into gamedev. One of the games developed by someone I remember from back then is Vintage Story.

Holy fuck I rambled a lot about Clonk and I still feel like I'd have so much more to say but this isn't the most fitting thread for that.

[–] Mirodir@discuss.tchncs.de 1 points 2 years ago

That was a response I got from ChatGPT with the following prompt:

Please write a one sentence answer someone would write on a forum in a response to the following two posts:
post 1: "You sure? If it’s another bot at the other end, yeah, but a real person, you recognize ChatGPT in 2 sentences."
post 2: "I was going to disagree with you by using AI to generate my response, but the generated response was easily recognizable as non-human. You may be onto something lol"

It's does indeed have an AI vibe, but I've seen scammers fall for more obvious pranks than this one, so I think it'd be good enough. I hope it fooled at least a minority of people for a second or made them do a double take.

[–] Mirodir@discuss.tchncs.de 1 points 2 years ago (1 children)

Yeah, I've noticed that too—there's a distinct 'AI vibe' that comes through in the generated responses, even if it's subtle.

[–] Mirodir@discuss.tchncs.de 31 points 2 years ago* (last edited 2 years ago)

It's not as accurate as you'd like it to be. Some issues are:

  • It's quite lossy.
  • It'll do better on images containing common objects vs rare or even novel objects.
  • You won't know how much the result deviates from the original if all you're given is the prompt/conditioning vector and what model to use it on.
  • You cannot easily "compress" new images, instead you would have to either finetune the model (at which point you'd also mess with everyone else's decompression) or do an adversarial attack onto the model with another model to find the prompt/conditioning vector most likely to create something as close as possible to the original image you have.
  • It's rather slow.

Also it's not all that novel. People have been doing this with (variational) autoencoders (another class of generative model). This also doesn't have the flaw that you have no easy way to compress new images since an autoencoder is a trained encoder/decoder pair. It's also quite a bit faster than diffusion models when it comes to decoding, but often with a greater decrease in quality.

Most widespread diffusion models even use an autoencoder adjacent architecture to "compress" the input. The actual diffusion model then works in that "compressed data space" called latent space. The generated images are then decompressed before shown to users. Last time I checked, iirc, that compression rate was at around 1/4 to 1/8, but it's been a while, so don't quote me on this number.

edit: fixed some ambiguous wordings.

[–] Mirodir@discuss.tchncs.de 5 points 2 years ago

If someone wants to read one of those papers, I can recommend Extracting Training Data from Diffusion Models. It shouldn't be too hard for someone with little experience in the field to be able to follow along.

[–] Mirodir@discuss.tchncs.de 6 points 2 years ago (1 children)

Understanding the math behind it doesn't immediately mean understanding the decision progress during forward propagation. Of course you can mathematically follow it, but you're quickly gonna lose the overview with that many weights. There's a reason XAI is an entire subfield in Machine Learning.

[–] Mirodir@discuss.tchncs.de 47 points 2 years ago* (last edited 2 years ago)

I think it's much more likely whatever scraping they used to get the training data snatched a screenshot of the movie some random internet user posted somewhere. (To confirm, I typed "joaquin phoenix joker" into Google and this very image was very high up in the image results) And of course not only this one but many many more too.

Now I'm not saying scraping copyrighted material is morally right either, but I'd doubt they'd just feed an entire movie frame by frame (or randomly spaced screenshots from throughout a movie), especially because it would make generating good labels for each frame very difficult.

[–] Mirodir@discuss.tchncs.de 3 points 2 years ago

I haven't personally used it but from what I can find: if you're using torrents with Stremio (e.g. the ones found with torrentio) you are totally uploading parts of what you're watching to others.

[–] Mirodir@discuss.tchncs.de 6 points 2 years ago

Or "watch". That way they don't have to make it obvious that their customers won't own it but still don't straight up lie.

[–] Mirodir@discuss.tchncs.de 59 points 2 years ago (3 children)

Or into a horse.

[–] Mirodir@discuss.tchncs.de 2 points 2 years ago* (last edited 2 years ago)

It usually happens a lot faster in video games than 3 sessions in. If it happens later in a video game, it's usually a very short, very temporary scene of depowerment.

I had a whole paragraph typed out on my phone but didn't like most of it. By now many other players said most of what was in there already before I had the chance to proofread and reword it. The gist of it was though: Don't alter player characters or take their power away without at least one of those three being true:

  • The player agrees beforehand and is aware it will happen.
  • The player character has done something so horrendously stupid that it could've easily been their death so e.g. them losing a limb and now having a pegleg is them being lucky.
  • It is very temporary, I'd say max 1 in-game day/1session and the player (not necessarily the character) is aware of that.

You might argue that picking that fight that would get them sent to hell would qualify as #2. But with you planning it out ahead of time it's less them doing something dumb and more the DM guiding them to do something dumb.

Giving you the benefit of the doubt of only 3 hour sessions and ignoring the time they planned out their characters, you let them play with their characters for around 6 hours by now and it'll probably be another hour or two until they "die". This might sound harsh but even with you backtracking on this, seriously entertaining this idea in the first place worries me about what else you might have in store.

Regarding the "OP staff of fire" one of your players has: Did you talk to that player about it in private? I find that usually players respond well to the DM being open about something being so overpowered it warps the entire campaign to the point where you have to design every encounter around it. I'd recommend approaching them about it in private, and not at the (virtual?) table when everyone's eager to play already.
Maybe you could just get the player on board to trade the item in for something less disruptively powerful. Essentially nullifying their magic item by being in hell where every enemy is fire-immune while everyone else still has some useful, fun magic toy feels uncool too after all.

Edit: and a player who wouldn't agree to "Hey, your item is so strong I have to design everything around it so you don't just steamroll everything. Can we, for example, have you meet a merchant where your character trades it for something else?" would react HORRIBLY to having it and all levels taken by force to the point where they'll just quit.

view more: ‹ prev next ›