Every time that a game has ambitions to be the next Doom, WoW, Apex, etc. I'm always reminded of Douglas Adams talking about how he wanted to be a member of Monty Python when he was younger. Cleese, specifically, and didn't really understand that the role was already taken until he was much older.
Oh this is delightful!
Honestly agree. I also found the usage demonstrated with "injecting" a little funky. It resembles so much the behaviour of just making a nodegroup with inputs. I have yet to mess around with it to understand the nuance.
Yeah I had a similar impression. It's easy to fall down that kind of hole. At a certain point though the attitude is its own problem.
I feel like the clincher is when you believe that you are indeed Super Special and everyone else is less than.
Goodbye social interactions, goodbye collaboration
The head of that team really went all in on showing his whole ass on twitter a while ago. Generally argumentative and "I think I'm smarter than everyone" vibes.
Gives me doubts about team cohesion over time. Even if the pitch was super promising.
This. This is the way.
Very Jar-jar Bynx coded
My wife got these knitted covers for hers, which seems to work pretty well!
So this exports as a skeletal animation? Or would it have to be a vdm?
Very cool!
I'd built something similar to this in Geometry Nodes a while back, with individual leaves (I thought Nanite was going to be better than it turned out).
In the end I realize something in-engine would be a lot better. The gap from modeling software to in-game render is too large for stylized work.
I mean from what I recall it also felt clunky and awkward.
Promising app but it's got real interface problems. I obsessively set up my Ps shortcuts to be executable on one hand of the keyboard. CSP has fewer commands but at least I can build thim with the action recorder and a custom shortcut.
Yeah so far the only people that I personally know whose work has changed noticeably is devs and students. The fuzzy finding, and language reshuffling to re-explain the same concept until it clicks, seem genuinely useful and utility that would survive the bubble as neat ways to help seek knowledge.
And, to be fair, these things are freaking hard. Knowing and commiting to what you are going to make so that you can move in a straight-ish line is **HARD**.
Now, AI seems pretty useful in fuzzy search, parsing large volumes of text, all stuff an exec probably does all day. I wonder if LLM's success for helping these roles leads to a fallacious conclusion that it must be that useful for *every* task.
The biggest cash sinks in games are challenges in direction and production.
Poor planning, lack of direction, and lack of creative conviction fueled by a poor understanding of what you are making and for whom - makes you burn dollars moving in circles.
None of those are solved by AI in production.
I've had difficulty with this too. When I first started seeing it I wanted to rip my hair out. Now that I know it won't affect expectations or how I work, I'm like "Ok have fun with your clip art."
Eager for things to get to a more sane place with all this.
Which I think gets to where I think things will land post-bubble. I hope we'll get away from the sloppy "just dump all the money/compute/water/energy and pray it'll do everything" and more towards. "We trained a ML system to tackle this specific problem according to our specific requirements".
I very much agree. For the latter group it does seem to make you more of a cyborg than reverse cyborg compared to creative work.
But even then I suspect for the latter group it makes more sense to have custom ML tooling instead of a general LLM integration. I wouldn't trust ChatGPT to analyze MRIs.
Interesting. What changed with the free release? I had assumed it was mostly UI/branding update. Accompanying the free license and tie-in to canva.
Affinity is *very* good. When I last used it, I had a few frustrations that stopped me from migrating to it.
Transforms don't visualize until you let go of the handle [full blocker]. Also if I recall correctly the keyboard shortcuts had some weird arbitrary limitations compared to Ps or CSP.
Choppy waters for sure, curious to see what this looks like 10+ years from now presuming the dust finally settles
Yeah the devs I have spoken to on this are mixed
What I've heard from devs is... Interesting?
More recent studies show consistent increase in productivity, but on the other hand you internalize and remember less of the code which makes debugging more difficult. Which increases risk. Then there's security risks.
But I'm no dev so.. grain of salt
Oh likewise! In my experience it literally has no greater or lesser impact on my work than some quick google or pinterest references (except for my own personal annoyance with the genAI). I still do my own research and explo. But, I can't control what makes the designer more comfortable ยฏ\_(ใ)_/ยฏ
I know some game designers who use ai when making a brief for a creature design for example. Never makes it to production but they find it helpful to get some visual that explains their intent.
Imho not particularly more useful than a reference board, but I guess it counts as "using" ai.
Whoah that's great to hear! I managed to get to almost ready to escape it with Clip Studio, but there's still that 1% of workflow I haven't managed to port over...
Absolutely nailed it. 10/10 art direction.
What I've been recommended by users at one point is to reply to my old post instead of repost.
So maybe reply with a tidbit about the piece (meaning, process, etc) to bump it up the timeline while keeping it "fresh"
That or find some way to prompt discussion in the thread?
Bluesky re-surfaces posts with the latest reply to the timeline, so it definitely favors conversation with back-and forth. People don't have nearly as much back-and-forth discussion on an art thread vs a meme or political post.
