100%... A game made by a human using tools and methods without AI can be as bad or worse than a game made by a human using AI assisted tools and methods.
The tool isn't the problem, the wielder is.
There was always slop, Human slop.
There's more slop now?
Sure, but it isn't AI slop, it's still human slop.
And for the argument of "It's taking away from people's jobs": The artist that got replaced can now replace the studio that cut it.
That person can now be the artist, the architect, the engineer, the marketing team, the testing crew, etc. And that doesn't have to exist in a vacuum. If you pair 2-3 people together that can do that but have a different specialty....phhhhh It's a wrap.
We need to also remember that just because we can be a solo wrecking crew that we also are still social creatures and even in the AI space working together is better. ๐ป
The anti-AI people in gaming are getting ridiculous.
Since when did creativity become about the tools you use instead of what you create?
AI is giving people the opportunity to build games they could never afford to make before.
And somehow that's a problem...
Every time I post something about AI in gaming, I get the weirdest and most ridiculous takes in the replies.
You don't have to like AI, but hating a game simply because AI was used to build it makes absolutely no sense.
Judge the GAME, not the tools used to create it.
Hey bro don't diminish your milestone!
Most people don't make it past creating an actual project folder after the open Blender or Godot and realize there's more to making a game than just playing one and having some cool ideas.
Congrats! Hopefully we'll cross paths on the road to success!
All I can say is...you did absolutely nothing that warranted Unvault taking as long as it did to launch. What you DID do however is fuck over a lot of projects that faithfully stood by your side because they believed keeping their project dropping on your platform was worth it.
This was no better than any launch Stonk facilitated that ultimately succumbed to the same pitfalls every other unorthodox attempt to be "different" yields.
I hope this isn't a foreshadowing of what's to come.
One of the things I almost never see talked about when discussing any sort of development with AI is the pursuit of understanding human language better.
We're at the point where you have to be equal parts architect and creative writer to be really leveraging the most out of these models.
Do any of you ever keep a thesaurus open when you write out your prompts? I feel like it is one of the best and unspoken tools when squeezing the most meaning and intent out of every word I write.
@Berna7224 Only if you make things you yourself want to play. Building for your own enjoyment ironically will yield better results (even if the goal is monetary) than if built for what the masses ask for right now.
If it's the former, hell yeah! Do it.
Miss me with the whole "Anthropic is better than OpenAI" bullshit.
That's like not hiring an electrician because you have a favorite carpenter.
Each excels at something different and both of them working together in their own loop is an entirely different experience than using either on its own.
The biggest limiting factors with both are the people using them.
Astra as the Architect + Sol as the worker
Just like
Fable as the Architect + Opus as the worker.
Right now I run a loop with me at the top delegating work to Fable + Astra through 1 .md file and a .py watcher that alerts both Astra and Fable when the .md file as been touched.
Both Fable and Astra pick the relevant model type for the task they need completed.
I run this around the clock on the $100 plan of each and almost never hit my limits.
I'm honestly amazed at how often Fable and Astra call the lower tier models like Haiku and Terra. And the results are on par or better than if I used just the higher models for a fraction of the cost.
I think we're at a point where our interjections into what we THINK is "optimal" are what cost us a lot of these wasted tokens.
I could just be full of shit though!
Astra as the Architect + Sol as the worker
Just like
Fable as the Architect + Opus as the worker.
Right now I run a loop with me at the top delegating work to Fable + Astra through 1 .md file and a .py watcher that alerts both Astra and Fable when the .md file as been touched.
Both Fable and Astra pick the relevant model type for the task they need completed.
I run this around the clock on the $100 plan of each and almost never hit my limits.
I'm honestly amazed at how often Fable and Astra call the lower tier models like Haiku and Terra. And the results are on par or better than if I used just the higher models for a fraction of the cost.
I think we're at a point where our interjections into what we THINK is "optimal" are what cost us a lot of these wasted tokens.
I could just be full of shit though!
@supahvee1234 Bet this guy gets mad when he sees carpenters using nail guns instead of hammers. Or mad when they use hammers instead of rocks.
These are the people that will be come irrelevant. Keep making, sir!
Judging by how poorly people tend to express their ideas on social media, I'm willing to bet a solid 90% of people's complaints about ChatGPT or Claude are user error.
I'm not as concerned with it in the normal gaming world as I am with the web3 gaming scene (if we can even call it a scene honestly). Browser based three.js games that keep users connected to a browser just introduce all sorts of attack vectors relating to wallet security.
We already have seen countless larps on Base and Robinhood over the last couple of years. And with the trend moving to gaming I can only assume web3 is about to get eaten alive with a new wave of disingenuous "devs".
@crazycarsnft@Berna7224 Things that were built with passion and a genuine eagerness to create a specific experience will always stand out, regardless of the tools used.
There's just as much good as there is shit. Just a matter of what lens we use to look at things!