AgenTic, not Agenic. As in, an AI that acts as your agent. Meaning, the goal would be to have an AI model that you could direct to perform certain tasks in the background while you focus on other things.
For example, youre in the middle of doing something or another when you remember that your oil change is due. You pull up KraftonAI and tell it to “book me a service appointment for my vehicle at the dealership this weekend”. The AI proceeds to work on that task in the background, only prompting you for input if it meets a road block it doesn’t understand.
I mean, that’s the goal of all of these AI companies. If you peruse any marketing material for Google, Microsoft, ChatGPT, Grok, etc. they almost all mention the “agentic capabilities” of their flavor of spyware.
Personally, an AI model which is capable of doing tasks like this would actually interest me. However, no organization (for profit or otherwise) is trustworthy enough to have access to all of the data on me that it would take to make an agentic AI actually useful, so, for me, it would have to be something I run locally. However, rather than invest all of the time, effort, and money into learning how to make that happen, I think I’ll just call the damn dealership and schedule an appointment. I may suffer from terminally online brain rot, but I’m not so paralyzed by human interaction I can’t make the occasional phone call.
$70 million on gpu cluster + $21 million yearly on employee AI tools and training. They could fund 21 $1 million indie games per year with zero expectations, but no push the money to the slop and get nothing in return. Fuck them.
Not saying krafton deserves the benefit of the doubt but:
Understand that “agentic AI” is almost entirely a buzzword that means “Microservices with an LLM somewhere in the mix”. Which… is what people are already doing.
Yes, there are some (idiots) who think that means EVERY single node in the graph needs to be an LLM and fuck the planet, Jensen needs a new zipper. But, by and large, what that means is they are using the exact same infrastructure they were last week but MAYBE added an LLM for preprocessing or postprocessing. It makes management happy because “We are using AI” and it makes everyone else happy because they can keep using the tools that actually work.
Glad to hear this. Years ago I worked with Obsidian engineers (who may or may not still be there, I have no idea) and I absolutely loved the systems they had built for their games. They were able to create games with a massive amount of content with a relatively small team of engineers, so I’ve always rooted for them.
I liked (didn’t love) and completed the first Outer Worlds but I thoroughly enjoyed Grounded - even though I never played through to the end game. I’ll definitely give this one a go.
Good news! You’re allowed to skip them and instead play the hundreds to thousands of older games you’d surely love, but that slipped under your radar at the time. Plenty to hold us over until publishers and studios learn their lesson.
There will always be passionate developers eager to put down the work and make good games. There will be a lot of shit they will try to shovel down our throats, but it’s just as easy to ignore it.
I think it will go the way of the NFT. People who don’t understand tech will hype it beyond belief and then the actual developers will go “this is useless” and not use it.
Well, maybe not exactly like NFTs because NFTs were actually useless while AI looks like it might have some actual niche use.
honestly for most people i feel like AI has already a solidified use, “the magic thing that answers all your questions and generates pretty pictures*”
compare that to NFTs which had strictly no use for the average person. i think what we’re seeing with AI is quite different
will the hype maintain itself when the AI bubble bursts, the VC money dries up, free options get removed and paid options significantly ramp up in price? that remains to be seen
*(the fact that the answers and pictures are often garbage is irrelevant. as long as they’re good enough often enough, it has value to many people)
Once the VC funding is gone and AI companies need to become profitable, I don’t see myself as an individual paying for some LLM, at best I’d try to setup a free LLM like deep seek on local hardware, but maybe corporations might pay for AI access for their employees if they think it provides some benefits, even if that’s not really true in the long run
pcgamer.com
Najnowsze