Divinity video game from Larian - may use AI trained on own assets and to help development

[Cergorach]

What's your point?

My point is that we need to question if what the fans like about Larian is BG3 or DOS1/2, not actually the studio, or the people behind it... I could easily see that when the game is released, and droves of BG3 fans don't actually like it, they'll blame AI or even the lack of AI... People are fickle!

 

[Ruin Explorer]

My point is that we need to question if what the fans like about Larian is BG3 or DOS1/2, not actually the studio, or the people behind it... I could easily see that when the game is released, and droves of BG3 fans don't actually like it, they'll blame AI or even the lack of AI... People are fickle!

I don't think so mate.

Unless there's some kind of huge and obviously AI-looking issue with Divinity (like, the art is really bland and overdetailed, and the dialogue is really space-y and odd), I don't think people will blame AI. Using AI openly like they are does lead to the risk that people will wonder if problems with your game are AI-caused.

However, I suspect the problems with Divinity will probably be distinctly not AI-caused, and based on all previous Larian games I think we can pretty easily guess what the major complaints with Divinity are likely to be:

1) The setting is lame and empty generic grimdark fantasy.

And that's also got an unfortunately flippant tone and has some really clumsy proper-noun stuff that kind of crushes the vibe rather than playing into the vibe (as most of the FR's proper-noun stuff does). There's really nothing to recommend Rivellon as a setting over, I dunno, literally making up a new setting (even a grimdark one). It's bad generic fantasy made worse by multiple other games using it to make bad Belgian jokes.

2) The characters are all total bastards who no-one really likes.

Now some people will like them - some people like the absolute vilest grimdark characters - but this is likely to be a major issue if they follow the DOS2 path particular. DOS2's characters are about 20-30% as likeable as BG3's characters. And note that when BG3 launched, BG3's characters were down towards that end of likeability - but huge pushback from the Early Access audience got them to change direction. As they probably won't do EA this time, and will get a lot of BG players, who expect likeable, even loveable characters, people are quite likely to be utter repulsed here.

3) The game is totally unbalanced and has confusing mechanics.

Mechanics that don't fit the setting or story very well at all, or even create "ludonarrative dissonance", and are themselves pretty complicated and weird. Certainly the case with DOS2, fully expect it to be the case here. Again, some people will love this because they only play for five hours in co-op or whatever and it's really funny to watch fire spread or explode barrels next to shopkeepers or something. But the sort of people who want a BG3-style experience are likely to be... very disappointed.

I expect it to get very positive reviews from a bunch of reviewers afraid to do anything else, and/or who are all very into grimdark and/or who are only playing it on co-op, sell extremely well initially because of BG3, and then drop off really hard due to word of mouth.

On the other hand, if they do go for Early Access, I think they'll be fighting the players, most of whom will be ex-BG3, every step of the way, but we'll likely end up with a game that's a hell of a lot better and more suited to a broader audience (and also less juvenile, frankly).

not actually the studio, or the people behind it

Which people though?

Only 100 of the 400 people at Larian were there for DOS1 or earlier. The rest were hired during or after DOS2's development, mostly during BG3's.

Neither of BG3's lead writers worked for Larian until after DOS2 had been released. So I think you can absolutely like "the people behind it" without liking the awful earlier Larian games. Just not Swen. But I've pointing this out for like, over six years at this point. Maybe a decade. Swen is a weird dude with many weird and frankly bad opinions (and a few good ones, but which he apparently doesn't hold very strongly). Those opinions got moderated in a helpful way when doing BG3 because he had to conform to the FR and D&D, rather than just his worst instincts. With Divinity it's extremely clear he didn't learn ANYTHING from that (see the recent Reddit AMA this thread stems from).

 

[MarkB]

From their AMA on Reddit:
View attachment 426854

They listened to fans, and there will be no concept art from generative AI.

So it looks like they will be using AI to quickly iterate to try things out and refine ideas, and if they use generative AI for art it will be trained only on assets they own.

I know there's been a lot of talk about what the studio that made Baldur's Gate 3 will do next, what do you think of this?

I guess they were persuaded not to commit a Divinity Derivative Sin.

 

[aiouh]

This is a tired argument. Look at the societal disruption around the Luddites, then look at the scale in which AI, if all goes to plan, will replace work.

If you want to end up a Government slave to UBI, that's on you.

So you're both overestimating the usefulness of AI (AGI replacing work, which it won't) while also being anti-AI? Maybe it is you who has drank the technocrat kool-aid they distribute to capture more capital attention.

 

[trappedslider]

Can't rage against the system here, so rage against the most recent symptom. 5 years ago it was block chain and NFTs, now it's A.I. I wonder what the next will be.

 

[Umbran]

Such as the weather prediction that benefits people and the environment? Thanks, you reminded me that we already have examples where what you are saying is incorrect.

Well, not really.

If you really want your argument to be taken seriously, it should show (not just assert, but demonstrate with evidence) that the benefits not only exist, but that they outweigh the detriments.

Given the list of detriments, that seems a tall order.

 

[Umbran]

ROFL. You've got this one nothing argument that you keep repeating.

Nope. One tiny counter-argument proves nothing. It's worthless. You're pissing in the wind. Get a real argument.

Mod note:

Hey, RE?

You can have this argument without breaking Wheaton's Law in the process. The site rules pretty much require it.

So, be better to people, or walk away, please and thanks.

 

[Blue]

Well, not really.

If you really want your argument to be taken seriously, it should show (not just assert, but demonstrate with evidence) that the benefits not only exist, but that they outweigh the detriments.

Given the list of detriments, that seems a tall order.

The argument was that they could not provide a major societal benefit nor that they could benefit anyone except oligarchs. I gave an example to show both were incorrect. There's value is to call out that absolutionist assertions are factually incorrect.

There are other examples but I'm not trying to convert them to pro-AI stance. This is just a side portion of the discussion if AI will stay around, which was my original point.

 

[Blue]

This is a tired argument. Look at the societal disruption around the Luddites, then look at the scale in which AI, if all goes to plan, will replace work.

If you want to end up a Government slave to UBI, that's on you.

Yes, an argument made time and time again as technology provides yet another step of efficiency can be describes as a tired argument.

Yet the world has continued.

It is definitely a time-worn argument. However, all that does is provide plenty of historic examples that it's true. If you disagree with it, please don't try to handwave it away with rhetoric but instead honestly debate it.

 

[Ruin Explorer]

Yet the world has continued.

And many times it's become horrific and nightmarish, but you seem to be discounting that. For the last 60 years we've been a few button presses away from the world very much not continuing in any sense that matters.

Again I note you haven't addressed the environmental destruction/climate collapse issues hard-linked to GenAI (because it's insanely power-hungry). It's fun that you're keen on it predicting the weather better, when it's going to cause the weather that ends civilization.

instead honestly debate it

Given you're actively avoiding debating major issues here, I don't think you're in a position to talk like that, I'm afraid. You might want to rephrase.

they could not provide a major societal benefit nor that they could benefit anyone except oligarchs. I gave an example to show both were incorrect

So you consider slightly lower power usage on climate models to be a "major" societal benefit? If so, the societal harms being done by stuff like the lie machine and the CSAM machine are far more than "major", they're all-encompassing, because the scale of harm they're doing is insanely larger than a few MWs here or there on a climate model.

As I pointed out, the amount of wasted energy from one hour of drivel from Claude (like, if we excluded any even semi-legitimate usage) is probably greater than the power saves on your example. But you've kept failing to link to evidence for your example so we can't even know for sure.