dirk

[Тsфdiиg @tsoding]

Understanding is the noblest of pursuits. People who actively strive to understand are pushing against uncertainty. People who give up on understanding are willingly falling into chaos.

[dirk @di_rk_]

@krishdotdev accurate

[dirk @di_rk_]

@gdechichi yeah

[dirk @di_rk_]

Cloud based is pretty unrealistic for the mass market, the infrastructure for this just doesn't exist.

There are so many limitations for cloud, like internet connections and metered play time. People will hate it, it will be like an arcade where you need to pay for play time.

It might work for some markets and game types but overall it's not really what most people want.

Back in the day arcades actually sucked and so did the games. Why is why as soon as people could game on their own devices they did and the arcades died swift and games got a lot better.

The push to digital only is something I personally support. Publishers mainly want it so they can avoid the reselling of games, which makes sense from a business perspective.

If i was a publisher or a game maker I would rather put my game on discount after a while so people can buy it that for a better price, than having middle men earn on resell so that others can get the game cheaper.

Also I think the amount of junk pysical copies create is unnecessary if the digital product is available.

It solves a lot of head aches with distribution, I remember going to gamestores in the early 2000s and they just didn't have copies of a game.

What a weird nowadays. Imagine wanting to buy a game and not being able to get it... now you just hitba button and its there in a few minutes.

Digital distribution is just vastly superior in the vast majority of cases.

[dirk @di_rk_]

Yeah maybe it's difficult to predict exactly where the failure points will be.

This time last year the consensus was that ai sucked at coding and now its essential.

Coding of course isn't the same as sculpting or texturing and so on, so it may not head along the same trajectory.

I predict that ai will get a lot better at creating assets over the next 5 years, but i am not sure how economically viable it will be. If it takes a week and many thousands of dollars worth of tokens it may not be much more viable than a human or a blend.

But who knows, Pixar spends huge amounts of money to render their movies, so maybe the neural token pipeline just becomes a different way of rendering in some cases.

Anyway, my assumption for the future is that humans augmented by ai will always beat pure ai processes, in most/many cases.

[dirk @di_rk_]

@sumanthvepa I agree to a certain degree, but what we must then conceed is that the models have an incredibly crippled conciousness.

Even though the outputs they create are highly intelligent, their conciousness is on the level of a worm or less.

[dirk @di_rk_]

You are totally right, it is not even close to being able to create high quality assets, dressing sets, animations, level design, quest design etc.

What it is good at is code, it's a great tool in that area and speeds up work for sure.

Will it get better at creating assets? I would think so, in a years time it will certainly be better at it.

The question is how economical is it?

I think maybe in a few years all you will need is concept artists, they draw cool characters and environments and the ai builds it out in 3d and gets 90% there and then you have a few more technical artists to clean it all up and polish it with neural render overlays.

[dirk @di_rk_]

Might be LPR (Laryngopharyngeal reflux), causes alot of issues with throat and upper airways.

It's pretty common, but the remedies for it are pretty confusing which usually means there are a variety of reasons why you can have it.

Some people say eating a low acid diet helps, others say increasing their stomach acid helps.

Some say diaphragmatic breathing and stress reduction helps.

I think calming the nervous system via meditation, probably helps.
The more difficult you find sitting still for 15 min the more the practice will benefit you in the long run.

I would also do a complete GI map and see what the landscape looks like and use probiotics and fiber to feed the beneficial bacteria which also helpsbto regulate the nervous system.

[Tesla @Tesla]

Real life, baby

[dirk @di_rk_]

@elonmusk amazing! so cool

[dirk @di_rk_]

These graphics look amazing, truly a generational leap.

The effort to produce this is still very high though. The GPU to render this in real time costs thousands. The neural rendering is an overlay on geometry and textures a human already meticulously crafted by hand.

Could that underlying geometry be done completely by ai? I would think yes, but the question is it cost effective?

I think it might turn out that letting the ai or robotics do a great deal of the work in conjunction with humans is far more cost effective than complete automation.

In a world where markets rule the cost of inference is the limiter on automation.

In a world where we decouple from market forces we bend reality for some time and the impractical seems possible.

[dirk @di_rk_]

That has been my point from the start.
You keep trying to "win" by misrepresenting my initial position.

My claim is that when you use Ai in the correct way - by using critical thinking - it enhances your abilities and the Ais outputs. This study shows that to be true.

It follows that if you use critical thinking when using Ai there will be no negative effects on your own ability to critically think.

Your position is that Ai use has negative association with critical thinking. Which seems to be true in specific scenarios. But in the studies you posted the conclusions where that this can be mitigated - which is my point.

So there you have it - I think my position is more than justified.

If you for some odd reason associate ai usage with right wing politics than I dunno, I think that's misplaced, your view is probably too US-centric.

I think the far larger concern is that Ai use gets so regulated that you cannot use critical thinking with ai anymore and that you need to accept their answers, that seems like a dystopian nightmare in my view.

Anyway I think that is enough back and forth. Have a good one.

[dirk @di_rk_]

a pre-print article is slop - yeah right - this is the same paper, just nicer format and published.

Makes sense that using critical thinking gives better outputs right?

The conclusion of the paper is that there is methodology where you can use ai in a way where not only do you get better results but your cognitive abilities are not harmed.
This mirrors my own experiences as well.

[dirk @di_rk_]

@KrissyQuark @antoineducros @FedericoTenga Even the cherry picked research you post draw the same conclusions as I do. I already agreed with you that there are negative associations in some scenarios.

Here is more proof that with correct handling critical thinking is maintained:
https://t.co/JQTNBP5AGV

[dirk @di_rk_]

That was a pre-print of an academic paper reporting six studies, led from Singapore, by researchers at Nanyang Technological University and Singapore Management University. Co-authors are at Purdue University (United States) and Monash University (Australia).

Here is the journal link:
https://t.co/nzEX3e8QG1

[dirk @di_rk_]

Funny how the conclusion of the research makes exactly the point I have been making to you:

"These insights contribute to the growing discourse on the cognitive implications of AI, suggesting that educators, policymakers, and technologists must work collaboratively to foster environments that balance the benefits of AI with the development of critical thinking. Future research should explore strategies to integrate AI tools in ways that enhance rather than hinder cognitive engagement."

[dirk @di_rk_]

@KrissyQuark @antoineducros @FedericoTenga You sure about that?
Computers in Human Behavior Reports, 2026 (six studies, over 1,300 people). Critical thinking about AI output, such as verifying and reflecting, was positively associated with how often people used AI.

https://t.co/4OJ5N5YMHY

[dirk @di_rk_]

@KrissyQuark @antoineducros @FedericoTenga Limitations: This study’s limitations include its reliance on self-reported measures and the potential for sample bias.

Not a great study. Far from empirical.

[dirk @di_rk_]

this is what the meta study has to say about critical thinking in particular:

"2.2.1. Critical Thinking
Research indicates that Gen-AI, leveraging its high interactivity and scaffolded learning features, can effectively foster students’ critical thinking. Guo et al. (2023) found that embedding AI chatbots in high school debate activities significantly enhanced students’ critical thinking, particularly in argument construction and evidence evaluation. Zheng et al. (2024) further reported that in collaborative learning environments, students strengthened their evaluative and reflective abilities by filtering and questioning AI-generated content and verifying information reliability. Bailey et al. (2021), drawing on cognitive development stage theory, emphasized that AI-generated prompts play a key role in guiding novice learners into critical reasoning processes."

You can follow the cited references to the studies.
Being Chinese literature is not a valid limitation.
Rather using the studies from Harvard, Stanford and meta analysis across the very strong Chinese academic landscape - makes a much stronger case.