What's new
IconEra

Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

Forums  ›  Gaming
Discussion 

voke

Veteran
Content Creator
Sircaw's Army
Vibe coding is changing the way people approach software development and lately I've been pretty fascinated with the topic. With AI tools capable of generating code from natural-language prompts, people with little to no programming experience can now build applications, automate workflows, and turn ideas into working prototypes over night (holy fuck)

but.... what does this mean for the software industry in the long run?

PROS

- Developers can build prototypes, test ideas, and ship features much faster.

- People without traditional coding skills can turn their ideas into real products.

- Experienced developers can use AI to handle repetitive tasks and focus on architecture, problem-solving, and innovation.

- Lower development costs could allow more startups, independent creators, and small businesses to compete.

CONS

- AI-generated code can contain bugs, security vulnerabilities, and architectural problems that aren't immediately obvious.

- If AI handles many entry-level coding tasks, how will new developers gain the experience they need to become senior engineers?

- Are we creating developers who can generate code but don't fully understand how it works?

- Could vibe coding reduce demand for certain programming roles, change salaries, or reshape what companies expect from their engineering teams? The value might shift away from writing code line by line toward understanding problems, designing systems, evaluating AI-generated solutions, and just checking the AI system for any flaws, which still requires knowledge to some extent.

At the same time, we shouldn't assume that faster code generation automatically means better software or more successful products, gaming is an artform at its core.

Seeing people create perfect ports of old games locked onto old hardware, or straight up stealing an indie developers code has been a real trip. Hoping to read some of you guys thoughts on the matter.
 
Last edited:
I expect every dev to implement AI agents, vibe coding etc…

And they’re going to lie about it until they get caught

I’d prefer being honest. Some games are more demanding than others and so are their users
 
Vibe coding
Technically that was the start the of how people started to code using AI (just trowing prompts) and it kept evolving through different steps or styles (like spec driven development, dedicated agents and subagets etc) but as I understand you mean to code using AI.

PROS

- Developers can build prototypes, test ideas, and ship features much faster.
Yes
- People without traditional coding skills can turn their ideas into real products.
Yes and no: unless spending a lot of time, they'll be able to have cool prototypes or MVPs that some experienced dev will need to complete, optimize, tweak and particularly fix.

- Experienced developers can use AI to handle repetitive tasks and focus on architecture, problem-solving, and innovation.
Yes

- Lower development costs could allow more startups, independent creators, and small businesses to compete.
Yes

CONS

- AI-generated code can contain bugs, security vulnerabilities, and architectural problems that aren't immediately obvious.
Yes (but this already happened pre-AI).

- If AI handles many entry-level coding tasks, how will new developers gain the experience they need to become senior engineers?
By adapting, learning the basics and starting, starting to become an expert in the new thing before than other people.

Traditionally experienced/old devs prefer to keep working with what they are used to, and don't like big change and learn very different stuff every few years (even if they are used to do it).

Companies start to demand devs who are experts in the new toy, and if seasoned folks don't learn it quick, younger kids get the job.

- Are we creating developers who can generate code but don't fully understand how it works?
Yes. Good planning, management, communication and tests are key in team work and game development. Having these bases well covered and knowing the basics it won't be needed to know what the code does internally in detail as long as their part -really- does well what it needs to do.

- Could vibe coding reduce demand for certain programming roles, change salaries, or reshape what companies expect from their engineering teams? The value might shift away from writing code line by line toward understanding problems, designing systems, evaluating AI-generated solutions, and just checking the AI system for any flaws, which still requires knowledge to some extent.
Programming roles still will be needed because tech and programming knowledge will always be needed to properly optimize and fix stuff.

At the same time, we shouldn't assume that faster code generation automatically means better software or more successful products, gaming is an artform at its core.

Seeing people create perfect ports of old games locked onto old hardware, or straight up stealing an indie developers code has been a real trip. Hoping to read some of you guys thoughts on the matter.
The quality of the AI generated code in some areas is better and in other ones is worse. The thing is, a good programmer using it knows how to tell it what best practices and coding conventions it has to follow, how and where optimize, what tests are needed, how to solve different issues that AI by itself doesn't know how to handle etc.

I worked in most gamedev areas outside audio and localization and yes, I consider programming an artform. And well, for what I tested both programming with AI or by typing it yourself manually you can keep tweaking the stuff until you have it as you want. So the artform will remain there, it's just achieving the same using a different tool/workflow.
 
I use it all the time to write simple conversion algorithms to translate information from one system to another.
 
It is quite amazing.

I use it in my own game development to resolve annoying shit like tracking down bugs/performance/optimisation.
Rather spend time making the actual content of the game.

In terms of the games industry:

I think we're gonna see a big boom of solo developed that actually compete in quality with AAA games.
I think the endgame is a game where it essentially vibecodes itself.

Then it'll cause a rejuveniation of VR where you have a game that can just generate unique VR experiences on the fly.
 
Seeing so many ports being released confuses me. Also, some are in clear alpha states. But you have sloptubers like Video Game Esoterica posting 5 videos a day.
 
I expect every dev to implement AI agents, vibe coding etc…

And they’re going to lie about it until they get caught

I’d prefer being honest. Some games are more demanding than others and so are their users
Why should they be honest about it? As a consumer I only care if the product is good. I don't give a shit how they do it.
 
It is quite amazing.

I use it in my own game development to resolve annoying shit like tracking down bugs/performance/optimisation.
Rather spend time making the actual content of the game.

In terms of the games industry:

I think we're gonna see a big boom of solo developed that actually compete in quality with AAA games.
I think the endgame is a game where it essentially vibecodes itself.

Then it'll cause a rejuveniation of VR where you have a game that can just generate unique VR experiences on the fly.
Awesome
 
It is quite amazing.

I use it in my own game development to resolve annoying shit like tracking down bugs/performance/optimisation.
Rather spend time making the actual content of the game.

In terms of the games industry:

I think we're gonna see a big boom of solo developed that actually compete in quality with AAA games.
I think the endgame is a game where it essentially vibecodes itself.

Then it'll cause a rejuveniation of VR where you have a game that can just generate unique VR experiences on the fly.
How realistic is it for someone to make sequels of old IPs with this? Or is it too far fetched (as of now)
 
A lot will change. Tim Sweeney kinda hints at all of it.

Smaller teams of software engineers and programmers, replaced by people without technical coding knowledge. This will result in less employees and lower wages.

Probably a small subset will need to be very highly technical as architects of what the AI is doing and ensuring it's integrated well into the tools. Eventually the training becomes so good that the tools are extremely good at programming "known" types of games, genres, physics models, graphics programming, etc etc.

Graphics will move further into Neural Rendering which will completely replaced traditional methods of rendering, resulting in much larger graphical gains despite diminishing returns on compute improvements.

So what the training lacks will be the novel types of games that haven't been made before, possibly as it relates to dynamic NPC encounters and behaviors that will all slowly improve and evolve over time. Art directors will be needed to regulate and fine tune the outputs of the model as it pertains to asset creation.

It couldn't come at a better time given the ballooning costs in the industry. I think we will find some agile small teams that end up producing incredible work, similar to expedition 33 on steroids.

I guess the problem will be an abundance of quality and options, I suspect. The market will be flooded and thus the bar will continue to have to be raised for teams to stand out.
 
A lot will change. Tim Sweeney kinda hints at all of it.

Smaller teams of software engineers and programmers, replaced by people without technical coding knowledge. This will result in less employees and lower wages.

Probably a small subset will need to be very highly technical as architects of what the AI is doing and ensuring it's integrated well into the tools. Eventually the training becomes so good that the tools are extremely good at programming "known" types of games, genres, physics models, graphics programming, etc etc.

Graphics will move further into Neural Rendering which will completely replaced traditional methods of rendering, resulting in much larger graphical gains despite diminishing returns on compute improvements.

So what the training lacks will be the novel types of games that haven't been made before, possibly as it relates to dynamic NPC encounters and behaviors that will all slowly improve and evolve over time. Art directors will be needed to regulate and fine tune the outputs of the model as it pertains to asset creation.

It couldn't come at a better time given the ballooning costs in the industry. I think we will find some agile small teams that end up producing incredible work, similar to expedition 33 on steroids.

I guess the problem will be an abundance of quality and options, I suspect. The market will be flooded and thus the bar will continue to have to be raised for teams to stand out.
Lol were you hating on this like a month ago. Calling AI a scam.
 
Why should they be honest about it? As a consumer I only care if the product is good. I don't give a shit how they do it.

It’s easy to imagine a resetera waging crusades against devs they suspect are using it. So may as well just be transparent about its use.

No, no one has to. But I expect the demonization of many.

Everything is on the line. Adapt or these devs all die off
 
It’s easy to imagine a resetera waging crusades against devs they suspect are using it. So may as well just be transparent about its use.

No, no one has to. But I expect the demonization of many.

Everything is on the line. Adapt or these devs all die off
New better ones will arise. It's just a name. I doubt ND is the same studio that made U2. So I don't give a shit if it was to die (assuming their games become sh.t)
 
It’s easy to imagine a resetera waging crusades against devs they suspect are using it. So may as well just be transparent about its use.

People should have the right to know whether a large chunk of a product they're buying is actually made by humans or not. It's really no different than creators being forced to disclose when they are doing a paid promotion.

Music streaming platforms and YouTube are slowly becoming unusable for me because I have zero interest in listening/watching shit created by a program, and the fact that it's not disclosed while these platforms actively push that shit into your algorithm is annoying as fuck. In YouTube's case it's still actual humans presenting the videos, but their scripts are obviously written by AI. I'm also seeing this AI-presentation template on so many YouTube channels now and it's easy to spot because two completely different channels will be using almost the same style but with different colors. For me it's simple: if you didn't bother to put effort into creating your content, why should I waste my time on consuming it?

There are tells but I'd rather just check a box saying "don't recommend content created with AI".
 
Use software to help you make better software.

People will produce slop, but people and game devs using it proper, it should help speed things up for sure.
 
I'm dead serious lol.

UC3 > UC2
TLOU > UC2
Uncharted 4 > UC2
The Lost Legacy> UC2
TLOU 2 > UC2
TLOU Remake > UC2
TLOU 2 Remastered > UC2

😎
U2 was an example. My point stands. I doubt that majority of people who worked on old ND games are still there.

As for your list

Imho TLOU > U2 (better flow than U3 but I agree it's close) > U3 > U1 > Crash games

Everything else is boring trash.
 
Back
Top