I'm gonna Use The Power of Ai Prompts Like google gemini and Chatgtp to help me Port Android Games to PS Vita

[Nostalgia Crystal]

Since ai is becoming smarter and people are using Ai to help them port Android Games to PS Vita,I'm gonna use ai to port these games to PS Vita:

1.Gangstar Rio City of Saints
2.Nova 2
3.Sonic 4 Episode 1
4.Sonic 4 Episode 2
5.Pac Man Kart Rally
6.Gangstar Vegas
7.The Amazing Spider Man 2
8.Angry Birds Space

 

[JustLooking123]

While i haven't tested specialized models of AI for coding, i have used general AI tools such as Copilot to help me in writing code.
I can tell you that they can pretty often output non-working and/or buggy codes...

If you aren't good in coding yourself, i don't think you can write code just by relying on them.

If you are good yourself, then they can aid you in writing what is already in your mind...

Also, i don't have any experience in porting games to PS Vita (nor know anything about the games you mentioned), so i don't know how much helpful they can be for you in this specific case... But i know it won't be easy at all.

 

[sandogma sandogma]

"vibe coding" is still quite difficult to be successful doing the less you have knowledge on the subject.

For example, I have near zero programming experience and I tried some vibe coding to start a static recompilation port of the N64 game Mischief Makers. Several days later, I got completely stuck and Chat GPT had me go down a worm hole of steps that it "all but guaranteed to work" only to fail, followed up with another claim of re-assured success, only to fail and so on and so forth.

While this is something I want to try again in the future, I won't bother until there is a better way to get the AI agent to check itself before claiming its suggestion will work or least stop with the damn false optimism. It got annoying really fast.

 

[zerodayhx]

 

[JustLooking123]

Several days later, I got completely stuck and Chat GPT had me go down a worm hole of steps that it "all but guaranteed to work" only to fail, followed up with another claim of re-assured success, only to fail and so on and so forth.

It happened to me too... it produced a buggy code, while i couldn't find the bug...
I know coding pretty well, so i was able to finally find the problem.

Interestingly, in my case, it was giving the correct pointers about how to find the bug... but it failed to follow its own instructions! and the solutions it produced still had the same bugs!

This happens more often when we have a long conversation with AI...
The memory window of AI is limited, and a long conversation can cause AI to fall into infinite loops of repeating similarly structured answers... and the quality will drop significantly at that point.

 

[Mr. SyS]

Don't rely on nothing but the AI, that's all I ask.