by 
I certainly don't need to tell you that AI is one of the biggest talking points of modern tech – you'll have picked up on that if you're a regular reader of this very site.
Generative AI has become a major source of contention across all walks of life, with artists, filmmakers and musicians divided over the ethical issues surrounding AI training data. For every person who feels anger that their hard work is being used to train robotic replacements, there are others who see it as the creative tool of the future: just tell it what you want, and it magically appears.
You should be aware of which side of the fence we personally stand on if you've read any of the other AI-related articles here on Time Extension, but it's alarming to see how quickly the tech is reaching into the world of retro gaming – and FPGA development is the latest battleground.
Esteemed FPGA developer @topapate has just highlighted the fact that a bunch of Gaelco arcade games are now playable on MiSTer FPGA, but there's a catch – these cores have been generated using AI and are based on existing MAME versions.
Given that FPGA is predominantly about accuracy – in fact, that's one of the key selling points over software-based emulation – it strikes me as rather odd that some developers are happy to simply entrust core production to an AI agent when such technology has been proven to make mistakes.
As you can see from the small snapshot of replies to the post on social media, this is sharply dividing opinion within the FPGA community. Some don't see the problem; after all, if a project is open source, there's no moral issue with AI being used to make ports. And besides, this means more games get FPGA cores, which is a win for players.
However, others have rightly pointed out that many 'vibe coded' ports are sometimes carried out behind closed doors, which makes it harder to assess their worth, and we've seen some AI-generated ports riddled with errors.
There's also the issue of AI use impacting the environment; while all computers have an environmental footprint to consider, AI consumes a lot of energy and requires water cooling – then there's the fact that the data centres being constructed to fuel this AI boom have massive ramifications on the local population. Welcoming the rise of AI-coded FPGA cores means embracing considerable harm on our already suffering planet – that's a fact.
Do you think AI-generated FPGA cores are the future? After all, the more games people get access to, the better, right? Or can you not see past the ethical concerns AI brings, and would rather see FPGA cores coded by humans with a painstaking focus on accuracy? What happens when AI agents become more accurate than humans for this kind of work – something that may already be the case now?
As ever, we want to know your opinion – so vote in the poll below and leave a (polite) comment.
Do you think AI-coded FPGA cores are a good or bad thing? (508 votes)
- Good
- Bad
- Undecided