On the weekend, Microsoft issued a Demonstrate technology From its AI Copilot Research Labs, displaying a productive AI producing 2 earthquakes from the beginning. Or something resembles it, at least, because the actual game has never felt as much nausea as it does. Nevertheless, when we will complain to the lead programmer, John Carmak, behind the ID software seminal game.
Fairly, I don’t think he is referring to the demo graphics or performance, as he specifically said on x“It’s an impressive research work!” There is a lot of criticism in response to anyone.
The research science journal Nature and someone like me, who joined the 3D graphics programming on the PC due to the choice of earthquake 2, reads a mysterious substance like some ancient alien script. I am certainly nowhere near the experienced in AI programming to decide the relative qualities of a large group of professionals.
Argument, Carmic Is The qualification is because if he says it is impressive, I certainly disagree. Keep you in mind, its interested in AI, starts an AGI (Artificial General Intelligence) company Deep technologiesBack in 2022.
But although the research is definitely advanced, I can’t help but feel that the final results are not too high. Yes, this is a very early tech demo and in recent years, we have all seen the realistic and accurate audio and video that has been surprised by the surprising nonsense of Generative A.
But this is not a slow -like frame rate or past rendering with which I have a problem. I am not worried about the fact that the demo is struggling to maintain a comprehensive grip in the earthquake, the game is about 30 years old. To me, the problem is the same as a very short but taken to create a ‘gameable’ demo. From a self -research dissertation:
“We extracted two datases, 7 maps and Sky Garden from the data provided by us through Ninja Theory. 7 maps contained 60,986 matches, which received about 500,000 individual players, with 27.89 tabs on the disk more than 7 years.”
The original earthquake was created by a handful of people – some designers, programmers and artists. To do this, they did not need 28 TB gaming data, just their own ease, creativity and knowledge. They did not need very expensive GPU servers, which requires many kilowatts of electricity, so that they can offer the graphics prepared.
I just have no problem in research for the sake of research (unless it is legal, morally and morally stable) and at the end of the day, it is Microsoft that has spent its money on the project, not on taxpayers. But if I had a shareholder, I wonder if this money was well spent, especially compared to how much a semi-decent game development team has to spend for the time period, which will start to create a similar earthquake 2-on-LLSD game.
I have no doubt that at some point in the near future, AI can produce much more impressive and playable for game, but it will definitely not be cheap – in terms of computing And Electricity power is needed – a group of talented people sits in front of some polite PC. If it ever comes I will be a lot of, Lot Influenced but also deeply concerned for the future of sports development.
I wonder what Carmak would say to him?
