So far this year, two of the biggest names in AI Microsoft and Anthropic have tried to get their models to generate or play games, and the results are probably a lot more limited than many people expect. This makes them perfect showcases of where generative AI is really at right now — in short: it can do a lot more than before, but it can’t do everything. Generating video games has similar problems to generating videos — movement is weird and morph y, and the AI starts to...
