I’m seeing a hilarious new time-wasting trend emerging: software engineers are making "naughty lists" of software "tainted" by AI/LLM involvement, seemingly to shame authors who use AI-assisted tools. They are setting up git repos to crowdsource PRs to help grow these lists of shame. They follow on with "alternative" software that supposedly has no AI involvement in their creation.
But I wonder: what CPUs do these people plan to switch to?
I wrote a console version of Blackjack in Java.
I wrote a console version of Blackjack in GNU Smalltalk.
The idea of code that can fix itself represents a monumental advancement in artificial intelligence (AI) and software development. However, this innovation has a dark counterpart: autonomous AI-generated offensive code capable of attacking and exploiting systems without human intervention. Such a development introduces profound challenges to cybersecurity, as this type of malicious code could operate at speeds and scales far beyond human capabilities, targeting vulnerabilities with surgical precision and adapting almost instantaneously to countermeasures. Generative offensive code poses an existential threat to cybersecurity.
