Thoughts on AI and AI-Veganism [2023] (www.marginalia.nu)

A recent commentary highlights the growing threat posed by large language models (LLMs) to traditional search engines, attributing this shift to the increasingly frustrating user experience of navigating modern websites filled with ads, pop-ups, and data requests. The author argues that LLMs provide a more streamlined and less irritating alternative, offering users a reprieve from the "dark patterns" prevalent on the web. This perspective raises important questions about the potential for AI to reshape online interactions and diminish the financial viability of ad-driven models, especially as independent websites may remain less affected. While acknowledging the concerns surrounding the sourcing of training data and potential exploitative practices in AI development, the commentary emphasizes the importance of understanding the broader implications of these technologies. The critique extends to the utility of AI in software development, suggesting that it may complicate rather than simplify the creation of large software systems. Although AI tools can alleviate mundane coding tasks, there is a risk that the speed introduced by these tools will lead to poorly designed software. Ultimately, the author warns of the implications of dependence on LLMs, noting that a limited number of providers may wield significant control over user access and costs in the future.