It was about twenty years ago that the world (of web developers) was getting excited about a concept called Web 2.0. That was my world at the time and, as I recall, the two key pillars were user-generated content and behind the scenes mark-up that not only presented things well on the surface but provided extra information about what the page was about. For an example of the latter, you might flag something as a review and include machine-readable (ie. easy to parse) information about what you were reviewing and your overall X out of Y score. In theory, that would allow someone else to search across numerous websites for reviews of a given thing and aggregate the results.

I’m not sure it has come to much. A lot of the web now consists of user-generated content, be it YouTube videos, Facebook posts or contributions to X (formerly known as Twitter). There are lots of gems out there but also a massive amount of twaddle, made worse by the profusion of ads. Some of those are targeted but most seem to work on a scattergun approach – they might get in front of people who will be interested but, by and large, they just leave the experience in tatters.

Meanwhile, that idea of semantic (ie. meaningful) markup also seems to have turned belly up. Instead, the modern trend seems to be to rely on a range of systems under the ‘AI’ heading to scan through plain old text and do clever things to figure out what people meant. Is that a good thing? It is certainly less work for content creators but I’m not sure the machines always pick up the right end of the stick. Worse still, they don’t seem to have any filter to indicate when they switch from carefully gleaned knowledge to drawing on weak sources or just making things up. Woe to the person who tries to get AI to write an essay which will be graded by someone who knows what they are talking about, and woe to all of us when that latter function also gets outsourced to silicon facsimiles.

It might be a bit premature to try and assess the history of the ‘web’ given that it is only about 35 years old (and in relatively widespread access for less than that) but I’m not convinced that we are demonstrating a proper process of learning and not just bolting new ideas on faster than we can cope with.