I remember when I was introduced to Google Search. I found myself – a young boy – slightly lost amongst the shelves of our local library. Turning a corner, I stopped to gaze at a cramped desk overladen with three CRT computer monitors. We didn’t have a home computer, but my father had, on the rare occasion, let me play Bow & Arrow on his work ThinkPad. A librarian must have noticed my curiosity and asked if I had used something called the web before. She explained that people could find things out on the web, and had me sit at a machine. “So how do I find things?” I asked, and she replied “You use a search engine. Most people use Google”. I don’t recall much of the next few minutes, other than it was rather marvellous and left me wanting more.

Fast forward to today, and I scarcely recognise the Google Search that I grew up with. Gone are the niche forums and esoteric blogs that were once prevalent; replaced with sites of a generic quality and pages with a veneer of information. It has begun to feel so transactional – trawling through long pages short on information that seemingly only exist for targeted advert impressions.

Using search as a way to learn and discover is now a nuisance. Social networks find themselves used as a refuge from cookie dialogues with searchers adding ‘reddit’ or ’twitter’ to queries in search of ‘real’ sentiment. Searching is no longer the wondrous activity it used to be. The SEO spammers have won – human content has self-evacuated to the islands of reddit, twitter, and facebook. The wasteland that is left has been taken over by commercial interests – businesses providing ‘content products’ and building content strategies. Commercial entities compete for human attention in search rankings and social networks and have buried non-commercial insight and wisdom.

Things we don’t know we don’t know

“A computer is […] the equivalent of a bicycle for our minds.” – Steve Jobs, 1995, talking about the future of libraries in the documentary film “Memory & Imagination: New pathways to the library of Congress”.

In 1995 Steve Jobs compared locomotive efficiency of a human on a bicycle to the mental efficiency of a human using a computer. In the context of the film, it was a remark about how computers can improve our access to knowledge and how knowledge institutions (libraries) might leverage computing in 1995.

Nearly thirty years later, that idea is relevant again. If search technology changed how we locate knowledge we are aware of, perhaps LLMs like ChatGPT will change how we discover knowledge that we are not aware of.

The limitation with search based learning is the user has to know what they’re looking for. They’re entirely reliant on the results to divulge information they weren’t previously aware of. In the context of modern search (commercial content products) the priority for a business is to efficiently distribute their effort between “relevant” content for advertisers and good search ranking. As such, a user may never discover new facets to their query when competing results contain the same ‘surface’ level of knowledge.

LLMs don’t have these constraints. At this time, ChatGPT has ingested a corpus of knowledge and over time OpenAI will ingest more. Users will shift from searchers to adventurers. ChatGPT allows users to quickly navigate topic topologies, discover things they did not know and advance their understanding as quickly as they can ask questions.