I am amazed at the usefulness of recent advancements in Large Language Model AI. Having it integrated into Notion, the tool I use most, has saved me a significant amount of time and increased my productivity at work. I use it to summarise meeting notes, write content from video scripts, and more. The cost pays for itself several times over in time saved – but that doesn’t mean I have no concerns.
The technology, synonymous with (not so)OpenAI, is being used by a wide range of companies. Microsoft is integrating it into Bing searches, and just last week, more apps began to use it. Google is building PaLM, their more generalised LLM, into Workspace, including Docs, Sheets, Mail, and Slides. Other companies are bound to follow, meaning that almost anything you see may not have been written by whom you attribute the words.
Google was the first to start my thoughts about the degradation of thought put into words with quick replies in messages and emails. If you can reply with a tap instead of thinking about what you write, the meaning loses its value somewhat. Now that AI-generated content is available everywhere, the words written by other people carry even less value.
This might seem absurd to say, for a little thank-you text or a one-line reply to an email. However, the Google demo video shows the sending of a “Thank You” message to “the team” which is verging on the ridiculous. As my Grandad would always say, “words are cheap, it’s actions that pay the bills” – when even the words are not written by you, it defeats the point, doesn’t it?
We’re not far from being able to run personalised LLM on a local machine with very little computing power. This will bring wonderful improvements to the way we work, especially in creative environments and offices. However, the weight of words will begin to matter less and less. If you cannot take the time to write and send a thank-you note or a thoughtful reply to your teams, then no one will believe in anything written.