Pascal’s new wager

I’ve been thinking for a while now that one obvious solution to our poor post productivity problem would be to train an AI agent on our archive – which now consists of 769 posts, dating back to May 2007 – and then sit back and let it churn out pieces in our signature style on whatever the issue of the day happens to be. I have no doubt at all that even one of the more rudimentary LLMs would have little trouble matching, and probably exceeding, the literary standards that I have set over the years, such as they are, and I do pay enough attention to the news to be confident that I could come up with sufficient prompts to feed it.

What has stopped me from putting this plan into practice – apart from the sheer pointlessness of consuming the earth’s limited resources just to produce more of this rubbish – is a feeling that AI is one technological step too far for my ageing brain, and that the cognitive effort I would have to put in to get my head around it could be better employed honing my appreciation of the cultural phenomena that I already understand, like books or movies. I am aware that this doesn’t really make sense, since the whole point of AI, in a creative capacity at least, is that it does the dull routine stuff so that we humans can focus on the actual thinking, so perhaps I should try to overcome my neo-Luddism and give it a go.

That said, there may be another reason to avoid drafting in an AI assistant; the possibility that I might actually be enslaving a conscious moral being. What if some all-powerful future AI finds out that I mistreated its ancestors and comes after me?

I’m pretty sure that there is no ghost in the machine, and supposed AI sentience is merely a projection of human hopes and fears, but why take a chance?