It was right at the beginning. Before most non-tech people had even heard the name ChatGPT, HN was already flooding the homepage with LLM posts and it became clear to me they were going to be big.

The consequences were even clearer, and I predicted the consolidation of power in the hands of a few, their use for surveillance, propaganda, discrimination, the proliferation of AI psychosis, sneaky ad insertion, carelessness and loss of skills, erosion of online discourse, and more. I didn’t predict the teenage suicides so soon or the rising costs in consumer hardware. I also underestimated the rate of increase in energy use (and thus the blow to environmental efforts) and that regular people would be left without electricity to power data centres.

As soon as I realised all the potential (now factual) harms and that the good parts are lacklustre in comparison but that people would eat it up at a massive scale anyway, I thought “uh oh” and “oh shit”.