

When I started doing this work it was scary how fast they progressed. I thought I’d be redundant in a few months. Then they hit a wall very hard. The models might even be getting worse now. That could be my perception because a big part of the job is steering them toward failure and correcting the mistakes, so now I’m in the habit of exploiting their weaknesses.
They aren’t going away, but if we can figure out the niches where they’re actually useful, maybe the big AI companies will stop pretending LLMs are a digital panacea.
My current location in rural Tasmania has some advantages. There would likely be a delay in the zombies getting here, so that gives us prep time. Low population density means fewer zombies to deal with when they do. Plenty of natural resources, food, and water.
My main issue is a lack of defensible buildings, but there are lots of wood mills so maybe something could be constructed.
The wildcard is Tassie devils, being a powerful scavenger species. Would they go for the walking corpses? If so, would they end up turning themselves, because I don’t fancy dealing with zombie nocturnal marsupials with a bite that crushes bone.