For weeks I’ve been trying to remember a particular series of books. For no particular reason, they just popped up in my brain during a sleepless stretch one night, and I got to thinking, yeah, I really enjoyed those books. Ought to see if the author wrote any more of them.
The trouble is, I read five or so of them in a bunch one summer, ages ago. Like late 80s or maybe early 90s? But I don’t remember the lead character’s name OR the author (not surprising, I’m lousy at retaining names) but I also don’t remember the titles of any if the books, or anything, really, other than random bits about the character and some incidents that I think were involved … and that I’ve very much enjoyed them, but doing the whole bunch in a row had somewhat fatigued me on them and I’d turned aside to other others/subjects and more or less forgotten them.
So, how to ID them?
Well, I thought about doing one of those appeals to the SD hive-mind, but what I had to go on was so vague I didn’t want to trouble y’all. But then it occurred to me that Google now throws up an AI based result at the start of their results, and who cares about wasting an AI’s time? So I typed in a search something like the following. (Sorry, didn’t think to save it, but this should be close,)
Crime caper novels. Robberies. Rock concert. Bank, Carefully plans elaborate crimes then assembles criminals with the skills he needs. Almost always goes bad. Noir-ish. Anti-hero.
And…the very first answer suggested in the AI part of the result was "The main character name novels by author’s name. And it was right! The second I saw the answer I knew it was what I was looking for! Score one for AI!
I’ve left out the answer here, in case any of you experts want to pit your wet-ware vs. ChatGPT or whatever Google is using. Think of it as John Henry vs. the steam powered pile driver.
And it’s not just energy. Generative AI systems need enormous amounts of fresh water to cool their processors and generate electricity. In West Des Moines, Iowa, a giant data-centre cluster serves OpenAI’s most advanced model, GPT-4. A lawsuit by local residents revealed that in July 2022, the month before OpenAI finished training the model, the cluster used about 6% of the district’s water. As Google and Microsoft prepared their Bard and Bing large language models, both had major spikes in water use — increases of 20% and 34%, respectively, in one year, according to the companies’ environmental reports. One preprint1suggests that, globally, the demand for water for AI could be half that of the United Kingdom by 2027. In another2, Facebook AI researchers called the environmental effects of the industry’s pursuit of scale the “elephant in the room”.
Not to pile on after Stranger’s important reminder, but John Henry competed against a steam drill. He was driving what was essentially a very large drill bit (several feet long and an inch or more in diameter) to drill holes for powder or dynamite. The drill, or steel, was held in place by the shaker, who would twist it at every blow. Thus the lines, “Shaker, you’d better pray, ‘Cause if I miss that little bitty piece of steel, Tomorrow be your burying day.”
Sorta the right author, wrong main character. It was Westlake writing as Richard Stark, the Parker novels. Good news for me, there are lots more I haven’t read, bad news Stark stopped writing when Westlake died.
Not going to defend AI in general, but what’s the problem with them using water for cooling their computers? You heat water up, it cools back down … it remains water, yes? Unless they’re polluting it with something in the process.
Close enough—I’m calling it a win! I’d forgotten about Westlake’s Stark novels, probably because I wasn’t a big fan of them, or Dortmunder. From your description in the OP, I thought Dortmunder was too tame to be the right answer, but again, forgot about Stark.
I’m keen to hear what happens to all that AI water.
They take water from municipal or groundwater sources at rates that are stressing those resources or are in some cases completely unsustainable. The water cannot be not returned directly back into ‘service’ because it is used, and is too hot to be recycled for cooling purposes, and data center and supercomputer operators don’t want to bear the expense of building massive cooling towers and holding ponds to evaporatively cool the water for reuse, so they just dump it and draw more cheap water (subsidized in the case of municipal sources; ‘free’ in the case of groundwater) for continuous cooling. How much water is used overall is something of a guess because the use is mostly unregulated and AI/supercomputer/datacenter operators are cagey about sharing but can be estimated by energy usage rates and the number of nodes doing the ‘compute’ for training models and providing data access.
