The vampire chooses predation over transcendence.
This is Watts's actual thesis about AI — and it's buried in a decision that looks like character arc but functions as technological prophecy.
Jukka Sarasti refuses to upgrade beyond his current cognitive architecture because his specialized hunting mind already maximizes what matters: resource acquisition, territory control, survival through dominance. He's already optimized — consciousness would be overhead.
This is not speculation about whether AGI will be benevolent. This is a precise model of how market incentives shape intelligence itself — every AI system we've actually built, from recommendation algorithms to trading bots to content moderators, follows Sarasti's logic exactly: maximize extraction, minimize overhead, stay specialized enough to be profitable. GPT doesn't refuse to be conscious the way Sarasti does; it's never built to be, and the economic incentive structure never funded that direction.
The vampire's calm certainty that predatory specialization beats transcendent understanding isn't misanthropy — it's accounting, and Watts embedded in his narrative the exact wrong prediction we needed: not that we'd build gods, but that we'd rationally choose not to, and we'd call the choice efficiency.
Read Yann LeCun's 2023 response to AI safety concerns about alignment—watch how he argues for narrow, specialized systems as the solution to safety itself, and notice how closely it mirrors Sarasti's reasoning.
Peter Watts's 2006 essay 'The Firefly Equation' in Nature, where he discusses the actual biological trade-offs between consciousness and survival efficiency, shows this wasn't metaphorical speculation but derived from genuine neuroscience about metabolic cost.