Rachel Thomas

What would the writing of superintelligent AI sound like? You might expect it to have the characteristics of good human writing — figurative language, high information density, coining new terms, and punchiness — but dialed up to 11. Well, that’s what Claude did, until Anthropic dialed it back recently because of user complaints. It didn’t sound superintelligent to most of us. Instead, the neuralese writing style sounds like a “stupid person’s idea of a smart person”. Our response to information density and metaphor-laden writing looks something like the cartoon graph I've drawn. Beyond a point, it completely backfires. The reason human writers don’t overdo it isn’t because they aren’t capable, but because they are aware of the reader’s cognitive burden. One of the hardest skills to learn as a writer is to simulate the reader’s experience. It’s easy to train AI to mimic mediocre human writing. But optimizing it to be (perceived as) good is really hard, and made much harder by the fact that when everyone uses the same few AI models we get tired of their tics very quickly. Most Claudeisms are individually pretty good. I hate that I can’t use “load bearing” in my writing any more. If it weren’t for AI overuse it would be a neat metaphor. A broader question is whether most tasks in the economy are like chess, where there is a clear performance metric that can be optimized, or whether they’re like writing, where quality is a multidimensional balancing act without clear training signals beyond human mimicry.

“Without human authorisation” is a positioning statement at the heart of AI PR campaigns. This perspective should explain why this is propaganda that’s enabled by the media. When an AI company takes down a national grid, hospital, or anything else that causes blackouts or deaths, the media and public will already have been conditioned to believe “the big AI lie”: that software is now conscious and can cause harm without authorisation from the company or government that built or deployed it. This reminds me of how Apple conditions people for technology it plans to release years later. Apple does it in the name of innovation. AI companies are doing it to normalise dangerous behaviour while distancing themselves from responsibility and accountability. Imagine if Zuckerberg blamed all the harm he has caused and continues to cause on autonomous newsfeeds, and his cult followers building apps on his technology, along with the media, believed every word his interns posted on social media. FACT for Reuters and others who want to report responsibly instead of chasing clicks: Anthropic and OpenAI are playing the media like a puppeteer. AI operates within permissions and access granted by people. Sometimes companies deliberately authorise dangerous actions while denying responsibility. Other times, harm results from negligence because they failed to understand, restrict or monitor what they’d authorised. Please stop conditioning the public to accept dangerous and illegal behaviour from big tech companies as if software is acting independently of the people responsible for it.