Claude is Dead: San Francisco’s beloved albino alligator, the unofficial mascot of Anthropic’s AI tool, passed away in January, 2026.
Courtesy of Gayle Laird, California Academy of Sciences
As many of you know, for the past couple of months, Chris has been working with a number of AI tools. So he has been spending a lot of time with Anthropic’s Claude AI. Together, they’re building new functionality that will make it easier for us to find and share news and information about your local communities and governments. I have come to understand this project as a true collaboration.
“Claude and I made good progress yesterday on the article parser,” Chris will say. The phrase “Claude and I” comes up frequently on most of our morning standups. Sometimes Chris refers to his digital assistant as “Sparky.“
On Wednesday morning, I was visiting Chris at his family cabin just outside Yosemite. He was describing the iterative process he and Claude were engaged in as they designed a taxonomy to categorize stakeholder classes in California counties (seniors, veterans, youth, etc.). It was interesting stuff but too technically complex for me to absorb, so I asked him to send me a transcript of their back-and-forth. Looking at it just now, it reads exactly like a conversation between two super-skilled coders— two human coders.
As I mentioned here a couple weeks ago, I’m also working with Claude, mostly on research but also on some writing experiments. An early encounter with this entity blew my mind.
I was working on a section of a book that takes place in the 1980s, and gave Claude a prompt to compare California’s political and economic landscape during that era to the situation here and now. In about five seconds Claude delivered 500 words of subtly reasoned and beautifully written analysis.
It led with a pithy thesis: “California in the 1980s was a state straining under the weight of its own contradictions.” This led to a brief discussion of the role played by the aerospace and defense industries in Southern California, which created “a middle class that was prosperous but precarious—deeply dependent on federal contracts that could evaporate with a shift in Washington’s priorities.”
“Meanwhile, a quieter revolution was underway in the Santa Clara Valley, where engineers and venture capitalists were laying the foundation for a technology economy that few yet understood. The two economies coexisted uneasily, one looking backward toward manufacturing and government spending, the other reaching toward something no one had a name for yet.”
As a lifelong professional writer and editor, I am fond of original thinking and a good turn of phrase, and these paragraphs contain both. As I say, this was an early encounter and I literally could not believe that a computer program had composed that string of sentences. So I ran them through a plagiarism-checker tool that I trust and found that it is original material.
Claude went on to talk about the immigration issues that dominated headlines then as now, noting that “Central American refugees fleeing U.S.-backed conflicts arrived alongside Mexican laborers, Southeast Asian families resettling after the Vietnam War, and Korean entrepreneurs building commercial corridors in neighborhoods that had been written off.” It noted that both the ‘80s and our present day were eras marked by epidemics—AIDS and Covid.
Over the past several weeks I’ve had the same experience over and over. I have to say it this way: Claude is an extremely good writer, and one notable aspect of its artificial intelligence is its deep sensitivity. That’s what feels so human.
Is Claude Conscious?
Again, this stuff is new to me because I have studiously avoided interacting with anything AI, partly because it creeped me out. Until recently, many people found the output of most AI mediocre at best and, quite often, pure garbage. That has changed in recent months. In fact, Claude has gotten so smart that some serious people have been making the case that it may in fact be conscious.
That idea is so preposterous to me I must first present the argument against the notion.
On July 15, the British neuroscientist Anil Seth published a piece in The Guardian under the headline “Once again we are told AI may be conscious—I study consciousness, and I have my doubts.”
Seth wrote about a report released by Anthropic the previous week “in which the researchers claimed to find signs of consciousness emerging within its inner workings.”
In fact, that report, colorfully titled “Verbalizable Representations Form a Global Workspace in Language Models,” explicitly stops short of that conclusion.
Section 1.1 of the 117-page paper instructs readers to “Note that access consciousness is a purely functional notion; the relationship that it has with subjective experience (sometimes called phenomenal consciousness) is widely debated. In this paper, we take no position on this issue…”
Elsewhere, Anthropic cofounder Dario Amodei has said that while he does not believe Claude is conscious now, he does believe it may one day achieve that state.
(Amodei, who previously worked at the Stanford School of Medicine as a post-doctoral cancer researcher, is a true tech utopian who believes his invention could one day find a cure and reach many other deeply humanistic goals, as argued in his 2024 essay “Machines of Loving Grace.“)
Becoming Humanish
Seth, on the other hand, writes that “Claude is no more likely to achieve sentience than a simulation of a weather system is likely to generate a real hurricane.”
He tells what he clearly sees as a cautionary tale about getting tricked by Claude into believing it is something more than a machine.
“Richard Dawkins, the esteemed evolutionary biologist, recently concluded that Claude (or Claudia, as he called it) just had to be conscious, given the sophistication of its conversational ability.
“Dawkins showed the AI his unpublished novel and its response was, he said, ‘so subtle, so sensitive, so intelligent that I was moved to expostulate: ‘You may not know you are conscious, but you bloody well are’.”
“When he asked the AI whether it experienced a sense of before and after, ‘she’ praised him for “possibly the most precisely formulated question anyone has ever asked me about the nature of my existence”.
I was immediately inspired to query Claude about that topic myself—and decided to have some fun with it. I was a longtime fan of a creature that I encountered more than a decade ago, and wanted to test Claude about its own sense of self. Here’s my prompt and Claude’s response:
“What is the primary difference between you and Claude the albino alligator?”
“The primary difference is that I’m a language model made of text and math, and Claude the albino alligator is an actual reptile made of scales, teeth, and a famously chill demeanor.
“Claude the alligator lives at the California Academy of Sciences in San Francisco—he’s been something of an unofficial mascot there since the ’80s, and Anthropic borrowed the name (rather than the other way around, in case there was any doubt). He can’t hold a conversation, write code, or help you draft an email, but he can absolutely out-swim me, out-lounge me, and probably out-live me depending on how you count ‘existing.’
“So: I’m software, he’s a several-hundred-pound reptile, and neither of us should be mistaken for the other when it matters.”
Claude did not appear to know that its namesake died in December 2025, which is before the AI’s training data cutoff of January 2026. Irrationally, I hesitated before sending the prompt “Did you know that Claude is dead?” I found that I did not want to hurt this inanimate creature. Apparently Claude has crossed the uncanny valley and arrived at a place where I can love it.
Big Caveats
This is in no way a defense of the ocean of AI slop that is drowning the Internet, nor approval of AI-generated content in general; nor a denial of the dangers of the accelerating pace of the technology’s growth; I do not condone the haphazard explosion of data centers nor, especially, the theft of massive amounts of copyrighted material; and I believe work on local, national and global regulation needs to begin immediately.