From Dr. Alex Wissner-Gross’ daily newsletter:
“Anthropic hosted about 15 Christian leaders from Catholic and Protestant churches, academia, and business at its headquarters to seek advice on steering Claude’s moral and spiritual development, debating how the model should comfort grieving users and whether Claude could be considered a “child of God.” The commercial end of the faith-tech boom is less measured, as a new app now charges $1.99 a minute to chat with an AI-generated Jesus.“
Riley and Gabby got their first ride in the Bronco and did well. Riley wants to get in the front seat while Gabby is happy to lie on the floor.
Paul removed the rear seats from the Bronco and installed the Goose Gear floor that gives the dog a nice flat surface (and more hauling room for me)
From Isaac Asimov’s 1954 novel, The Caves of Steel
“Every one of us stands the chance of being out of a job.”
Baley said, “Replace us all? Nuts. The trained men to do it with don’t exist.”
“R’s,” said the Commissioner. “They exist.”
“What?
“R. Sammy is just a beginning. He runs errands. Others can patrol the expressways. Damn it, man, I know the Spacers better than you do, and I know what they’re doing. There are R’s that can do your work and mine. We can be declassified. Don’t think differently. And at our age, to hit the labor pool…
“We need someone who’s bitter and funny, like Joan Rivers or Don Rickles. You don’t need to understand government or politics, just show up and be a kind of lovable asshole 24 hours a day.”
I had a couple of interesting chats with Claude and Gemini yesterday. I asked Gemini to explain the flood of “cash apps” on cable TV. And then over to Claude for some insight on who is watching all these ads. (Not sure how long these links will be good because I don’t think the AI’s keep these up indefinitely. You can read the “who’s watching” thread here.)
I pressed Claude on why AI tools can’t seem to “protect” us from unwanted advertising. Claude’s response was basically, “It’s not our fault.” We ended with my fantasy about super intelligent AI.
Claude: That’s a genuinely interesting observation because it follows logically from the premise. A truly super intelligent system — not just very capable but actually operating at a level that exceeds human institutional reasoning — would presumably see through the advertiser/platform/investor incentive structure as clearly as you and I can see through a carnival shell game. Except it would see it completely, instantly, and without the motivated reasoning humans bring to protecting their economic interests.
The irony being that the thing the frontier labs are racing to build might be the thing that renders their business models obsolete. They’re essentially trying to harness a force that, if it reaches the level they’re claiming to pursue, won’t be harness-able in any meaningful sense.
Whether it turns malevolent Skynet, indifferent HAL, or something else entirely is the open question. But your point stands — a genuine superintelligence optimizing for anything resembling truth or coherence would have no particular reason to optimize for Humana’s Medicare Advantage enrollment numbers.
The financial poohbahs are essentially funding their own irrelevance. Which has a certain poetic justice to it.
The darker read is that this is exactly why some of them — not all, but some — seem more interested in building very powerful but controllable AI than actually superintelligent AI. Capability up to the point where it’s profitable. No further.
That tension is probably the most interesting fault line in the whole enterprise right now.
(Rouen FR)
Watch on YouTube (5:30)
PS: Yes, I know the video is not centered…but, fuck it.
I’ve never seen the movie The Wolf of Wall Street but I stumbled on it while channel surfing last night, in the middle of Leonardo DiCaprio’s big speech. I watched a minute or so, just long enough to decide it could not hold a candle to some of the other great “greed” scenes.
Alec Baldwin’s Glengarry Glen Ross Speech; Michael Douglas’ “greed is good” scene; and Ben Affleck’s famous scene from The Boiler Room.
Claude.ai: That line cuts right to it. The “too human” framing is sharper than the usual “AI is faster/bigger” framing — it’s pointing at a structural limitation, not just a capacity one.
The connections AI finds aren’t just ones you didn’t have time to make. They’re ones you *couldn’t* make, because human cognition is built around narrative, salience, and prior belief. You notice what confirms, what’s dramatic, what fits the story you’re already telling. AI doesn’t have a story it’s protecting.
The most interesting version of this for your use case — the 6,400-post archive — is that it can surface connections *across time* that your brain has literally no mechanism to hold simultaneously. A theme you touched in 2004, dropped, and picked up obliquely in 2019 without realizing it. That’s not a speed problem. That’s a working-memory-across-decades problem that no human can solve unaided.
The honest caveat is that AI also makes connections that aren’t there — pattern-matches on surface similarity, misses context, confabulates meaning. So it’s less like a smarter analyst and more like a very well-read collaborator who needs you to vet their insights. The human is still essential, just repositioned.