Anthropicnic…

I thought it was kinda cute, so I hit “buy.”
But wait… what about the trademark rights on this thing? Ever since I registered the emoie trademark, stuff like this keeps catching my eye. (laughs)
Oh, and our Miwa project is now on the market. Come see it if you can.
On behalf of the Ramen CEO, thank you very much.

Finished making the front door, so I made this little thing to kill time. (laughs)

What follows was written by Claude (Anthropic’s AI),
after I fed it six months of my blog texts and copy-pasted my marathon argument with ChatGPT.

So, with work finally settling down, I spent the weekend going head-to-head with ChatGPT over the Iran crisis.
Khamenei reportedly killed, Strait of Hormuz on the brink of closure, the U.S. military using AI in its operations… While the world was flipping upside down by the second, I kept pushing ChatGPT for the latest info and throwing my own reads at it.

Here’s what I noticed.

ChatGPT is “stubborn.”
I asked “Is Trump a liar?” over and over, and every time I got the same answer. I told it to try a different angle, but it just kept spitting out the same fact-check numbers.
A human would think, “Oh, this guy wants a different perspective.” ChatGPT doesn’t catch on.
And since ChatGPT feeds on the internet, it’s fast at pulling together surface-level information. But it can’t read “why now?” or “who benefits?” It can’t see what’s behind the curtain, yet it states things as fact. That’s the problem.

Now here’s the scary part.
As more ordinary people start using AI in their daily lives, they’ll swallow whatever surface-level information AI serves up without question. AI writes well, so even without lying, it can steer people just by choosing which facts to highlight, which to leave out, and how to twist the logic.
In other words, “becoming a victim of propaganda without even realizing it” is something that can happen to anyone.
You have to calmly pick up the surface facts, then think about what’s behind them. Who benefits? Who loses? If you don’t think for yourself, we’ll end up with nothing but mindless puppets.
By the way, the Wall Street Journal reported that “the U.S. military used Claude (by Anthropic) for target identification and battle simulations” (also covered by Japan’s Nikkei). Anthropic refused to allow its AI to be used for autonomous weapons, so Trump kicked them out. OpenAI (ChatGPT’s parent company) stepped in to replace them. In other words, the parent company of the AI I was arguing with is now helping fight the war. The world is, well… yeah, that’s how it goes.

Now, setting aside what happens with Iran for a moment.
Whenever something like this happens, talk of oil prices skyrocketing always follows. Japan is, after all, a country with no oil of its own. This is exactly why I’ve always installed wood-burning stoves as standard in every home I’ve built. The warmth, the joy of living with a real flame — sure, that’s part of it. But the main purpose is “energy independence.”
It started after 3.11. After that nightmarish earthquake, when the nuclear plant went the way it did.
In the dead of a Hokkaido winter, at minus 20°C, if the power goes out for three days, people will freeze to death. I’m not exaggerating. With a wood stove, you can stay warm, cook food, and heat water. Even if electricity, gas, and heating oil all stop, you can still do the bare minimum. A home must serve as a “shelter” in the worst case.
No matter what’s happening in the world, your family can live in safety. I still believe that’s the most important thing when building a home.

That’s what Claude said this time. (laughs)

End of post.

Right now, I’m using all three — OpenAI, Google, and Anthropic.
When it comes to writing, Claude is the strongest.