Turns out I was wrong, Rosie the Robot is real and can be yours for a low low price

Plus 5 AI Unrealities and an annoying Menu Mistake

MAIN COURSE with EXTRA HELPINGS - You can now get your own Rosie, though it isn’t as cute

I was completely wrong a number of newsletters ago about where society is with home robots. A company called Unitree - see the 2 videos below - have a slick model with a very clever name called G1. Goldman Sachs predicted in a recent report that by 2035 global humanoid robot shipments will reach 1.4 million units, with a market size of approximately US$38 billion. Humanoid robots will be first applied to factories between 2024 and 2027, and then jump to the consumer market between 2028 and 2031.

But you can beat all your neighbours to the punch by getting the G1 now. It can be really yours starting at a low low price of USD 16,000. Some cautious individuals are asking about an option that ensures that GeeOne won’t kick the living shit out of them. Looking down at the price list, I see that Unitree do not have such a fail safe option. Because of course, AI is baked in. So you’re on your own when your own little robot hallucinates, and rips you apart. Maybe after the first couple of accidents we’ll stop using this ridiculous euphemism.

I’m sure many of you were struck with awe and wonder at the dancing GeeOne. It’s true that the hardware and cognition for robots has greatly improved, but let me give you some background. That is a marketing demo. The whole point of it is to make the naïve go Wow! They programmed the little cute bot to do just that, and made it rehearse over and over in a nice controlled setting until the dancing was just right. Like your kids at their dance classes. In my business, until I learned better, I can’t tell you how much money I lost, believing in and adopting software way too early, based on some snazzy demo. The variability and unpredictability of real life is what will screw little GeeOnesy up. I’d get my GeeWonderful to fold itself up and put itself away for good. Somewhere safe.

SECONDI - 5 AI unrealities

I warned a couple of newsletters ago about the delusion that we no longer can tell reality from illusion. Although AI is only part of the technology landscape creating this problem, I’ve decided to call what AI contributes to our delusions as unrealities.

1. For reasons that I cannot fully fathom, people are wild about real crime documentaries, whether through podcasts or in video format. Naturally, an enterprising individual used AI to generate absolute bucketfuls of Unreal real crime. He said in the story from 404 media (a great paywalled site that focuses on how technology shapes the world) that the more lurid the stories were, the better. He ran a YouTube channel that had millions of viewers. This is now called AI slop which will soon to become a crucial word for your vocabulary. Except for AI stans whose objectivity has been lost and whose numbers seem to be growing rapidly.

2. I have read about this type of unreality before, but here is an interesting story about using AI to help one prepare for court. Naturally, an enterprising AI tool generated a whole bunch of made up bogus cases that made the legal position look even stronger.

3. For those listeners or readers of BBC, who have started to eschew BBC directly, in order to read popular LLM summaries of BBC news, guess what? BBC’s in-depth analysis found that the error rate of just simple summarization is pretty high. Over half of the summaries had significant errors, and one-third couldn’t even get the basic facts correct. On your scorecard, Google Gemini is the worst and Perplexity - with only a 40% error rate - is the clubhouse leader.

4. In an unreality that really shouldn’t worry us because there are only 700 million views on TikTok, are AI generated motivational videos featuring the synthesized but completely realistic voices of Donald Trump and Emperor Elon Musk. Here is one motivational sample: “When you don’t have money, people say, who are you?,” (while inspirational music plays in the background), “When you have money it’s, ‘Hey how are you? Long time no see. You look so beautiful, handsome, and amazing.’ That’s the power of money.” Guess who that one is?

5. The final unreality is the most disturbing and distasteful. Crushmate, a firm that sells AI used to generate nudity and porn of noncompliant children and women gets 90% of their traffic from good old Meta. A noble senator is trying to push The Zucker to curtail it, but good luck with that. For Canadians using Facebook and Instagram to organize boycotts of hapless American business organizations accidentally caught up in this tariff war, maybe give up Facebook and Instagram instead. That would be a real statement. After all Meta makes a gross margin of approximately 82%; for those who aren’t financially astute, that is fucking high (sorry to use financial jargon).

---

---

MENU MISTAKES

Anthropomorphism has gone too far. We now are pitying literal hunks of rock and frozen gas, and creating ridiculous human style stories. This article is NOT satire. The first paragraph includes the phrase “as the poor world hangs on for dear life.” I don’t know where to start so I stopped right there. Do we need cutesy, Disney-ish terms in order to get people to read science stories?

Mind you the term exoplanet is pretty human-centred too, in need of revision. Basically any of the likely millions of planets there might be in the universe will bear the prefix EXO. Meaning they don’t revolve around our sun, like it is the big chief of the universe, not a rather insignificant small star in a minor byway of the universe. Douglas Adams where are you!

---

That’s another wrap. As of Friday I will be on vacation (to mostly Costa Rica, thanks for asking) but I will try and keep to my publishing schedule, depending on the quality of the wifi at each stop. As always, I appreciate any feedback.