Tesla settles, Amazon shows what's behind the curtain, and Globant shows human intelligence and not the artificial kind.
I no longer have to say "what a week" in emerging tech innovation; the news cycle is consistently a blitzkrieg regarding emerging tech, which is matched only by geopolitical movements. I think that is not a coincidental parallel.
However, here are some tech innovation nuggets, Gentle Readers, for your consumption. We know you are curious and care, and it helps to track what's shaping up in the world of emerging tech.
Tesla has settled its first wrongful death case.
Musk famously declares that he will never settle cases while quietly settling them behind closed doors. University of South Carolina law professor Bryant Walker Smith observes that it's a significant signal that they are willing to settle these cases. Human factors engineers will tell you this is a problem well-known in automation since it emerged. Semi-vigilance is perilous and creates more distraction and strain on the human brain. If it doesn't know it can shut off completely, the brain is under stress, and its performance deteriorates over time. It's quantal, like that. This research is available to anyone and goes back all the way to the 50s. The most clearly documented ideas on the perils of automation were written by Lisanne Bainbridge in the 1980s when Musk was writing video games for computer contests, and Altman was a pre-spermatozoa. Either people haven't read it, or they have, and they think they can get away with it. Neither scenario speaks volumes about their ability to lead responsibly.
Amazon admits its AI was 1,000 Indian contractors.
This wasn't from this past week but from a few weeks ago. However, tracking how people spin the truth on emerging innovation is getting more critical so intelligent people don't get taken in even more than they already have. Amazon's grocery AI turned out to be human actors working behind the curtain of "AI" like a Kansas humbug. If Tesla's settlement is an example of the perils of automation that is not entirely automated, here is another "semi" prefix you see everywhere. "Semi-supervised" learning. If you claim you know how that works, you are lying because the only people who do are under NDA. That's right. When does "fine-tuning," "customization," and "maintenance" actually involve rules-based programming instead of neural nets, and when does it actually involve thousands of contractors working in an offshore location? Contractor is sometimes a generous word for what these humans are. There is another word for it.
In better news, Globant is a company that is displaying intelligence and giving its employees permanent remote status.
All the research shows that RTO mandates do not improve productivity, profitability, or "collaboration." All the research indicates that RTO mandates have a dramatic impact. The thing is, it's all negative. What is genuinely driving these rigid mandates? I have my theories. Do you? Again. Either they aren't reading the research or betting you haven't. I can't decide which is worse.
It's a little bit of a short one this week. I took a break and unplugged, except for Duolingo and some puzzles and games I like to play. I was on a boat. There was sun and sand and salt, and good wine and Jamaican rum. It was great. Get out of the office and unplug if you don't have plans to. I'm already planning my next one. One week a quarter is all I need, as well as an occasional long weekend.
Stay Curious!
JP
