Are you ready for aging escape velocity? AI is bringing it.
PLUS Five things returns to add to your AI concerns, a company that is reverting to people, the dead speak in court and a new high tech product to guess. And more!
Greetings to all my Canadian readers on this stat holiday where traditionally people plant flowers and vegetables. Where we are it is too cold to enjoy anything outside. Which means I can focus on the societal perils of technological impacts. I do have a great dinner planned for this evening but before we get Cookin’ withTech I have an announcement. I have decided to go back to once per week format, publishing just on Mondays. This is to free up more time for my other writing. Let’s get right to this week’s new product but we will now have to wait a week for the answer.
EXPERIMENTAL RECIPES
MAIN COURSE: We are just a few short years from living forever
As a boy I hated the magazine Popular Mechanics. It was stuffed with DIY stories, including detailed instructions, for building things from scratch, like mini bikes and patio roofs. Issues featuring 50 great tips for your home and shop caused me to tremble. I couldn’t use hammers or saws, and I failed scissors in kindergarten. Some of my friends loved it and one guy did build his own minibike. I could explain the principles of motors and engines, but God help me if I had to pull a spark plug.
I thought Popular Mechanics would have been long gone by now, especially in a world where magazines are mostly dead. Found out this week it was still breathing. I assumed it would doing things like providing detailed plans on how to improve SpaceX’s Starship rocket, to prevent it from blowing up.
But no. Instead it had an insane article about how AI will be doubling our lifespan, in just a few short years. This catchy subject was based on an interview with Dario Amodei, the head of AI company Anthropic. Dario has trampolined into first place on the all crazy Silicon Valley leaderboard. His key message is that by 2030, with the rapid developments we are making with AI, it will have found many new medical discoveries that we live to 150. After that who knows where we can go, as we reach an escape velocity from mortality. He has made this proclamation a few times so he has had time to say to himself, “Am I over-stepping just a wee bit?”
No, he lives with others steeped in the Silly Con ethos, unmoored from anything empirical, where evidently you can just believe something in being. Soon we will get vibe longevity and vibe space travel. Why are they all scared of earth and dying? Reality check: none of you TechBros are ever leaving our planet for Mars and you are going to dance the mortal coil shuffle.
Some of Dario’s thinking is based on flawed statistics. He purportedly has a PhD in physics, which in my day would have included enough mathematics to not misunderstand averages vs variability, or believe simplistic extrapolations like we have have doubled out life span since 1900 why can’t we double it again right away? Or understand that marginal improvements in mice studies don’t translate to eternal life.
The article actually has some brilliant insights on actual medical research and biology but a headline like AI Leader says more Crackpot Things doesn’t get as many clicks for the obviously struggling Popular Mechanics as AI will double the Human Lifespan by 2030 does.
Where can I get a job like Dario’s, where I can make millions and basically say any wack-a-do thought that pops into my head? Oh right, the newly crypto-monetized American presidency.
5 CONCERNING FINDINGS FROM RECENT AI STUDIES
No reader submissions this week so returning to an oldy but goodie format. The quotes are taken from an excellent article by Jing Hu about the latest research on AI and human behaviour. I recommend her publication for those interested in deeper looks at what is happening with AI
“You and I have been conditioned—yes, conditioned—to view AI with a sense of objectivity or even superiority. "It's just math," you might think, "it runs on logic!"
“You think you're making choices, but your perceptual framework itself is being recalibrated through hundreds of micro-interactions with algorithms designed primarily for engagement, not truth”
“Participants underestimated the substantial impact of the biased algorithm on their judgement, which could leave them more susceptible to its influence”.
“AI systems, once trained, tend to be remarkably consistent, even if they’re consistently wrong or biased.”
“When I say this represents an evolutionary threat: for the first time in history, we've created systems that can bypass our natural skepticism toward other humans and implant biases directly into our decision-making processes at scale.”
A LITTLE SPICE
“The 21st-century information environment might be termed “moronogenic”, tending to reduce deep thought, sustained attention and cognitive function”
James Marriott
SPECIAL DISH: We are allowing the dead to “AI speak”. Where will it end?
Maybe you saw this story about how the sister of a man killed in a road rage murder was allowed to develop and present an AI video replica of her brother, using his voice, to address the court during the sentencing phase. The avatar, introduced a series of videos about the dead man, and then was allowed to conclude with his own impact statement. The sister and her husband had the technical chops to pull this off.
Now that this door has apparently been opened where we let someone speak for the dead, we are going to permit ourselves go anywhere with this, especially as the “authenticity” in terms of voice and video will make it ever so hard to close it.
Here is one hypothetical example to just show the endless possibilities of what we can now do. In the pardon trial for the Menendez brothers, Jose’s older sisters could put together an AI video of him saying that he wasn’t anything what his son’s have portrayed him, and that his boys were sadistic monsters. The sons themselves could put together a competing avatar using their mother saying that in fact she was deeply concerned with her husband’s behaviour and that her sons were lovely boys that couldn’t see a way out. A Fidel Castro AI recreation could admit in court that he was reprehensible and harmed Jose’s early Cuban upbringing irreparably. A dead neighbour from Beverley Hills could say that she thought that there was something off about the whole family. One of the housemaids, now deceased…………
Sorry, I was off staring into the abyss. Let’s end with a quote from this landmark trial, real words spoken by the dead man’s very much alive sister: “Our goal was to make the judge cry.”
MENU MISTAKES
No one is immune!
QUICKBYTES: One company admits to an AI mistake
I have written about my personal experiences where AI implementations (RBC Visa and Canva support) create lots of problems for customers. There are many more of these AI implementation mistakes. Here is one company that is going back.
Klarna, a buy-now-pay-later company (yes kids these really are loans), decided to abstain from marketing and customer service people in favour of digital agents. Not only that but they did it openly and arrogantly, trumpeting how much money they had saved and claiming that “its automated customer service agents could do the work of "700 full-time agents."”.
Klarna never said out loud we made a mistake, instead muttering about too much emphasis on cost and ending up with lower quality. Well let me say it for you Klarna: You made a mistake. As are other business organizations rushing ahead with crucial business applications centred on AI without thinking through better implementations. I’m not against AI - as some think - but I spent 45 years implementing large business solutions. I remember all of the lessons. Those thinking AI is magic will be always be surprised when they learn that it isn’t really so.1
That’s it for this week. There are just so many tech stories. If you see one please send it along. Thanks so much for reading. If you feel so moved leave a comment, a like, forward to a friend, or restack it on Substack. It helps me grow.
For those that are interested in two weeks time I will start releasing a series called Shining Light on the Consulting Industry & its AI magic.
There is just so much to keep up with regarding AI. Every time I scroll on my Google feed it's some new AI product that's gonna solve all your problems. I still have to look at that program you sent me through LinkedIn. I'm still going to do that. And I used napkin again in my article that I wrote today. I used a couple different graphs that I created. You might want to take a peek. You can even edit the output so that makes it nice. So far I'm pretty impressed with it.
I love how you walk us through the insanity and inanity of this all. Moronogenic is funny but apt. I appreciate your conversation with Jing Hu a lot. Much to ponder here. Thanks!