Monday Morning Tonic
Fresh Takes on a tiny bit of last week’s Technology news including AI, Zuckerberg, and Amazon
The Headliners: AI
Another corporate announcement (yawn) of a AI product: This time Salesforce
Salesforce announces its new AI product called Agentforce, their “third generation Agentic AI”. Why does everything in the AI world sound so marketing brainstorming team contrived. Yes, for those of you interested they did mention fourth and fifth generations. The name sounds more like a new branch of the CIA to me. It is built on the Atlas Reasoning Engine which sounds like it is holding up the entire world. But in this case it is merely holding up their platform. As their CEO of Salesforce AI said Clara Shih, CEO of Salesforce AI, explained: “The power of Agentforce is that it’s built from within. It is now part of every Salesforce cloud, every industry cloud, from financial services cloud to Health Cloud to government cloud and media cloud”. They didn’t list all 47 or so different versions of their clouds. Now I haven’t used Salesforce hands-on in about 15 years, but I always found these multiple different clouds slightly unnerving, especially because in those days they weren’t integrated. The secret agent will be showcased at next week’s conference (another marketing name: Dreamforce) but they have learned from much from the rollouts of the other AI players whose offerings originally astonished only to then lose much of their lustre by keeping customer expectations reasonable: “…to give you the best artificial intelligence experience you could possibly have,” he said. “The most accurate, the most capable, the easiest and the lowest cost, and no other AI will be able to deliver that, because only we are delivering this platform approach, these autonomous agents that are learning and are taking action.” Whew, no betting the house on there. In their pop icon Marc Benioff’s words, “This is going to be the most important thing we have ever done at Salesforce.” I’m taking bets that this won’t end well.
Yes but what about the coming AI apocalypse?
But at the beginning of last week a new book Nexus written by the famous Yuval Noah Harari came out. This book about the history of information follows his well worn format of starting with the discovery of sand tablets, proceeding through every conceivable human information network ending with a third section on the digital world especially AI, and the coming collapse of everything we love. His name and track record will mean that these ideas will be percolating all this autumn. So who cares about yet another routine corporate marketing announcement.
Well there are some reviewers (here are a just two examples from the Wall Street Journal and The Guardian) who think that he might be out of his depth and that some of these prognostications might be unfounded. But a much under the radar announcement might better quell our qualms. An article in the MIT Technology Review introduced us all to Lazy Robotics. This is a field that is trying to get the software controlling robots to do less. So people who were worried about AI reaching AGI (Artificial General Intelligence) with images of computers smarter that Einstein or Dawkins can relax now.
As Professor van de Molengraft the person advocating for this new approach says, “I think that the best robot is a lazy robot,” he says. “They should be lazy by default, just like we are.”
So in addition to hallucinating AI will now be lying down on the job. That means that AGI is actually going to be more like the intelligence of mid-1990s slackers.
This well intentioned move is to reduce not only the power consumption of robots and the mammoth AI data centres, but to create better focus on the multitude of incoming inputs as cars and robots navigate through the world. This is being tested on a very big problem: getting robots to play soccer. Evidently in the recent Robocup the lazy robot team won. No word on when they will field a team for the Eurocup.
Secondi
Mark Zuckerberg’s 20 year mistake
This article immediately caught my eye: Mark opening up about his biggest regret. I was on tenterhooks. Which of the numerous major societal problems was he taking responsibility for: Instagram’s diabolical targeting of teenage women about their fragile emotions, cheerfully selling everything they have gathered and formulated about all their users like a base commodity, or my favorite, his supposedly noble Free Basics internet plan to help citizens in lesser developed nations have access to the internet; well mainly tethering them for life to Facebook’s products primarily. But I was mistaken about his mistake. His regret is actually spending 20 years listening to any criticism or pushback on his exemplary products. When you are incredibly rich and isolated from ordinary life it is easy to become completely detached from reality.
The Roar
Guess what? Home deliveries do have an negative environmental effect
Wired reports that a new study by Stand.earth found that the company’s delivery emissions were up 194 per cent since 2019. That was the year that Jeff Bezos proudly announced their Climate Pledge. Now they did say that there was some tricky arithmetic involved, that they needed to make several assumptions and Amazon doesn’t provide data in easy enough breakdowns and formats to make the study easy. Of course Amazon pushed back in the strongest terms, well at least about the study:
“Stand.earth’s work is based on inaccurate data, a broad mischaracterization of our operations, and by their own admission, a methodology based on assumptions and unverified information”
The rest of their rebuttal was high level marketing-ese riffs that appear to be gravely serious but are as generic as horoscopes (sorry believers in astrology).
This is another case of the virtuous and the good (that is, us) outsourcing a major societal problem to a big institution: Amazon isn’t living up to their green commitments harming our progress against climate rescue. Really? This isn’t complicated at all. Show a chart correlating deliveries against emissions. I’m guessing that it will be quite tight. The study doesn’t even include the grossness of the excess packaging - some of it including non-recyclable plastics (note to those with their heads in the clouds almost all plastics aren’t really recyclable) - and the splitting of one order into different packages to ensure speed and promptness, as if these were medical emergencies we were purchasing for. Look maybe each and every one of us should think before placing the third order of the day because it is easy and convenient to do. Maybe get organized and drop 2 hours of scrolling for a weekly local shop. We are the consumers and we are driving the emissions.
FULL STOP