Your Weekend in Doom: Retail Regret, Metaverse Meltdown and Deep Fakes
The uncomfortable truths about AI, plus: Meta's creepy glasses & the death of "Go” stores.
Dear Friend—
Jeffrey and I like to talk about “the future as a paradox”—when you ask a room full of people, some will tell you they believe the future is bright, amazing, and here; others will respond with caution, seeing the future as uncertain and sometimes outright scary. I often look at the radical Weekend Briefing and see the very same—signals showing an amazing future ahead of us, while others indicate at least a bumpy road (and sometimes a complete train wreck) toward this future.
Let’s see what this week holds for us—shall we?
Have a beautiful day!
P ツ
The Thin Wisps of Tomorrow
Poof Goes the Store — Remember AMAZON’s Go Stores? The future of grocery retail where you walk into a store, grab what you want from the shelf, and head out without ever pulling out your credit card—with the store “seeing” what you bought, identifying you, and charging your credit card? Well, it looks like that utopian/dystopian version of the future might just not be ready yet. AMAZON is largely pulling out of the technology and closing stores, mostly due to the fact that the tech is not doing the one thing it was supposed to do: save time. A good reminder that breakthroughs take time (and sometimes are simply dead ends—which I, personally, don’t think this tech ultimately is…) (link ↗)
Double-Poof Goes the Metaverse — If you know me, you know that I was always extremely bearish on the whole Metaverse trend. I just could not see how the world would migrate, en masse, into an environment that was originally conceived as a cynical, dystopian version of our future selves, requiring you to wear a sweaty, nausea-inducing headset, and that doesn’t even render your limbs properly. But for a while, the world clearly thought differently—and even invented the ultimate, apex-predator-style C-suite title of “Chief Metaverse Officer” (seriously, LOL!). That seems to have come to an end now… Which makes you wonder how many of those execs will simply leave you guessing what the CMO in their LinkedIn resume stands for. (link ↗)
Deep Fakes Are an Actual Problem — When it comes to AI-generated deep fakes, a lot of the conversation centers on the misuse of AI image, sound, and increasingly, video generators in elections, financial scams, and pornography. Hurricane Helene, which hit the US just a few days ago, highlighted a whole new category of deep fakes: images of people affected by disaster—with troubling real-world consequences for those who are impacted in real life by disaster. Yet another indicator that we need to improve our defense systems and become better at sifting through what’s real and what isn’t. (link ↗)
Surveillance Powered by Ray-Ban — Talking about problems: Students at Harvard University hacked together a demo that takes Meta’s spiffy Ray-Ban AR glasses (which have a built-in camera), grabs the image feed from those glasses, runs the video through some AI and search algorithms to provide the wearer with background information on the person they are looking at. Imagine a person looking at you and instantly knowing who you are, where you live, where you work, what your political leanings are (based on your social media posts), etc. Not creepy at all. (link ↗)
Where Are the Productivity Gains? — The AI mantra “it will make you more productive” has been repeated ad nauseam, especially when it comes to one of the earliest (and thus longest established) areas of AI assistants: coding. AI was supposed to turn the average programmer into the fabled “10x rockstar” coder—alas, as (actually good developer) Gary Marcus points out: “Data, though, is where hype goes to die.” And the data isn’t pretty: at best, AI coding assistants lead to modest gains in productivity, decrease code quality, and introduce bugs. Now, AI will get better—alas, we might be well served to consider those fantastical AI claims with a grain of salt. (link ↗)
The Smarter the AI, the More It Will Lie to You — Here is a fascinating insight into AI and its tendency to hallucinate: The more sophisticated the model, the more likely it is to lie to you. The reason: We told them to. It turns out that the way we train(ed) LLMs, which is to give us answers and not respond with the equivalent of a shoulder shrug and a quietly uttered “I don’t know,” has trained our models to provide an answer, any answer, regardless of how wrong it is. (link ↗)
Your Boss Truly Cares About You — Nothing says “I care about you” more than an AI-generated avatar of yourself telling your employee how much you value them. Or so Zoom thinks… Instead of cleaning up the UX disaster that its namesake software has become (seriously, have you looked at the settings menu in Zoom lately?), the company seems to prefer to add one more level of creepy, uncanny valley AI-BS (or is it “aibs”?) to its video chat software. You know that you can achieve the same thing by simply talking into your webcam or phone and sending your people an actual message from an actual human being, right? (link ↗)
Climate Change: It’s Complicated — Thought liquefied natural gas was cleaner than coal? Once you take into account the energy needed to supercool and transport it, its carbon footprint is worse than coal. (link ↗) And methane emissions from dairy farms are five times higher than previously thought. (link ↗) The ray of light? You can turn that methane into £400 million a year in fuel cost savings.
What We Are Reading
🧠 How Your Brain Tells Speech And Music Apart How exactly does the human brain so effortlessly and instantaneously distinguish between speech and music? @Jane
🛰️ How Nasa Astronauts Vote From Space Aboard International Space Station Out of this world! Astronauts, believe it or not, can vote from outer space by filling out an electronic ballot aboard the orbiting laboratory; the document flows through NASA’s Tracking and Data Relay Satellite System to a ground antenna. @Mafe
🏢 Revenge Of The Office Executives want employees back in the office en masse, but their arguments for ending hybrid work arrangements aren’t entirely persuasive. The data remain mixed, and for workers with leverage, the office is likely to remain contested territory. @Jeffrey
📲 Please Don’t Make Me Download Another App Although slightly dramatized, this appropriately raises the point that over-assigning apps to highly specific aspects of jobs to be done increases our flood of apps. Careful curation of what app is truly needed on the user’s side shouldn’t be forgotten in navigating it. @Julian
♻️ Thred Up Resale Report America’s used clothing market hit $43 billion last year, up from $23 billion in 2018. Even large retail brands like Banana Republic have jumped on the trend with in-house resale businesses. @Pedro
👗 Why I’ve Tracked Every Single Piece Of Clothing I’ve Worn For Three Years What happens when you track every single piece of clothing you have and how often you wear it? You gain some unique insights into not only what you like but also what makes a good or bad garment purchase decision. @Pascal
(Random) Bits & Pixels
🧠 The fruit fly’s brain is now fully mapped. Fantabulous. (link ↗)
💻 A working Turing machine made out of LEGOs? Heck, yes. (link ↗)