Note: Most Thursdays, I take readers on a deep dive into a topic I hope you'll find interesting, important, or at least amusing. These essays are made possible by — and are exclusive to — our VIP supporters. If you'd like to join us, take advantage of our 60% off promotion.
It all started a quarter-century ago with Jennifer Lopez and the infamous Versace gown she barely had on at the 42nd Grammy Awards in 2000. The sheer green jungle-print dress featured a V-neck so deep it plunged almost to the floor, which is also where many jaws were located.
Almost immediately, "Jennifer Lopez green dress" became the most-searched-for term in Google's brief history. Google had gone live only about 18 months earlier, but Lopez and Versace probably drove more people to the fledgling search engine than anything that had come before.
I couldn't find any source saying how long "Jennifer Lopez green dress" reigned supreme among search terms, but it did spur Google to launch Google Image Search in 2001. It seems almost impossible today, but the Year of Our Lord 2000 version of Google produced only text results — links to sites that might show Lopez nearly coming out of the dress you could pretty much see through.
"People wanted more than just text," former CEO Eric Schmidt later said of the company's moonshot effort to design, build, and launch the first image search engine as quickly as humanly possible. "Google Image Search was born" on July 12, 2001, Schmidt later recalled.
Google Image Search immediately became one of the company's most-loved and most-used tools, and constant refinements over the last 25 years helped keep it that way. I moved on from Google years ago — mostly because I despise the company's business practices — but when the copycats fail, I find myself back on GIS.
Then this week — on Image Search's 25th birthday — Google just had to go and ruin it.
Say it with me now: "By adding AI nobody asked for."
But this isn't really about Google, Image Search, or even AI. It's about tools. To see what I mean, let’s take a quick look at how Google broke one of its most popular features.
For now, Image Search retains its delightfully minimalist interface, but Ars Technica reported this week that Google says that soon it "will feature a gallery of images from across the web before you’ve even searched for anything."
Wait, wut?
"Google says this gallery will be updated continuously based on your interests," according to Ars, and your interests in the revised interface mean "your web and search history on Google. So the things you look up and interact with online will inform what content Google suggests in this new interface." Google wants to do this even — and it's difficult for me to say this without getting angry on behalf of people who still use Google — when they aren't searching for images.
If you searched for "Jennifer Lopez green dress" when you started reading this essay (just to jog your memory, I'm sure), you might want to scrub your search history before letting any of the young'uns borrow your laptop. You just don't know what photos Google might decide to serve up before young Bobby or Betty can even type "backpacks for school."
Based on what I could glean from Ars Technica, the "improvements" are a subtle way — a tool, really — to get people using Google's Gemini AI. "If the sheer volume of existing images on the Internet isn’t doing it for you," Ars added, "Google is making it easier to generate new images with AI."
"If you want more AI images in your search results, just ask for one in your query. Google’s AI will generate and place it in the AI Overview that occupies an increasingly expansive portion of the results page."
You don't have to be paranoid to suspect that the next step might be auto-generating AI images based on every little detail Google knows about you, whether you asked for images or not — and to keep you scrolling. "Anything to increase engagement" is the unofficial motto of every Big Tech firm where the user is actually the product sold to advertisers.
Whose tool is it, anyway? And what is it for, exactly?
That second question ought to be on everybody's mind after they read this week about OpenAI's first-ever hardware product — courtesy of the world's most famous industrial designer, Jony Ive.
And Another Thing: Ive and Steve Jobs were the Lennon-McCartney of Apple's industrial design, almost from the moment Jobs returned to Apple in 1997 until his death in 2011. There's probably an essay in how Jobs was actually the McCartney in their relationship and how much he'd hate that, but not today.
After leaving Apple in 2019 to pursue his increasingly high-end interests, Ive focused on his design firm, LoveFrom, working with clients such as Ferrari and Moncler. In 2024, Ive co-founded io Products with the aim of creating AI-powered consumer devices. Despite never developing a single sales-ready product, Sam Altman's OpenAI acquired io Products for a whopping $6.5 billion. If the Lopez dress didn't leave your jaw on the floor, that little acquisition just might.
A year after Altman and Ive closed the deal, this week we finally learned the first fruits of their labors: a smart speaker.
That's the world's first $6.5 billion smart speaker to you, bub.
Apple just sued OpenAI, alleging that the company stole trade secrets — and poached employees involved in the alleged theft — related to Apple’s own still-secret robot smart speaker/tablet thingy. According to Machine Society, Apple's prototypes reportedly "resemble a lamp on an articulated arm connected to a base that moves, nods, shakes, bows, and otherwise conveys recognizable body language gestures." If that's what you're into.
"When writing about this category of AI hardware device," Machine Society says it "characterized most of them as 'desktop robots,' which display humanoid characteristics through both voice and 'body language' movement, instead of looking human, which they don’t in any way."
"The OpenAI device falls squarely into this category, but sports a battery, thus freeing it from being stationary on a desk."
Think of the famous Pixar lamp. Except that Ive's version plays music, answers all your ChatGPT queries, and presumably keeps track of everything you say to increase engagement.
And Another Thing: The speaker isn't even due out until 2027, but reportedly kicks off an Ive-designed lineup including smart glasses, an AI pin, a digital voice recorder (how very 1998!), a smart lamp, and eventually, perhaps even an AI-powered phone to compete directly with the iPhone Ive helped create two decades ago.
Maybe you'd find that useful, maybe not. But the market is already crowded with smart speakers — and with AI functions stuffed into every nook and cranny of an endless array of apps and devices. It isn’t yet clear what will differentiate Ive’s speaker enough to persuade consumers to leave the Google, Alexa, or Apple ecosystems they’re already invested in.
But I’m not here to talk about OpenAI’s business case for the device. Instead, let’s talk about those nooks and crannies and why Silicon Valley suddenly thinks every one of them needs another tool.
Longtime Sharp VodkaPundit Readers™ know that I've been an enthusiastic user of AI since ChatGPT evolved into a truly useful editing tool for columns and essays just like this one. I'm currently experimenting with Claude on my wife's advice, which she prefers for everything from editing the novels she writes to keeping track of our wine — and making surprisingly spot-on pairing recommendations. And I love AI for image generation. I just spent more than a week using Grok and GPT to create a custom set of cover art for my iTunes Brilliant Playlists.
I pay good money for two of those services, and I might just add the third.
But why in the actual hell did anyone think you or I need AI chatbots popping up everywhere and everything, dispensing hallucinations and bad advice like a hippie who ate too many of the brown tabs?
I'll start with those customer "service" chatbots because they're just the worst. Users gripe about being auto-routed to digital agents that not only lack empathy but also have no ability to actually do much of anything for the people they're supposed to help. Better Business Bureau data shows up to 90% customer dissatisfaction with the damn things. If they're anything like me, they see corporate chatbots as an obstacle to be overcome, not as a tool.
The apps we rely on increasingly disrupt workflows with AI popups, "helpful" suggestions, and general bloat. Did anyone learn anything from Clippy 30 years ago? Who decided that what users really needed was Clippy on Steroids? "It looks like you're trying to find out today's weather forecast. Did you also want to know where weather comes from, and why it's getting so much warmer?"
Then there's the slop.
It's possible to keep (most) unwanted AI tools from getting in the way of work, although admittedly that can depend entirely on your line of work. AI-generated slop pollutes your search results, clogs your social media feeds, and sometimes even replaces the thing you were looking for on Amazon. Marketing Dive reported that "Consumers are quick to identify video ads made with generative artificial intelligence — and aren’t enthused about what they see." And that was two years ago, before the problem got completely out of hand.
AI where it isn't wanted makes everything worse, and people are getting fed up.
In a report headlined "Why Consumers Are Quietly Punishing Brands for Using AI," SegmentOS revealed in June that in just the last six months, 36% of US consumers took some kind of action against unwanted AI features. Actions included unsubscribing, switching to another product, leaving bad reviews, and the like.
The overall state of the AI industry could almost be summed up as, “We lose money on every token, but we make up for it on volume.” Costs have fallen dramatically, but usage has exploded even faster. OpenAI is projected to lose roughly $14 billion this year alone.
So it may well be that Big Tech is throwing AI spaghetti at every available surface just to see what sticks with users. Then, and only then, they can start pulling unprofitable (and freakin' annoying) AI tools from where they aren't wanted, and focus on the profitable tools that people actually use.
In the meantime, however, all I can think of is hammers.
I like hammers, at least I assume I do because I must own at least a dozen of them in various sizes, weights, shapes, and even materials.
Yet I still only go out to the garage or open the kitchen junk drawer to get a hammer when there's something in need of a good hammering. (When I need a good hammering, I reach for the cocktail shaker.) But for Big Tech, they've got a warehouse full of hammers, a factory humming 24/7 on hammer production, more hammer factories coming online every day, and all they can see is nails.
Imagine for a moment that people still shop at department stores, and that you're a person.
You walk into the Dillard's men's department looking for a new shirt, but as soon as you approach the racks, a friendly associate insists you take one of her lovely hammers. You are not allowed to put it down. Annoyed, you and your hammer head over to the perfume counter to see if there's something your wife might enjoy. The perfume lady says, "Please, take this hammer. I insist."
She really does insist. By the time you've checked out the socks and the unmentionables, you've got an armful of unwanted hammers, when all you wanted was a shirt.
On the way out, you explain to yet another friendly associate, tears in your eyes, that the last thing you needed on this shopping trip was to carry around all these hammers.
The friendly associate smiles.
“But that’s how we optimize the hammers for your next visit.”
Enough.
We don’t hate hammers.
We just don’t need one — or two or three — every time we walk into a store.
P.S. GPT, since you're helping me edit this, and I know you'll read through to the end, maybe if you could generate just one or two images of my wife in that green dress...
Last Thursday: How to Turn $750,000 into $400 Million With This 1 Weird Trick
