I’ve been following the story of generative AI a bit too obsessively over the past nine months, and while the story’s cooled a bit, I don’t think it any less important. If you’re like me, you’ll want to check out MIT Tech Review’s interview with Mustafa Suleyman, founder and CEO of Inflection AI (makers of the Pi chatbot). (Suleyman previously co-founded DeepMind, which Google purchased for life-changing money back in 2014.)
Inflection is among a platoon of companies chasing the consumer AI pot of gold known as conversational agents – services like ChatGPT, Google’s Bard, Microsoft’s BingChat, Anthropic’s Claude, and so on. Tens of billions have been poured into these upstarts in the past 18 months, and while it’s been less than a year into since ChatGPT launched, the mania over genAI’s potential impact has yet to abate. The conversation seems to have moved from “this is going to change everything” to “how should we regulate it” in record time, but what I’ve found frustrating is how little attention has been paid to the fundamental, if perhaps a bit less exciting, question of what form these generative AI agents might take in our lives. Who will they work for, their corporate owners, or …us? Who controls the data they interact with – the consumer, or, as has been the case over the past 20 years – the corporate entity?
Those of you who’ve been reading for a while may have noticed a break in my regular posts – it’s August, and that means vacation. I’ll be back at it after Labor Day, but an interesting story from The Information today is worth a brief note.
Titled How Google is Planning to Beat OpenAI, the piece details the progress of Google’s Gemini project, formed four months ago when the company merged its UK-based DeepMind unit with its Google Brain research group. Both groups were working on sophisticated AI projects, including LLMs, but with unique cultures, leadership, and code bases, they had little else in common. Alphabet CEO Sundar Pichai combined their efforts in an effort to speed his company’s time to market in the face of stiff competition from OpenAI and Microsoft.
Today I’m going to write about the college course booklet, an artifact of another time. I hope along the way we might learn something about digital technology, information design, and why we keep getting in our own way when it comes to applying the lessons of the past to the possibilities of the future. But to do that, we have to start with a story.
Forty years ago this summer I was a rising Freshman at UC Berkeley. Like most 17- or 18- year olds in the pre-digital era, I wasn’t particularly focused on my academic career, and I wasn’t much of a planner either. As befit the era, my parents, while Berkeley alums, were not the type to hover – it wasn’t their job to ensure I read through the registration materials the university had sent in the mail – that was my job. Those materials included a several-hundred-page university catalog laying out majors, required courses, and descriptions of nearly every class offered by each of the departments. But that was all background – what really mattered, I learned from word of mouth, was the course schedule, which was published as a roughly 100-page booklet a few weeks before classes started.
This past Monday NewsGuard, a journalism rating platform that also analyzes and identifies AI-driven misinformation, announced it had identified hundreds of junk news sites powered by generative AI. The focus of NewsGuard’s release was how major brands were funding these spam sites through the indifference of programmatic advertising, but what I found interesting was how low that number was – 250 or so sites. I’d have guessed they’d find tens of thousands of these bottom feeders – but maybe I’m just too cynical about the state of news on the open web. I have a hunch my cynicism will be rewarded in due time, once the costs of AI decline and the inevitable economic incentives that have always driven hucksters kick in.
Given 250 is a manageable number for a mere mortal, I decided to ask the good folks at NewsGuard, where I’m an advisor, for a copy of their listings. Nothing like a tour through the post-apocalyptic hellscape of our AI future, right?
What I found was…disappointing. Most of the sites were beyond shoddy – barely literate, obviously automated, full of errors and content warnings, and utterly devoid of any sense of organizational structure. The most common message, upon clicking on a story link, was some variation of an OpenAI violation:
Not exactly a compelling headline. The next most common experience was this:
This of course is evidence that the scammers are rotating URLs to avoid blacklisting, unburdened of any concern about building audience loyalty. Beyond OpenAPI warnings and 404s, there’s the browser warnings that the site you’re about to visit is, well, seedy:
When you do get an actual news experiences, it becomes clear that these publishers have little interest in passing as “real news sites,” IE publications a sane person might intend to visit. They are instead built as SEO chum in the hopes that Google’s indexes might favor them with some low quality traffic, or worse, as destinations for bot traffic destined for arbitrage inside the darker regions of the programmatic ad universe. The editorial decisions on the various home pages I visited were, well, hilariously inchoate:
Perhaps that’s what we should expect with the first phase of this particular genre, but I found their general awfulness depressing: Most reporters will look at these sites and dismiss them. But they shouldn’t.
Traditional “made for advertising” sites already control 21 percent of all programmatic advertising revenues, and these sites tend to dominate Google search results, enshittifying the open web with low-calorie crap that, one would hope, actually good AI might help us avoid. But the relatively low volume of AI sites indicates, at least anecdotally, that so far the economics of replacing human-built content with AI-driven drivel have yet to kick it. Put simply, it’s still too expensive to replace sites like Geeky Post or Explore Reference with AI. For now.
But when costs come down, I expect made for advertising sites will pivot to AI almost overnight. And I wonder if that’s a bad thing. Once the web’s worst sites all shift to AI-driven output, perhaps they’ll find themselves in a positive spiral of competition for actual human attention. If these sites start to create reasonably high quality content, and search and social start to reward them with real traffic that converts to revenue, perhaps we can simply automate away the shitshow that the open web has become.
I recently caught up with a pal who happens to be working at the center of the AI storm. This person is one of the very few folks in this industry whose point of view I explicitly trust: They’ve been working in the space for decades, and possess both a seasoned eye for product as well as the extraordinary gift of interpretation.
This gave me a chance to ask one of my biggest “stupid questions” about how we all might use chatbots. When I first grokked LLM-driven tools like ChatGPT, it struck me that one of its most valuable uses would be to focus its abilities on a bounded data set. For example, I’d love to ask a chatbot like Google Bard to ingest the entire corpus of Searchblog posts, then answer questions I might have about, say, the topics I’ve written about the most. (I’ve been writing here for 20 years, and I’ve forgotten more of it than I care to admit). This of course only scratches the surface of what I’d want from a tool like Bard when combined with a data set like the Searchblog archives, but it’s a start.
My friend explained that my wish is not possible now, despite what Bard confidently told me when I asked it directly:
Well, no. Bard hallucinated all manner of bullshit in its answer. Yes, I write about technology, but not the Internet of things. I guess I write about society, but mainly in the context of policy and consumer data, not “education, healthcare, and the environment.” Culture? When’s the last time you’ve seen me write about movies?! And if I ever start writing about “personal development,” please put one between my eyes.
Bard’s list of supposed articles was even funnier – it reads like an eighth-grade book report culled from poorly constructed LinkedIn clickbait. Bard is a confident simpleton, despite its claim to be able query specific domains (in this case, battellemedia.com). I responded to Bard with this new prompt: “This is not right. That site does not cover music, movies. Nor does it do motivation, well being, productivity. Why did you answer that way?” Bard’s answer was … pretty much the same, though it did clumsily incorporate my corrections in its response:
Gah. My next prompt was an attempt to clarify where Bard was getting its answers, since it was clearly not using the battellemedia.com domain. “Are you actually referring to content on the site to do these answers?”
Ok, then, at least we’re getting some honesty. I decided to try one last time:
Now this was quite the freshly whipped bullshit: Actual percentages of how the content on my site breaks down! Unbeknownst to me, more than one in ten of my posts are about cybersecurity – a topic I’ve rarely if ever written about here.
Ok, enough beating up on poor Bard. My well-placed friend explained that while it’s currently out of scope for a standard chatbot like Bard or ChatGPT to do what I’m asking of it, “domain specific” queries was a hot area of development for all LLMs. So when will it happen? My friend didn’t commit to an answer on that, but I did get the sense it’s coming soon. The ability to apply LLM-level intelligence to large data sets is just too big an opportunity – in both B2C as well as B2B/enterprise markets.
A big reason this is taking more time than I’d like is cost. Noted AI investor Andreessen Horowitz recently posted a long explainer on the state of LLM models, but it all comes down to this money quote: “Today, even linear scaling (the best theoretical outcome) would be cost-prohibitive for many applications. A single GPT-4 query over 10,000 pages would cost hundreds of dollars at current API rates.” By my estimates, this cost would need to come down at least four orders of magnitude – from hundreds of dollars per query to pennies – to unlock the kind of magic that I’ve been dreaming about over the past few months. Not to mention all the technological machinations related to prompt handling, vector database management, orchestration frameworks, and other stuff that makes my brain hurt. But the good news, despite my rather pessimistic post from earlier this week, is that the good shit’s coming – we just need to be a bit more patient.
Not since the iPhone, in the mid aughts. No, not since the rise of the browser and the original web, in the early nineties. No, not since the introduction of the PC, in the 1980s. Ah hell, honestly, not since the Gutenberg printing press in the 15th century – or, fuck it, let’s just go there: Not since the invention of language, which as far as we know marked the moment when homo sapiens first branched from its primate cousins.
That’s how big a deal AI is, according to academics, politicians, and a rapt technology and capital ecosystem starved for The Next Big Thing.
I tend to agree. First we created language, then we created its digital doppelganger with computer code, and with generative AI, we’re melding the two into a shimmering and molten fun house mirror, one that forces us to question our very consciousness. What the hell does it mean to be human when we’ve created machines that seem to transcend humanity?
“…the Digital Revolution is whipping through our lives like a Bengali typhoon…[bringing] social changes so profound their only parallel is probably the discovery of fire.”
Ah, fire. I forgot about fire, which likely preceded language by a good 50,000 years. Those lines introduced the very first issue of Wired magazine 30 years ago. As founders we were convinced every aspect of society would be reshaped – our culture, our economy, our social lives, our faiths, our sense of self. In those early days we were essentially a cult, a non-denominational sect stoned on a buoyant certainty that we were right – that technology offered all of us an offramp from the tired shit-show of the industrial revolution. Of course the Internet was going to rewire everything – it was obvious. If you didn’t see that coming, you just weren’t paying attention. Our job was to slap you into seeing what was right in front of our eyes: The future, coming fast, screaming into our face with possibility and promise.
And now, here we are. The starting gun has been fired once again- this time the release of ChatGPT. After a decade of trillion-dollar platform consolidation based on surveillance capitalism and trickle-down innovation, tech once again brims with optimism, with that original possibility and promise.
If, that is, we don’t fuck it up by forcing our new tools into the structures of the past.
Yesterday Fred posted about voice input over on AVC, and it reminded me how long it takes for consumers to adopt truly new behaviors, regardless of how enthusiastic we might get about a particular technology’s potential. As Fred points out, voice input has been around for a decade or so, and yet just a fraction of us use it for much more than responding to texts or emails on our phones.
While tens of millions of us have begun to use generative AI in various ways, its “paradigm shifting” impacts are likely years away. That’s because while consumers would love to have AI genies flitting around negotiating complex tasks on our behalf, first an ecosystem of developers and entrepreneurs will have to do the painstaking work of clearing the considerable brush which clogs our current technology landscape – and it’s not even certain they’ll be able to.
Some historical context is worth considering. When the World Wide Web hit in 1993, I was convinced this new platform would change everything about, well, everything. Culture, business, government – all would be revolutionized. 1993 was the year Wired first published, and we took to the technology with abandon. We launched Hotwired, one of the first commercial websites, in 1994- but quickly realized the limitations of the early Web. There was no way to collect payment, serve advertising, or even identify who was visiting the site. All of those things and more had to be invented from scratch, and it took several years before the entrepreneurial ecosystem ramped up to the challenge. Then, of course, the hype overwhelmed the technology’s ability to deliver, and it all came crashing down in 2001.
Fast forward to the launch of the iPhone in 2007, and once again, everyone was convinced the world was going to change dramatically. But Airbnb launched in late 2008, Uber in 2009, and both didn’t gain widespread traction until 2011 or 2012. It took another seven to nine years for these two stalwarts of the mobile revolution go public. Along the way tens of thousands of smaller companies were building apps, exploring new opportunities, and generally laying the groundwork for the world as we know it today. But to win, they learned that they had to play by the increasingly rigid policies of the dominant platforms: Apple, Google, Amazon, and Facebook. The dream of “Web 2” – where the Internet would be an open platform allowing innovation to flourish – never truly materialized. The platforms became some of the largest corporations ever to roam the earth, and quite predictably, enshittification followed.
So while many of us are currently enraptured with the rise of generative AI, it’s worth remembering that despite the technology’s huge potential, this will all take time. And unlike 1993, when the Internet was literally a blue ocean opportunity, or 2007, when smart phones were as well, this time everyone’s in on the joke. Yes, billions upon billions of venture capital is now being deployed against what feel like unlimited opportunities in the space, but these new startups will have to battle deeply entrenched incumbents with almost no interest in seeing their moats breached.
Thirty years after the first issue of Wired, it’s still making for one hell of a story.
A few weeks ago I was genuinely thunderstruck. My co-editor atP&G Signal (thanks Stan!) introduced me to a new company – one that promised to give consumers control over their personal data in new and innovative ways. At first I was skeptical – I’d seen quite a few “personal data lockers” come and go over the past decade or so. I even invested in one way back in 2012. Alas, that didn’t work out.
For as long as I can remember, I’ve been writing –over andover andover – about how the Internet’s central problem is the lack of leverage that consumers have over the data they co-create with the hundreds of apps, sites, and platforms they use. But data lockers never got any traction – most were confusing to install and run, and they all suffered from a lack of tangible consumer benefits. Sure, having a copy of all my personal data sounds great, but in the end, what can it do for me? Up till now, the answer was not much.
It was with all those caveats – and honestly pretty low expectations – that I took a meeting with Sumit Agarwal and his team at Palo Alto, CA-basedGather, an early stage startup still in its first year of operation. Fifteen minutes later I was hooked – here was a company that was addressing the “what can my data do for me” problem by building out a generative AI agent that just might spark the kind of personal data revolution I’ve been writing about for more than a decade. And this was no fly-by-night startup – the company’s founders, team, and investors are all deeply experienced in AI, Internet security, scaled engineering, product design, marketing, and much more.
Before diving in, a caveat: Gather is still at a very early stage, as is the overheated AI ecosystem in which Gather’s products will eventually live. Agarwal told me he’s not even sure if his company will be called Gather by the time its first product becomes available later this year. In addition, the company faces fearsome obstacles to success – including entrenched platform players like Google, Amazon, and Apple, whose business interests do not align with the concept of a newly empowered consumer base. While I usually like to write about companies and products that readers can use immediately, I’m breaking that rule for Gather. No matter the business you’re in, it will pay to understand the shift in consumer behavior that tools like Gather could unlock. And as I said before, this company is the first I’ve seen that has assembled the team, vision, and execution chops to pull it off.
Timing Is Everything
With startups, timing is everything. There’s only so much you can control – what you make, how you spend your investors’ money, the people you hire. But nearly everything else is driven by externalities you must navigate. Is the technology ecosystem capable of supporting your vision, or is your product ahead of its time? Netflix, for instance, had to wait until broadband was pervasive enough to launch its streaming service. Are consumers ready for your idea, or is it out of sync with their expectations? Uber and Airbnb faced this challenge in their early years. Will huge competitors copy your idea, or change their policies and make it impossible for you to thrive? Ask Yelp how it feels about Google’s review summaries, or ask Epic Games about Apple’s 30 percent tax in the app store.
Gather faces all these timing challenges and more, but the company does have one huge tailwind: AI is hot, and investors can’t get enough of it. This past month alone,VCs poured more than $11 billion into AI startups, up 86 percent from a year ago. But while AI funding tipped into a frenzy with OpenAI’s launch of ChatGPT last November, Gather managed to raise an impressive seed round five months earlier, in June of 2022. Agarwal and several of his co-founders were already seasoned operators with a billion-dollar exit in the Internet security sector (Shape Security, sold to F5 three years ago). Gather’s $9 million round was led by general partners at respected firms Bain, Floodgate, and Wing Ventures, with participation from experienced Valley angels like Gokul Rajaram – an investor and director at The Trade Desk, Pinterest, and Coinbase, among many others – and Vivek Sharma, the co-founder and CEO of Movable Ink.
“I’ve known Sumit for about 15 years,” said Gaurav Garg, founder of Wing Venture Capital and early investor and board member at Gather. “He has a deep background in consumer and enterprise products used by hundreds of millions of people, as well as exposure to government policy, across technology areas including security, identity, privacy, and e-commerce.” Garg also noted that Sumit has the attributes of a great founder – drive, persuasiveness, perspective, and a learning mindset – crucial when you’re looking to reimagine something as big as how consumers will interact with AI and the Internet.
“The team are all exceptional founders,” added Bain Capital Venture’s Ajay Agarwal, who’s known and worked with several of the Gather team over the past 25 years. He added a key observation: The tech ecosystem is at an inflection point – mainstream devices like phones and computers can now power distributed platforms like Gather, large language models have evolved to conversational levels, and there’s even just the right amount of government regulation to create conditions for a sea change in how consumers control their data. We’ll get into all of that, but first, the product.
First, Gather The Data
Gather acts as your trusted and secure agent, logging into various data-hoarding services like Google (think Maps, Search, Mail, Android Play, YouTube), Amazon, Uber, Strava, and many more. At your direction, Gather then downloads copies of your data from each service to your local device – a right codified into law by the 2018European General Data Protection Regulation (GDPR) and adopted, in broad strokes, by several states in the US –California chief among them. Till Gather came along, no one had built a service that automates what is otherwise a tedious and frankly pretty pointless process – almost no one actually downloads copies of their data from online services, because, as we’ve already established, there was simply no use case for doing so.
But once you have a critical mass of that data, and it’s organized in a way where questions can be asked of it, a whole new world opens up. Gather added me to a pre-release version of its platform, and it was magical to watch the service engage with Amazon, Uber, Google, Twitter, Venmo, Strava, and many others. Within minutes, copies of my data were presented and organized in my Gather app. And that’s when things start to get interesting. As Gather marketing lead Niki Aggarwal pointed out, now that I had the data in my control, I could start to ask it questions. What kind of bias, if any, might be evident in the stories I was reading from Google News? How much did I spend on Amazon each month, and would I have saved money if I had bought those same items at Walmart? Was there an optimal time of day to get a personal best when riding on Strava?
Of course, this is only the tip of the proverbial iceberg when it comes to what’s possible with data sets like these. Sure, it’s cool to have a copy on my own device, in my control. But the real fun will start when you add a personalized AI agent capable of instantaneously answering those initial questions, as well as conjuring up ones you’ve not thought to ask. Even more exciting, imagine that same agent as a trusted confidant, leveraging your data as it interacts with the rest of the online world. Now that’s a consumer benefit – a personal agent that knows my preferences and can, say, plan a complicated business trip, or negotiate for the best price for an item I want to buy online. The time savings alone make the idea compelling. I’ve come to call such agents “genies” – because they work only for you, and they can produce all kinds of magic (and as a bonus, they aren’t limited to three wishes!).
Where Genies Play
Gather hasn’t released its “genie” yet, but it’s working on it. Codenamed “Sidekick,” the product will consist of several elements. First is your personal datastore, which lives on your own device and remains under your control at all times. Second is the Sidekick agent, which Agarwal describes as “an AI product that proactively and intuitively helps you.” He continues: “We are trying to get away from ‘blinking cursor in an empty text box’ and get to ‘intelligent character that thinks about your needs, intuits your desires, [and] acts on your behalf.'” Gather’s third element is a platform that manages how outside organizations interact and create value with your data.
Agarwal offers an example of how Sidekick might work: “You visit Amazon Music, and there’s an offer for six months of free service if you upload your Spotify play history. Your Gather Sidekick allows you to upload with a single click. Your data moves to an external service (Amazon Music) and you get value. You can think of the same example in many contexts – food/dieting, fitness, other entertainment apps, medical apps, etc. The concept is simply that a specific – and sometimes complex – “slice” of your data needs to move somewhere in order for you to receive some value (economic or otherwise).” (This example calls to mind my 2018 piece, in which I imaginedhow Walmart might compete with Amazon online by leveraging a consumer’s Amazon purchase history.)
Agarwal’s second example envisions an instance where you don’t necessarily trust the service that wants to leverage your data. Imagine, for example, that a third-party developer has created “the ultimate music recommendation app.” It sounds appealing, but you’re wary of uploading your Spotify or Apple Music data to a service that has yet to prove it’s trustworthy. In this case, Gather becomes a secure platform that runs on your device. “With Gather,” Agarwal explains, “that recommendation app can run locally in your environment. This is a win for you because you have no data sharing concerns, so you can comfortably let the app engage with your data to get the very best recommendations. The Gather platform keeps your data local but publishes the schema so developers know how to interact with the platform – without seeing your data.”
I think of Gather’s platform as like Apple’s app store or Google Play, but with one critical difference: The power to decide who gets access to the platform resides not with a massive corporation, but with you, the consumer. This seemingly small distinction is in fact a massive shift in power, agency, and value from the centralized model of Web 2 toward a decentralized vision more aligned with the original architecture of the Internet –pushing intelligence and control to the edge of the network.
The Iceberg Metaphor
Over the course of many emails, calls and Zoom meetings with Agarwal and his co-founders Sudhir Kandula and Mengmeng Chen, both colleagues with Agarwal at Shape Security, we touched on topics as varied as security and privacy, Internet history, and information theory. Gather emerged from its founders’ dissatisfaction with what author and Internet OG Cory Doctorow calls the Internet’s “ensh*ttification.” As large companies have consolidated control of our online lives, our experiences have begun to degrade. This is why so many technology observers are excited by Microsoft’s integration of ChatGPT into its Bing search service. Google search is so clogged up with low-quality results and ads, Microsoft’s “conversational search” promises a better, clutter-free user experience. But Agarwal sees many more use cases beyond search – and to understand how it might work, it’s worth a dive into what he calls “the iceberg metaphor,” a visualization of how humans might best communicate with infinitely capable AIs.
As we know, 90 percent of an iceberg is underwater. At its tip – the 10 percent – is human interaction with AI – the prompts we type, or soon, the words we utter. That interaction is limited by our ability to speak or type – which compared to machines, is very low bandwidth, about 50 words per minute. But human speech is richly nuanced, and informed by executive function – this is where decision making occurs.
It’s in the 90 percent underwater where AI can excel. Machines can speak to other machines at mind bending speeds – one AI genie speaking to countless others, negotiating information demands, price comparisons, complex, multi-step transactions like scheduling a meeting or building a travel itinerary. “Underneath the water the GPT is listening, watching, reading, and comprehending on your behalf,” Agarwal says. “It’s unconstrained by our puny 50 word-per-minute input.”
The key to the iceberg model is that ten percent – no substantive decisions are taken, no meaningful action, until the human in charge says so. Good genies will surface questions and clarifications at the speed of human language, then dive back below the surface to negotiate next steps in the underwater world of machine-to-machine communication.
Will Tech Giants Let It Happen? The Plaid Example.
For Gather to scale, it needs hundreds of thousands, if not millions, of people to engage with its platform. Agarwal is reserved when pressed on the use cases that might drive that engagement, but it’s not hard to imagine any number of “killer apps” that could get the startup to its first million users. But as challenging as scaling an initial user base may be, Gather faces an even larger threat:The data use policies of big platforms like Amazon and Google. Regulations like GDPR guarantee a user’s right to access and download their own data, but most big tech platforms have “terms of service” policies that prohibit automated retrieval of user data. These policies are ostensibly in place to counter malicious actors who are spoofing real user’s accounts, but they could also be employed to stymie Gather’s work on behalf of its user base.
This is where the Gather team’s experience at Shape Security comes into play. Shape’s core business was to help large financial institutions fight automated attacks against the banking industry’s consumer portals. Agarwal and his colleagues spent years understanding and perfecting defenses against sophisticated “attack vectors” in a sector where the stakes are high – people do not like to lose their money. For much of their time at Shape, one of their most vexing opponents was a company called Plaid – then a startup, but now a $15 billion industry leader that offers consumers a platform to retrieve, manage, and gain value from their personal financial data across a majority of banking institutions online. If you’ve ever used RobinHood, or moved money from your bank to Venmo, you’ve used Plaid. Like Gather, Plaid works as an agent on behalf of its individual customers. For years big banks fought against the idea of their own consumers taking control of their data, and Shape Security was one of their most potent weapons. While Shape was able to win at the tactical level – stopping Plaid from accessing data on behalf of clients like Capital One –Plaid managed to win the overall war, because its ardent users pressured their financial institutions (and regulators) to allow them access to their own data.
The Plaid example can’t but be front and center in the Gather team’s minds as they embark on the next phase of their journey. Agarwal, a former Network Warfare Officer with the United States Air Force, often uses military terminology when describing the state of consumer data rights. “In the military there’s a term called preparing the battlefield – months and months of preparatory work before you commit,” Agarwal says. “As soon as we finished up at our acquirer, we started brainstorming about how to protect more users in more profound ways.” The idea for Gather, he said, hit him “like a lightning bolt” and work on Gather began almost immediately afterward.
What’s In It For Brands?
Agarwal is committed to making Gather free for its users, a tactic that will certainly help the company garner its initial user base. Once a critical mass of consumers are on the platform, he envisions charging enterprises API fees when they reach out to consumers and request consumer data, with the consumer’s permission, of course. As Agarwal imagines it, brands might want to offer promotions much like the example he mentioned above – where Amazon Music offers six months free in exchange for a user’s Spotify data. Walmart might do the same in a bid to lure away Amazon customers. But the examples can get even more granular – McDonald’s might offer otherwise hard to reach consumers in a certain zip code free delivery via DoorDash, or a company like P&G might pilot a Pampers subscription service based on a user’s past purchase data. The possibilities are infinite – if Gather gets to scale.
Should he succeed, Agarwal and team are hoping to jumpstart an entirely new value equation for consumer-driven data, one that just might force all businesses to abandon today’s dominant model of hoarding data and steering consumers into a limited set of choices. “Today, our data is so siloed, but it’s so valuable,” Agarwal told me. “We can change the power balance from the platform back to the user.”
Long time readers know how I feel about Agarwal’s sentiment – I believe unleashing the consumer data economy could drive a huge increase in economic innovation and flourishing. I suspect this won’t be the last time I write about Gather, and certainly not the last time I write about the sea change it hopes to spark. In the meantime, Agarwal will be presenting his vision for Gather at P&G Signal this coming July 12th. You can find a registration link for the event here.
Well that was something. Yesterday the Center for AI Safety, which didn’t exist last year, released a powerful 22-word statement that sent the world’s journalists into a predictable paroxysm of hand-wringing:
“Mitigating the risk of extinction from A.I. should be a global priority alongside other societal-scale risks, such as pandemics and nuclear war.”
Oh my. I mean, NUCLEAR WAR folks. I mean, the END OF THE WORLD! And thank God, I mean, really, thank the ever-loving LORD that tech’s new crop of genius Great Men – leaders from Microsoft, Google, OpenAI and the like – have all come together to proclaim that indeed, this is a VERY BIG PROBLEM and not to worry, they all very much WANT TO BE REGULATED, as soon as humanly possible, please.
The image at top is how CNN responded in its widely read “Reliable Sources” media industry newsletter, which is as good a barometer of media groupthink as the front page of The New York Times, which also prominently featured the story (along with a requisite “the cool kids are now doing AI in a rented mansion” fluff piece. Same ice cream, different flavor).
But as is often the case, the press is once again failing to see the bigger story here. The easy win of a form-fitting narrative is just too damn tasty – confirmation bias be damned, full steam ahead!
So I want to call a little bullshit on this whole enterprise, if I may.
First, a caveat. Of course we want to mitigate the risk of AI. I mean, duh. My goal in writing this post is not to join the ranks of those who believe AI will never pose a dire threat to humanity, or of those waiting by their keyboards to join the singularity. My point is simply this: When a group of industry folks drop what looks like an opportunistic A-bomb on the willing press, it kind of makes sense to think through the why of it all.
Let’s review a few facts. First and foremost, the statement served as a coming out party for The Center for AI Safety, a five-month old organization that lists no funders, no phone number, and just a smattering of staff members (none of whom are well known beyond its academic director, a PhD from Berkeley who also joined five months ago). Its mission is “to equip policymakers, business leaders, and the broader world with the understanding and tools necessary to manage AI risk.” Well Ok, that’s nice but…who exactly is doing all that equipping? And where might their loyalties and incentives lie? And do they have any experience working with real life governments or policy?
Hmm. Did The New York Times, CNN, or The Verge ask about this in their coverage yesterday? Nope. Strange, given the last time we saw a similar effort, it turned out that the organization behind it was funded in part by Elon Musk. The golden rule of journalism is Follow The Damn Money.
OK, next. Look at the signatories. A ton of well-meaning academics and scientists, and plenty of previously vocal critics of AI (Geoffrey Hinton being the most notable among them). OpenAI’s network is all over the list – there are nearly 40 signatories from that company alone. OpenAI partner Microsoft only mustered two, but they were the two that mattered – the company’s CSO and CTO. Google clocked in with nearly 20. But not a one from Meta, nor Amazon, Apple, IBM, Nvidia, or Snowflake.
Hmmm. Did any of the mainstream media pieces note those prominent non-signatories, or opine on what they might imply? Only in the case of Meta’s Yann LeCun, who is already on record stating that AI doomsday scenarios are “completely ridiculous.”
So what’s this really all about? Well, in a well-timed blog post just last week about how best to regulate AI, OpenAI’s CEO Sam Altman called for “an International Atomic Energy Agency for superintelligence efforts.” There’s that nuclear angle, once again – this AI stuff is not only supremely complicated and above the paygrade of mere mortals, it’s also as dangerous as nuclear fissile material, and needs to be managed as such!
Altman’s testimony at Congress two weeks ago, his blog post equating AI with nukes the week after, and then this week, the newly minted Center for AI Safety’s explosive statement – come on, journalists: Can you not see a high-level communications operation playing out directly in front of your credulous eyes?
Before I rant any further, two apparently contrary ideas can in fact both be true. I am told by folks who know Altman that he truly believes “super-intelligent” AI poses an existential risk to humanity, and that his efforts to slap Congress, the press, and the public awake are in fact deeply earnest.
But it can also be true that companies in the Valley have a deep history of using calls for regulatory oversight as a strategy to pull the ladder up behind themselves, insuring they alone have the right to exploit technologies and business models that otherwise might encourage robust innovation and by extension, competition. (Cough cough privacy and GDPR, cough cough). Were I in charge of comms and policy at OpenAI, Google, or Microsoft, the three current leaders in consumer and enterprise AI, I’d be nothing short of tickled pink with Altman’s heartfelt call to arms. Power Rangers, Unite!
I’ve written before, and certainly will write again, that thanks to AI, we stand on the precipice of unf*cking the Internet from its backasswards, centralized business and data models. But if we declare that only a few government-licensed AI companies can pursue the market – well, all we’ve done is extend tech’s current oligarchy, crushing what could be the most transformative and innovative economy in humankind’s history. I’ll save more on that for the next post, but for now, I respectfully call bullshit on AI’s “Oppenheimer Moment.” It’s nothing of the sort.
The tech press has breathlessly speculated that, freshly invigorated thanks to ChatGPT, Microsoft’s Bing might steal a major distribution partner from Google. First it was Samsung (wrong), then it was Apple (unlikely), and always there was Firefox, with its 200 million monthly users and its tumultuous relationship with its Googley paymaster.
But The Information’s reporting includes a twist: While Google and OpenAI declined comment, Firefox Chief Product Officer Steve Teixeira is quoted saying his company could add AI-driven chat without violating its current deal with Google – or any future deal, given the Google deal is reportedly up later this year. From today’s story: “Teixeira said his company’s search deal with Google doesn’t pertain to conversational technologies, and the chatbot arrangement would be separate from traditional search.”
Teixeira goes on to make any number of claims about how search and conversational interfaces are essentially different use cases – a useful fantasy that feels increasingly hard to defend. “…the mainstream is really accustomed to getting a search engine results page, with lots of results, and changing that behavior for mainstream people is going to take some time.”
Michellebegs to differ. A huge chunk of search is already poised to be swallowed by chat interfaces, and I’d argue it’s only a matter of time before the ten blue links becomes a secondary destination – one you go to only after consulting your AI agent. One of my principal frustrations with early chatbots like Pi is that they don’t have the ability to quickly search something in the context of our chats. As I’ve argued before, search is on its way to commoditization, and more likely than not, we’ll all end up paying for chat bots the way we pay for other valuable information services – as a monthly subscription.
To get there, major platforms like Google, Amazon, Apple, and Facebook will need to roll out chat integrations inside search – and three of four of those have no reason not to – only Google has a cannibalization dilemma with its core search model. But for now, Google has the lion’s share of search-related attention, and therefore, the most to both win and/or lose as consumer behavior shifts toward chat. Which is all a long way of saying this: When Firefox does announce its first chat integration, more likely than not it’ll come from Google.
Last week I was traveling – and being in four places in six days does not make for a good writing vibe. But today I’m back – and while the pace is picking up for the annual Signal conference I co-produce with P&G, I wanted to take a minute to reflect on last week’s news – no, not that CNN shitshow, but Google’s big I/O conference, where the company finally revealed its plans around search, AI, and a whole lot more.
Leading tech analyst Ben Thompson summarized how most of the pundit-ocracy responded to Google I/O: “the ‘lethargic search monopoly’ has woken up.” He also noted something critical: “AI is in fact a sustaining technology for all of Big Tech, including Google.” Put another way, the bar has been reset and no one company is going to own a moat around AI – at least not yet. Over time, of course, moats can and will be built, just as they were with core technologies like the microprocessor, the Internet itself, and the mobile phone. But for now, it’s a race without clear winners.
Head to The Verge if you want a summary of what went down at I/O – beyond AI, Google doubled down on devices – positioning itself as a serious competitor to Apple (I’ve been a Google Pixel user for years, and all I want is for the two companies to figure out how to deliver a text…).
But we’re all about AI and search here at Searchblog, and damn, there was finally some real talk about how the peanut butter and chocolate would be combined. As The Verge put it, “The single most visited page on the internet is undergoing its most radical change in 25 years.” From the story:
Called the Search Generative Experience (SGE), the new interface makes it so that when you type a query into the search box, the so-called “10 blue links” that we’re all familiar with appear for only a brief moment before being pushed off the page by a colorful new shade with AI-generated information. The shade pushes the rest of Google’s links far down the page you’re looking at — and when I say far, I mean almost entirely off the screen.
That seems radical, but for commercial searches (Google defines that term, not me), ads will still be front and center. And that’s a key distinction. The majority of searches are not commercial, and for those, Google promised an AI-driven summary at the top – an evolution of the “snippets” and one boxes that we’ve become accustomed to. It’s a clever hack – for most of our searches, it’ll seem like Google’s familiar ten blue link interface has been replaced. But commercial searches will continue to feel, well, commercial. Google’s really pushing what that will look like, for a taste, check out this short video:
These changes have yet to be rolled out (they’re coming “soon” and you can sign up for the sandbox with a personal Gmail account here), but with these revelations certainly gave Google’s extended ecosystem a new set of opaque signals to read. Millions of SEO-driven websites will have to recalibrate against Google’s black box algorithms and hope they can continue to win placement and the revenue that comes along with it. To you all, I say bon chance.