You’ve probably already noticed your search engines are starting to evolve. Google and Bing have already added both AI-generated results and conversational chatbots to their respective search engines. The Browser Company, a startup that made a big early splash thanks to its mission statement of building a better internet browser, has launched an AI summary search. And OpenAI is reportedly building its own search engine to compete directly with Google. Even Reddit, one of the last oases of human-generated advice on the web, will soon start selling its user data to an AI company.
Curiously, though, at no point amid our current AI arms race have the companies stuffing AI into our search engines and browsers have offered any guidance as to what happens to the web if this truly is the future of the way we find things online. And it may be the best evidence yet that the AI industry is still completely engulfed in hype.
Or as Glitch CEO Anil Dash tells me, “Why is everyone in the industry lemmings.”
The first shot fired in the war to build an AI portal to the web was from Microsoft. The company invested a billion dollars in OpenAI in 2019 and after the company’s buzzy generative-AI tools, DALL-E 2 and ChatGPT, were released to the public, Microsoft began trying to cram them into their respective suite of software.
In May of 2023, I was invited to the unveiling of Bing’s new AI search, powered by OpenAI’s GPT-4. I stood among a row of tablets in a Manhattan loft and asked the search engine basic questions, which didn’t get me very far. Eventually an employee working the event came over and showed me how to ask it in specific ways that could generate huge summaries of existing web content. Which sort of defeated the point of talking to a chatbot like it’s a person, if you ask me. Then, a few months later in July, Google began rolling out its own AI-powered search, Bard, worldwide and the race was on.
These AI-hybrid browsers and search engines have progressed considerably in the last year. Google’s Bard has since been renamed Gemini, and continues adding more features. A recent, extremely impressive demo of Gemini 1.5 Pro revealed that it’s now capable of using image recognition to search for specific scenes and movements in videos. Still, the summaries they generate aren’t necessarily better written than the blogs and websites they pull their information from. Nor do they even populate the chat window faster than a normal search would. But when they aren’t hallucinating, what they’re capable of is still impressive. Though, it’s a bit like watching a dog walk around on two legs — fun, but not exactly an efficient way to get around.
Dash, a long-time defender of the open web, has been particularly outspoken on social media and in the press about the rise of AI search and the danger it poses. “The industry needs to have new things to chase all the time because it sort of is dependent on hype cycles now,” Dash tells me. “Substance isn’t enough.”
He says that generative AI, right now, is not totally dissimilar from what happened during the cryptocurrency bubble during the height of the pandemic: Hundreds of startups, flush with cash from a bull market, started trying to build crypto-backed consumer products after they had already decided the technology was the future — not the other way around.
Dash says he’s also seen new startups make similarly mistaken pivots to AI just as a previous generation of rising tech companies pivoted to crypto.
“When the, you know, up-and-coming startup that everybody’s rooting for, that has the good vibes, does the thing that obviously is showing that they’ve got FOMO brain and does it wrong,” he says. “There was the Brave browser. And they added crypto.”
And now there’s the Arc Browser.
For years The Browser Company has been promising to save the internet. Its Arc Browser is a smart refresh of what a modern gateway to the web should look and feel like and it generated a lot of good will with early users. And then, earlier this month, they released their AI-powered search app, which “browses the internet for you.” The Browser Company did not respond to a request for comment.
The Browser Company’s new app lets you ask semantic questions to a chatbot, which then summarizes live internet results in a simulation of a conversation. Which is great, in theory, as long as you don’t have any concerns about whether what it’s saying is accurate, don’t care where that information is coming from or who wrote it, and don’t think through the long-term feasibility of a product like this even a little bit. Or, as Dash put it, “It’s the parasite that kills the host.”
The base logic of something like Arc’s AI search doesn’t even really make sense. As Engadget recently asked in their excellent teardown of Arc’s AI search pivot, “Who makes money when AI reads the internet for us?” But let’s take a step even further here. Why even bother making new websites if no one’s going to see them? At least with the Web3 hype cycle, there were vague platitudes about ownership and financial freedom for content creators. To even entertain the idea of building AI-powered search engines means, in some sense, that you are comfortable with eventually being the reason those creators no longer exist. It is an undeniably apocalyptic project, but not just for the web as we know it, but also your own product. Unless you plan on subsidizing an entire internet’s worth of constantly new content with the revenue from your AI chatbot, the information it’s spitting out will get worse as people stop contributing to the network. Which is something that’s already starting to happen.
“The biggest use case [for generative AI] so far is probably computer programming, but there is evidence that the quality of code is declining as a consequence,” Gary Marcus, cognitive scientist and CEO of Geometric Intelligence, tells me. “It’s also great for quickly writing knock-off books — Amazon is now flooded with them — and fake reviews and creating deepfake porn, etc. But that’s not a wonderful set of use cases.”
Marcus calls generative AI “a solution in search of a problem.” He also points out that no matter how good AI-driven search has gotten since Bing first rolled out its ChatGPT integration last year, it’s still constantly making mistakes.
“Last night, a tech CEO absolutely raved about how Google’s latest model summarized a MrBeast video,” he says. “And then within a few hours somebody checked carefully and the summary was full of errors.”
Making matters worse, if you’re hoping to prevent the eventual death of search, there won’t be a before and after moment where suddenly AI replaces our existing search engines. We’ve already seen how AI development works. It slowly optimizes itself in drips and drops, subtly worming its way into our various widgets and windows. And it’s likely we’re already living in the world of AI search and we just don’t fully grasp how pervasive it is yet.
Which means it’s not about saving the web we had, but trying to steer our AI future in the direction we want. Unless, like the Web3 bust, we’re about to watch this entire industry go over a cliff this year. Though, Kevin Donnellan, a writer who focuses on AI, doesn’t think that’s likely.
“People are using AI tools to help with the mundane stuff,” Donnellan says. “They may not shout it from the rooftops, they may feel uneasy about where this is heading, but if it’s shaving an hour off their working day, then what harm?”
Instead, Donnellan sees two possibilities for where this is all headed: Datasets and answers get better, compensation for creators materializes somehow, and we have a brand new — and actually better — way to find content online; or, we get a “ubiquitous AI search tool that repeats all the commerce-driven problems we see with today’s traditional search engines.” Under that latter scenario, he says, we’d see “SEO-hacks for the AI era, paid placements shunting out the best answers, and money trumping ingenuity every time.”
And it’s hard to shake the sinking feeling we already know where we’re headed.
