I used AI to do all my holiday shopping

In their initial responses, ChatGPT did not provide any product links. But it easily delivered them when I asked, and although I didn’t click on every single one, none seemed to hallucinations. Claude, on the other hand, apologized and said it “can’t actually link directly to websites or products.” Anthropic hasn’t released a web search feature for Claude yet, but the company says it’s working on it.

This technically made Claude the least useful chatbot I tested for shopping. But it also means that Anthropic has so far avoided entering the ethically murky territory of allowing its AI chatbots to scrub human-written product reviews from the web. Instead, Claude bases its product comparisons on the existing data set. Perplexity, on the other hand, says that thanks to Buy with Pro, people “no longer have to scroll through countless product reviews.”

When I asked Perplexity what to get my editor/musician friend, he recommended a set of solar bike lights (I also noted that he’s a cyclist). It wasn’t a bad idea, but it wasn’t exactly an important birthday present. I kept changing my prompt. How about a custom leather guitar strap? Down the rabbit hole I went.

I was beginning to understand that Perplexity’s purpose in blowing up its shopping features wasn’t just to help me come up with fresh ideas or come up with extremely thoughtful gifts. Perplexity is playing a long game, slowly shifting our attention away from competing corners of the web, gaining a better understanding of how people like me use its platform, and feeding that data into its ever-evolving AI models. Every time I needed to refine my searches, because the initial results were often missing, I was left in Perplexity’s app, which meant I wasn’t on Amazon and I wasn’t on Google (although I ended up on both sites). Perplexity Pro is not a full-fledged e-commerce site, nor is it an “agency” in any real way yet, but I am one of the millions of people who provide the information it needs to become these things.

When I turned to Google Gemini, I found that the gifts suggested for my 16-year-old niece weren’t bad per se, just uncreative and, in one case, confusing. It said I should buy her a “cat blanket to snuggle up with a good book,” but it wasn’t clear if the blanket was for her or her cat. Kindle was a good idea. But I’m terrified of what she’ll send me if I send her the SAT prep book suggested by Gemini (probably “thx” and nothing else). The app ideas for my editor/musician friend were equally uninspiring, with “Vinyl Records” and “High-end Headphones” among them.

I used the annual Gemini versionbut earlier this month Google began rolling out a newer version, Gemini 2.0, to developers and limited testers. The new AI model will “think several steps ahead and take action on your behalf,” the company said says. For now, that means taking action on behalf of developers — taking the next step in their coding workflows — but I’m looking forward to the day it can dig into my shopping list.

ChatGPT eventually led me to an online spice shop where I bought some special baking ingredients for my friend, who at that point I had built up in my mind as a finalist in The Great British Bake Off. I ended up talking to the AI ​​bots for so long that many of the gifts I chose wouldn’t arrive until after Christmas. My niece will receive money on a card. My search for an important birthday gift for a friend was inconclusive. I decided to continue the task until January, a month full of novelty and agent determination.