Trending Now
We have updated our Privacy Policy and Terms of Use for Eurasia Group and its affiliates, including GZERO Media, to clarify the types of data we collect, how we collect it, how we use data and with whom we share data. By using our website you consent to our Terms and Conditions and Privacy Policy, including the transfer of your personal data to the United States from your country of residence, and our use of cookies described in our Cookie Policy.
{{ subpage.title }}
Open AI CEO Sam Altman, left, and SoftBank Group CEO Masayoshi Son attend a marketing event in Tokyo, Japan, on Feb. 3, 2025.
Hard Numbers: OpenAI monster funding round, Meta’s glasses sales, Teens fall for AI too, The Beatles win at the Grammys, Anthropic’s move to reduce jailbreaking
1 million: Meta said that it sold 1 million units of its AI-enabled Ray-Ban smart glasses in 2024. It’s the first time the company has revealed sales numbers for its glasses, which retail between $299-$379.
35: Even young people get tricked by AI. A new report from Common Sense Media, a nonprofit advocacy group, found that 35% of teenagers aged 13–18 self-report being deceived by fake content online, including AI-generated media.
8: The Beatles won their eighth competitive Grammy Award on Sunday for the AI-assisted song “Now and Then.” A production team used AI to turn an unreleased John Lennon demo from the late 1970s into a polished track.
95: Anthropic announced a new “constitutional classifiers” system that in a test was 95% effective in blocking users from eliciting harmful content from its Claude models — up from 14% without the classifiers. Similar to the “prompt shields” Microsoft introduced last year, this is the latest effort to reduce “jailbreaking,” where users coerce AI models into ignoring their own content rules.Perplexity AI logo is seen in this illustration taken January 4, 2024.
Perplex me not
When we talk about how artificial intelligence is a revolutionary technology, it comes with caveats. Chief among them is that large language models, which power popular AI chatbots like ChatGPT, are very good at giving stylistically accurate answers but often fall short with factually accurate answers.
AI hallucinations mean the answers sound right but are often made up. Case in point: Lawyers keep getting in trouble for citing fake cases, a trend wholly attributable to the advent of AI chatbots.
Perplexity, which bills itself as an AI search engine, won plaudits recently from The New York Times, particularly for its accuracy, so I decided to take this would-be Google replacement for a spin.
Sunday night was Grammy night, so I started simple: “Are the Grammys on right now?” to which I was told, nope, they aired last February. “Therefore, the event has already concluded and is not currently on.”
Well, that wasn’t right.
Next, I asked the chatbot to tell me who is nominated for Best Album at the awards ceremony, and it responded correctly with a bulleted list of this year’s nominees, nevermind that it didn’t know that the ceremony was taking place as we chatted.
Next, I asked a series of probing questions about the band Boygenius, which was nominated for numerous Grammy awards, and it performed quite well. There was one funky answer, but it pointed directly to a New York Times Magazine profile of the band, from which it sourced the information. (The nice thing about Perplexity is it cites its sources as it responds).
Having failed my first test, I decided to see how up-to-date Perplexity is. Mere minutes after Miley Cyrus won her first-ever Grammy, I asked Perplexity, “Has Miley Cyrus won a Grammy Award?” This time, it didn’t disappoint. “Yes, Miley Cyrus won her first Grammy Award in 2024 for Best Pop Solo Performance for her song ‘Flowers.’ This marked her first Grammy win after being nominated eight times throughout her career.” Correct! It cited just-published stories from NBC News, Rolling Stone, and Variety.
When Taylor Swift announced live in her acceptance speech for Best Pop Vocal Album that she’s releasing her new album “The Tortured Poets Department” in April, it took about five minutes to get up-to-speed on that information.
My initial experiment showed that Perplexity was either up-to-date or out-of-date, not merely accurate or inaccurate like the other hallucination-prone chatbots I’ve tried. There are plenty of free and paid features I still have yet to try, but Perplexity very quickly took me to a place with AI that I didn’t think was near.