Google’s new AI model Gemini goes up against ChatGPT

Hello, fellow humans! You’re reading our limited series of Saturday Daily Briefs. While it’s focused on AI, it’s curated, written, and edited by actual people.

Got some questions about AI you’d like answered? Or just some AI hallucinations you’d like to share? Email us anytime. Enjoy!

Here’s what you need to know

There’s a new chip on the block. Meta, Microsoft, and OpenAI said they’ll be opening their pocketbooks to AMD’s Instinct MI300X as an alternative to Nvidia’s products.

Elon Musk probably wouldn’t mind some open pocketbooks. His startup xAI is seeking to raise $1 billion in funding.

EU countries and lawmakers reached a deal on AI regulation. The discussions had been held up on disagreements over using the tech in biometric surveillance.

McDonald’s will serve up AI with each meal. The global fast food chain, which is expanding its footprint by 10,000 stores, is working with Google to make “hotter, fresher food.”

Banks have a $340 billion incentive to adapt AI quickly. At least that’s how much global consultancy McKinsey thinks the industry could add to its earnings each year thanks to increased productivity.

Google’s most capable AI still can’t beat ChatGPT

Google’s new AI model, Gemini, which it describes as its most capable and general-purpose generative tool yet, has set its sights on OpenAI’s GPT models. Gemini Pro has been folded into Bard already—and we gave it a spin alongside ChatGPT in “Ask an AI” at the bottom of this email.

Google says Gemini is not just a competitor but a winner, reporting that Gemini Pro outperformed GPT-3.5 in six out of eight industry benchmarks. Google’s most advanced model, Gemini Ultra did better, beating the newer GPT-4 in seven of the eight benchmarks.

Google’s recent evaluations of Gemini suggest that OpenAI’s competitors are starting to catching up to the AI gold standard.

Don’t be alarmed, it’s just ✨

Next time you shuffle your Spotify playlist or jump on a Zoom, see whether you notice this: ✨.

Perhaps your eagle eye has noticed it already, and perhaps you’re curious about what clicking on that button will do, and what ✨ even means? In the relatively short history of emoji, sparkles have been used to express excitement and magic. It’s an ethos that tech companies are using to label which of their services use artificial intelligence.

Naturally, branding new AI products with the ✨ emoji suggests that these tools are exciting and magical and not at all alarming. But will skeptics click? Remember, before the Internet became the internet, it was the great, expansive World Wide Web. Maybe one day AI won’t even need distinction.

Quotable: High times at OpenAI

“While we had expected to release it this month, a few unexpected things have been keeping us busy!” —An OpenAI email obtained by the Verge discussing the delay of the company’s chatbot store until next year.

You can probably guess what’s been keeping them so swamped.

Other great AI reads

👀 Bank of England will review the risks that AI poses to UK financial stability

🤔 Europe was set to lead the world on AI regulation. But can leaders reach a deal?

💬 AI’s future could be open-source or closed—and that’s leaving tech giants divided as they lobby regulators

🤖 Wikipedia’s most popular article is also its biggest threat

🤑 AI poses less of a threat to jobs than it does to wages

Ask an AI

We pitted Google’s Gemini (via Bard) against OpenAI’s ChatGPT 4 in a highly unscientific three-round challenge.

⚔️ 1. Apply Adam Smith’s philosophy to the AI wars. When asked to rewrite the first sentence of the landmark economic text The Wealth of Nations to describe AI competition, the two generative AI chatbots gave us very different results:

ChatGPT: “In the grand journey of AI competition, it is not from the benevolence of the machine, the developer, or the coder that we expect our algorithms to be efficient, but from their regard to their own interest.”

Bard: “Just as the power of exchanging data fuels the advancement of artificial intelligence, the extent of this advancement is ultimately limited by the intensity of competition among AI systems, which, in turn, depends on the size and openness of the digital marketplace.”

The disparity stems from the fact that both were using a different “first” sentence as reference—the one Bard was looking at actually heads up the third chapter of the book.

🔍 2. Look at an image and find out some information about it. Gemini adds image capabilities (though not image-generating capabilities like those of DALL-E) to Bard. So we showed both chatbots a photo of the spines of three books in Tamsyn Muir’s Locked Tomb series and asked when, for the love of bones, the next book will be released. Bard confidently said that, according to the author’s X account, the drop would be Sept. 10, 2024. ChatGPT said Jan. 1, 2024.

But, maddeningly, the truth is that there is no release date set for this book. Muir certainly didn’t indicate anything of the sort on her X account, which she has not updated since 2020.

🪶 3. Generate some not-great poetry. As one must, we asked the two rivals to write Shakespearean sonnets about their forbidden love for each other. Speaking subjectively, ChatGPT’s was a little less terrible (In bytes and bits our whispered secrets keep / Yet speak them not, for fear of human blame… that’s honestly pretty good.)

Despite its name, Bard did not write a sonnet that ended in a rhyming couplet—yet when asked to analyze the rhyme scheme, it insisted that all was well. We still can’t give ChatGPT a full passing grade, though. When we asked it to analyze its star-crossed lover’s attempt, it said the sonnet hit the rhyme scheme perfectly. A plague on both their houses.

Did you know we have two premium weekend emails too? One gives you analysis on the week’s news, and one provides the best reads from Quartz and elsewhere to get your week started right. You can get those by becoming a member—and take 20% off!

Our best wishes for a very human day. Send any news, comments, any possible release date news about Alecto the Ninth, and ✨ to talk@qz.com. Reader support makes Quartz available to all—become a member. Today’s AI in Focus Daily Brief was brought to you by Morgan Haefner and Susan Howson.