GoKawiilTests to determine whether Google’s agentic AI system Gemini Spark can deliver on the promises made on stage at last month’s I/O event shows that, for the most part, it can.
Since this is the model Apple will be using to power the new Siri, that’s equally good news for the Cupertino company and its customers …
There was a huge argument a little over a year ago when Apple commentator John Gruber launched a blistering attack on the iPhone maker’s failure to deliver on its new Siri promises. He said the company had done nothing more than show concept videos of Apple Intelligence features they couldn’t actually demonstrate, even in carefully-controlled conditions.
When Google introduced its agentic AI Gemini Spark, the company performed live demos on stage. That’s a massive step forward from a video simulation, but there’s still a sizeable gulf between a carefully-planned demo and real-life usage.
The Verge’s Jay Peters decided to try the demonstrated features for himself, turning them into real-life tasks on his own data, starting with this one.
I asked Gemini to draft an email to my wife that compiles our total monthly average grocery spending in 2026. I figured this test would tell me a few things: Could Spark figure out who my wife was (without me giving Spark her name), could it determine where our budget spreadsheet is in Drive (which does not have “budget” in the file name), and could it actually draft an email in Gmail?
People sometimes use the phrase “scarily good” in a colloquial way, but in this case, I think it applies rather literally.
When I got the result from Spark shortly after, I really said: “Wow, that’s actually nuts.” Spark found my wife’s email address, pulled the right information from our 2026 budget spreadsheet, grabbed the monthly grocery totals including the incomplete data from May (which still wasn’t over when I ran the test), averaged the totals, and put it all in a draft email in my Gmail. The text of the email addressed my wife by her first name, even though her email address does not contain her first name. It even included a sign-off that we use just for each other.
It didn’t fully deliver on everything demonstrated, but he said that he was “floored by the results, though they were imperfect.” The full piece discussing the other examples is definitely worth reading.
What Google demonstrated – and Peters found to mostly work in real life – was exactly the type of features Apple showed off in its concept video. The new Siri may be taking an extremely long time to materialize, but this experience does suggest that it really will live up to Apple’s promises, even if those promises are actually fulfilled by Google.