It's 2025, and companies still don't know what AI is good for
I got this idea from this year's CES, which features AI-powered kitchen appliances, Baby cribsAnd other products that didn't really call for AI.
Behold: Spicerr, an “intelligent” touchscreen-equipped spice dispenser that learns your tastes to recommend unique recipes as you cook.
Spicerr's usefulness is a bit questionable to begin with. Spicerr doesn't grind, and it takes $15-$20 proprietary capsules that aren't refillable. All that aside, are people really itching to start with the recommended salt and pepper shaker?
Elsewhere at the show, Drew had the Chefmaker 2, an AI-powered air fryer. Yes, you read that right — the AI-powered air fryer.
The concept isn't as outlandish as Spicer's, mind you. ChefMaker 2 can extract recipes from cookbooks with a page-scanning feature and even handle the complex math of calculating cooking times and temperatures.
But is cookbook scanning really a feature that the air-fryer-buying public demands? Speaking as a member myself, I can't say this has ever happened to me – and it seems to True for most people.
Notably, CES had more weird AI products in store.
Razer's Project Ava, inexplicably named after the killer robot from the 2014 movie “Ex Machina,” is an “AI gaming copilot,” as the company describes it. Ava essentially plays games for you instead of playing games for you. With permission, Ava captures still images of your computer screen, then provides pointers (such as “dodge while rotating blades”).
As Sean Hollister of The Verge writesAva disputes that it was clearly trained in gaming guides, yet does not credit the authors. It's also confusing. At least in its current form, Ava has a delay of several seconds, and it interrupts the game's audio to give instructions.
I have to ask again: Who was screaming for this, exactly? Who is going to use it regularly, much less pay for it?
As far as I can tell, AI products outside of CES are a symptom of the industry's rampant hype. AI companies raised $97 billion in the US alone last year, enough for 42 buyouts the sphere. What's stuck is that vendors are throwing AI spaghetti at the wall, because there's little downside to doing so — and huge potential upside.
In many cases, they are also running up against the limitations of AI as we know it. A daunting challenge for industry is figuring out which areas of AI are technically feasible to use. Often, this leads to over-promising – under-delivering. ChatGPT still gets some errors. Image generators are historically inaccurate. And characters in AI videos merge into each other's bodies.
So we're stuck with IRL AI slop: air fryers, spice dispensers, and “AI gaming copilots.” They are not what most of us want, but what is possible today with relatively little R&D lift.
Here's to a good next year.