Early efforts to create hardware specifically for artificial intelligence smart homes have been criticized as wasteful. But here’s an AI gadget-in-the-making that’s all about waste, literally: Finnish startup Binit is applying large language models’ (LLMs) image processing capabilities to track household waste.
AI for sorting discarded items to increase recycling efficiency at the municipal or commercial level has received the attention of entrepreneurs for a while now (see startups like Greyparrot, TrashBot, Glacier). But Binit’s founder, Borut Grgic, considers household waste tracking to be an area where it doesn’t work.
“We’re producing the first household waste tracker,” TechCrunch said, likening the upcoming AI gadget to a sleep tracker but for littering habits. “It’s a vision camera technology powered by a neural network. So we tapped LLM to identify ordinary household waste objects.
The early-stage startup, which was founded during the pandemic and received almost $3M in funding from angel investors, is building AI hardware designed to live (and look cool) in the kitchen – mounted on a cabinet or wall near the trash can. actions related to doing so. Battery-powered gadgets already have cameras and other sensors on board to wake up when someone is nearby, allowing them to scan items before putting them in the trash.
Grgic said that he relies on integration with commercial LLMs – especially GPT OpenAI – to perform image recognition. Binit then tracks what the household throws away – providing analytics, feedback, and gamification through the app, such as a weekly waste score, all aimed at encouraging users to reduce the amount they throw away.
The team initially tried to train their own AI model to do garbage recognition but the accuracy was low (about 40%). So, they came up with the idea of ​​using OpenAI’s image recognition capabilities. Grgic claims he achieved almost 98% accurate garbage recognition after combining his LLM.
Binit’s founder says he “doesn’t understand” why it works so well. It’s not clear if there are a lot of junk images in the OpenAI training data or if it’s just able to recognize a lot of stuff due to the large amount of training data. It has been achieved in tests with the OpenAI model that the item can be scanned as a “public object”.
“You can even say, with relative accuracy, whether the coffee cup has a layer, because it recognizes the brand,” he said, adding: “So, what the user has to do is pass the object in front of the camera. So it forces them to stabilize in front of the camera for a little while the camera takes pictures from all angles.
The trash data that the user scans will be uploaded to the cloud where Binit can analyze it and generate feedback for the user. Basic analytics will be free but we plan to introduce premium features through subscription.
The startup is also positioning itself to be a provider of data on things thrown away – which could be valuable intel for entities like packaging entities, assuming they can leverage it at scale.
Still, the obvious criticism is do people really need a high-tech gadget to tell them they’re throwing away plastic? Don’t we all know what we consume – and we should try not to produce waste?
“It’s a habit,” he argued. “I think I understand – but I don’t have to.
“We also knew it was possible to sleep, but then I put on a sleep tracker and I slept more, even though it didn’t teach me. whatever that I don’t know yet.”
During tests in the US, Binit also said that there was a reduction of about 40% in mixed waste because users participated in the waste transparency provided by the product. So, a transparent approach and gamification can help people change their existing habits.
Binit wants the app to be a place where users get analytics and information to help reduce the amount they spend. For the last Grgic said he also plans to tap LLMs for advice – factoring in the user’s location to make personalized recommendations.
“The way it works is – let’s take the packaging, for example – so that every piece of packaging the user scans there is a small card that is formed in your application and on the card it says this that has been thrown away (for example, plastic bottles. )… and in your area this is an alternative that you can think of to reduce plastic consumption,” he said.
He also looked at the scope of the partnership, such as the impact of reducing food waste.
Grgic argues that another new product is “anti-unhinged consumption”, as he puts it. The start is in line with growing sustainability awareness and action. The feeling that the culture of single consumption must be discarded, and replaced with more conscious consumption, reuse and recycling, to preserve the environment for future generations.
“I feel like we’re on the cusp (of something),” he suggested. “I think people are starting to ask themselves: Is it really necessary to throw everything away? Or can we start thinking about repairing (and reusing)?”
Isn’t Binit’s only use case to be a smartphone app? Grgic says it depends. He said some households like to use their smartphones in the kitchen when their hands get dirty while preparing a meal, for example, but others see the value of having a dedicated hands-free waste scanner.
It is worth noting that they also plan to offer the scanning feature through the app for free, so they will offer both options.
So far, the startup has tested the AI ​​waste scanner in five cities in the US (NYC; Austin, Texas; San Francisco; Oakland and Miami) and four in Europe (Paris, Helsniki, Lisbon and Ljubjlana, in Slovakia, where Grgic is from. of).
They say they’re working on a commercial launch this fall — likely in the US. The targeted price point for the AI ​​hardware is around $199, which is described as the “sweet spot” for smart home devices.
