Early attempts to create dedicated hardware for artificial intelligence smarts have been criticized, as well as, a bit rubbish. But here’s an AI gadget-in-the-making that’s all about trash, literally: Finnish startup Binit is applying large language models (LLMs) image processing capabilities to track household trash.
AI for sorting the stuff we throw away to increase recycling efficiency at the municipal or commercial level has captured the attention of entrepreneurs for some time (see startups like GrapeRot, Trashbot, Glacier). But Binit’s founder, Borut Gargic, believes tracking household trash is untapped territory.
“We’re building the first household waste tracker,” he tells TechCrunch, likening the upcoming AI gadgetry to a sleep tracker but for your littering habits. “It is a camera vision technology supported by a neural network. So we are tapping LLM for identification of regular household waste.”
The early-stage startup, which was founded during the pandemic and has pulled in nearly $3M in funding from an angel investor, is building AI hardware that’s designed to live (and look good) in the kitchen — where a cabinet or wall is near the sink. is mounted on. – Relevant action is taken. The battery-powered gadget has an on-board camera and other sensors so it can wake up when someone is nearby, allowing it to scan items before tossing them in the trash.
Gargic says they rely on integrating with commercial LLMs — primarily OpenAI’s GPT — to perform image recognition. Binit then tracks what the household is throwing away — providing analytics, feedback and gamification through the app, such as a weekly waste score, aimed at encouraging users to reduce how much they toss out.
The team originally tried to train their own AI model to identify trash but the accuracy was low (around 40%). So they hit on the idea of using OpenAI’s image recognition capabilities. Grgic claims they are achieving about 98% accurate trash identification after integrating LLM.
Binit’s founder says he has “no idea” why it’s doing so well. It’s not clear whether so many images of trash were in OpenAI’s training data or whether it was able to recognize so much content as it was trained. The results achieved in testing with OpenAI’s model may be down to items scanned as “normal items”.
“It’s also able to tell with relative accuracy whether a coffee cup has a lining or not, because it recognizes the brand,” he continues, “so basically, what we have to do is for the user to pass the object in front of the camera. So it forces them to freeze it in front of the camera for a while while the camera is capturing the image from all angles.
Data on trash scanned by users is uploaded to the cloud where Binit is able to analyze it and generate feedback for users. Basic Analytics will be free but it intends to introduce premium features through subscription.
The startup is also positioning itself to be a data provider on the things people are throwing away — which could be valuable intel for organizations like packaging entities, assuming it can scale consumption.
However, an obvious criticism is that do people really need a high-tech gadget to tell them they’re throwing away too much plastic? Don’t we all know what we eat – and we need to try not to produce so much waste?
“It’s habits,” he argues. “I think we’re aware of it — but we don’t necessarily act on it.
“We also know that sleeping is probably good, but then I put on a sleep tracker and I slept a lot more, even though it wasn’t teaching me. anything Which I didn’t already know.”
Binnit also says that during tests in the US, there has been a reduction of about 40% in mixed bin waste as users engage with the transparency of the bins the product provides. So it believes its transparency and gamification approach can help people change internal habits.
Binit wants the app to be a place where users can find both analytics and information to help them reduce how much they throw away. For the latter, Grgic says they also plan to tap LLM for recommendations — factoring in a user’s location to personalize recommendations.
“The way it works is – for example, let’s take packaging – so the user scans each piece of packaging that your app has a little card designed and on that card it says this is what you threw away ( eg plastic bottles )… and these are options in your area that you can consider to reduce your plastic intake,” he explains.
He also sees scope for partnerships, such as with influencers to reduce food waste.
Another innovation of the product, Grgic argues, is that it is “anti-unhinged consumption,” as he puts it. The startup is aligning itself with the growing awareness and sustainability action. To protect the environment for future generations, our throwaway culture of single-use consumption needs to be overcome, and replaced with more mindful consumption, reuse and recycling.
“I think we’re on the cusp of (something),” he suggests. “I think people are starting to ask themselves questions: Is it really necessary to throw everything away? Or can we start thinking about repair (and reuse)?
However, Binit’s use-case can’t just be a smartphone app? Gargick argues that this depends. He says some households are happy to use a smartphone in the kitchen while they’re getting their hands dirty during meal prep, for example, but others see the value of having a dedicated hands-free trash scanner.
It is worth noting that they also plan to offer scanning facility for free through their app so they are going to offer both options.
So far the startup is in five US cities (NYC; Austin, Texas; San Francisco; Oakland and Miami) and four in Europe (Paris, Helsinki, Lisbon and Ljubljana, Slovakia, where Grgic is from).
He says he’s working on a commercial launch this fall — possibly in the US. The price-point they’re targeting for the AI hardware is around $199, which it describes as the “sweet spot” for smart home devices.