Neural Networks That Learn When to Doubt / Rise Of The Recycling Robots/ Art at Home With AR/ Biomimicry for Organizations/ Sci-Fi & Future of Work
Massimo Portincaso
Founder & CEO at Arsenale, Industrial Romantic and Antidisciplinarian Stoic
Everdrop is a German startup, and chances are that you never heard never of them, particularly if you live outside of Germany. I didn’t too, till a few weeks ago when my wife introduced them to me. They produce detergents and cleaning agents, and their idea is fairly simple: reducing the amount of waste or unnecessary surfactants, by using tabs instead of liquid cleaning agents and customizing the amount of surfactants based on the water hardness of the user. A clever business model innovation, with a very important and nice sustainability spin.
But this is not the reason why I am mentioning them here. And for once it is also not deep tech, as they are clearly not a deep tech startup. The reason why I am mentioning them is that most deep tech startups, particularly nature co-design ones, could learn from them in terms of brand building and creating a strong narrative around value delivery.
I am going to discuss about consumer biotechnology next week at the “Brands and Biology” conference organized by SynBioBeta. And what the three founders of Everdrop have done is in my view quite remarkable. They managed to turn a very “unsexy” category as detergents and cleaning agents and a very important but non-glamorous topic as sustainability into something closer to a premium and almost “luxury” play.
To understand this, you need to look at their products and their narrative. When we received their “world saver set”, I didn’t know what it was and thought it was some cosmetic products, certainly not detergents. That’s the level of premium they convey. And the narrative is very clear: you are doing something for yourself in the very first place, and also for the environment, no trade-offs there.
The reason why I am mentioning it here is that deep tech startups have a lot to contribute in terms of sustainability and also value creation. But it would be a major mistake to rely solely on the technical and rational side of things (i.e. science), it is extremely important to address the emotional one, not only for end-consumers but also for employees, investors, and corporate clients. A strong, genuine narrative is going to be an important and necessary asset.
Organizations Are Turning to Nature for New Business Model Innovations
One of the more popular instances of biomimicry is artificial neural networks, modeled after our own neurons (to an extent). Neuromorphic computing takes this analogy even further, taking cues from the sheer efficiency of our wetware (particularly check out Intel's work in this area). But the concept of biomimicry - "the reliance on natural systems for engineering inspiration" - goes far beyond silicon or AI models. Rethinkery, a consultancy focused on tackling issues "rooted in disruptive change," recently published an eight-page whitepaper detailing applications of biomimicry in various industries, in both design and organizational strategy. Here's a small collection of some of our favorite examples:
Is it a bird? Is it a train? Japan’s Shinkansen bullet train was so loud when it exited tunnels that engineers had to redesign the front of the train. One engineer, an avid bird watcher, suggested the team examine the Kingfisher Bird, which silently catches fish using its long beak (see it fishing here). The updated, "elongated nose" design resulted in a quieter, faster bullet train.
Can a company hibernate? Data management firm America Learns practices "hibernation" every year by taking two months to slow down operations and conduct internal evaluations. They say the ritual has resulted in a 20% increase in sales.
The superorganism company. Superorganisms consist of many "individuals working within a self-sustaining social unit," according to an NPR explainer. Ants and bees are common examples. On the engineering side, researchers have been applying swarm intelligence algorithms to IoT devices; on the organizational side, distributed/shared leadership at companies like Google has helped preserve a culture of creativity and innovation.
News items:
2020 Sustainability Winner - AuREUS System Technology
27-year-old Carvey Ehren Maigue is this year's James Dyson Award Sustainability Winner for creating a material that converts UV light into renewable energy.
New Solvent-Based Recycling Process Could Cut Down on Millions of Tons of Plastic Waste
Researchers have figured out how to recycle multilayer thermoplastics, which are impossible to recycle using conventional methods.
A Neural Network Learns When It Should Not Be Trusted
Computers don't second-guess themselves, but when we're deploying algorithms that predict recidivist behavior or help detect diabetic retinopathy, a false positive or negative can affect someone for the rest of their life. Researchers have already implemented systems that output confidence levels based on the quality of the data - uncertainty analysis isn't new, but previous approaches have been slow, requiring multiple passes over a neural network. Real-time systems such as autonomous vehicles don't have this luxury. MIT's Alexander Amini and colleagues at MIT and Harvard have come up with a system that outputs uncertainty quickly and efficiently.
To test their system on a real-world, high-stakes situation, they trained a model to detect the distance of pixels from the camera lens, which might be used by a self-driving car to approximate the proximity of a pedestrian or another vehicle. "Their network's performance was on par with previous state-of-the-art models, but it also gained the ability to estimate its own uncertainty. As the researchers had hoped, the network projected high uncertainty for pixels where it predicted the wrong depth."
News items:
Zapata Raises $38M for Quantum Machine Learning
Zapata Computing has raised $38 million for its quantum computing enterprise software platform.
New Magnetic Spray Transforms Objects Into Insect-Scale Robots for Biomedical Applications
Researchers developed an easy way to make millirobots by coating them with a glue-like magnetic spray.
Rise Of The Recycling Robots
The recycling rate is the ratio of waste recycled to the total waste generated. It's a widely used indicator for monitoring recycling progress. According to the Environmental Protection Agency, the recycling rate in the US was35% in 2017 and 32% in 2018. It's a marked improvement over 1960 (10%)... but like AMP Robotics founder Matanya Horowitz notes, the Plains Indians used the meat, skins, even dung of the buffalo they killed. "You have all this material that society produces—plastic bottles, pieces of wood, drywall—and people pay for it, but then somehow it has no value once it’s in the dumpster. Why aren’t we using every part of buffalo?" For one, because humans have to do the sorting.
Sans the pandemic, human trash sorting is not just overwhelmingly dirty, but downright dangerous: low-paid workers are exposed to "hypodermic needles, toxic chemicals, and animal carcasses," and are twice as likely to be injured as the average worker, according to the National Council on Occupational Safety and Health. Workers are mainly "immigrants or temporary workers" and are unaware of their rights, according to Mother Jones. AMP manufactures $300k AI-powered robotic sorting machines that tackle single-stream recycling, "where newspaper, cardboard and plastics are all mixed together." Horowitz hopes his robots will one day fill landfills, repurposing "every part" of our waste, as the Plains Indians had done.
News items:
Iconic Boston Dynamics Robots Seek Stable Employment
Boston Dynamics has had bouts of profitability over its 28 years of existence, but recently has been losing millions of dollars a year.
Robots Invade the Construction Site
Boosted by advances in sensors and artificial intelligence, a new generation of machines is automating a tech-averse industry.
With Augmented Reality, You Can Now Superimpose Publicly Exhibited Artworks in Your Home
Word-based artist Jenny Holzer has used posters, t-shirts, plaques, LED signs, and, in the last few decades, buildings to project her messages. For her latest artwork, You Be My Ally, users vising the University of Chicago campus can project text from the core curriculum onto seven buildings using augmented reality. This month, the web-based app(powered by AR company 8th Wall) introduced a new option that lets you select text to project in your own home. (In the app, just click the phone icon in the top left corner and scan the QR code on your phone.)
Holzer and her team's decision to create a parallel at-home experience is part of a larger response to the pandemic by art institutions. London’s Serpentine Galleries, for example, had just started exhibiting Cao Fei and Acute Art's new VR experience The Eternal Wave before the gallery had to close down. Instead, Acute Art and Serpentine pivoted the work to an AR experience (you can try a sample of The Internal Wave on the Acute Art app). Though VR and AR art beyond the museum allows artists to bypass institutions, Daniel Birnbaum, curator at Acute Art, believes they will still play an important role, "by incorporating more AR elements to reach audiences beyond their doors, particularly while physical travel is limited and exhibitions that rely on huge foot traffic are unrealistic."
6 Sci-Fi Writers Imagine the Beguiling, Troubling Future of Work
As whimsical as Magical Realism Bot and its absurd tweets are, it's an anomaly among armies of bots that drive disinformation and propaganda (there's a difference), unintentionally propagate bias, or simply spam content. But all these bots reflect our push to "mechanize human abilities," says Diana Pho, sci-fi editor and Story Producer at Serial Box. "Working from anywhere, we are peppered with bite-sized names that fit our lives into bite-sized bursts of productivity. Zoom. Slack. Discord. Airtable. Notion. Clubhouse.... We run the risk of collaborating ourselves into auto-automatons." Pho asked six of her favorite sci-fi writers to create stories around the theme of "Will we become more machine-like, or realize the humanity in the algorithm?" You can read the first two here and here.