I’ve been predicting the future of Hardware for about 10 years now which started with:
Predictions that came true
- Research in robotics was going to be accelerated. This was based on the fact that in India education in robotics, though a profitable business, was teaching the basics. There was an eventual slowdown in the number of companies being started in education because students in electronics and mechanical were their target customers. This meant with the efforts I was putting in to bring it to a wider audience including computer science students through JMoon Makerspace, and showcasing how they all worked together, the impact was going to create opportunities for people to go into research and a similar trend was emerging worldwide. Canada is at this interesting intersection where both education and robot applications can become better. So in some respects it is behind India, but in others, because of its proximity with the US market, it’s ahead in deployment and potential use cases.
- When vacuum robots were launched I immediately saw the rise of mopping robots because the majority of the world doesn’t use vacuums and instead uses mops to clean their houses and even schools and corporate buildings. Even though vacuum cleaners work on marble and tile flooring, in Indian houses and commercial buildings people are used to using mops with disinfectants mixed in which means they will never be happy with a simple vacuum no matter how well it performs because they are so used to a perception of cleanliness that vacuums alone do not provide. I tried to have my team work on this in 2016 but at the time there were other important problems to focus on in the 4 startups I was running.
- For a wider social robot adoption in most households, there will be a step that goes from your Amazon Echo (Alexa) device to full-fledged butler robots. This step consists of an “Alexa on wheels”. This robot needs to have conversational abilities, a camera for recognition, navigation, and obstacle avoidance, as well as wheels for mobility. It can optionally have storage capabilities which has been achieved by Astro having space for holding a cup or small objects.
However, there is a pricing issue that needs to be resolved. A robot that’s smaller in size, limited in features and better priced might be a better fit to increase mass adoption. Think of this as an enhanced bigger version of Cosmo (which lacks the conversational aspect). While most stationary social robots had impressive features and capabilities, public perception of robots still dictates that they be mobile in some way (wheels, legs or robot arms) which is why these companies did not scale. Their assumption that people would pay for a stationary social companion did not scale outside of early adopters who love the tech.
Creating a similar robot but targeting it as an educational tool or toy is also a limiting approach because, like Echo, Miko can serve as a device for everyone in the house. It can play Chess thanks to Bhavya and Atur from SquareOff (side note: I saw the prototype of their chess board in action before it was even fully developed in Mumbai) so it can certainly do more.
What’s better, a robot educational toy that kids will grow out of using after a few years never to be bought again, OR a robot with conversational LLMs that becomes part of family life so it is integrated into daily activities of everyone to the point that people will need to replace it if damaged or obsolete?! Why couldn’t it also play chess with your retired grandparent or help keep them entertained? The answer is it could, but due to its marketing targeting children and parents of young children, it is possibly missing out on a completely relevant audience.
Following are my new predictions (new for the public but I’ve seen them coming for a long time) and reasons.
Predictions for the Future
1. Retrofitting Household Devices that allow people to automate their existing homes is a better solution than creating devices that are made to be automatic.
If you are creating an auto-cleaning dishwasher for people who already have dishwashers with a 10-year life cycle, you will only sell to them when they need a new dishwasher for a new home or are replacing a broken one. So make sure you keep your sales cycles in mind as they may not be seasonally dependent. In this case, your sales will also be dependent on the housing market and new construction, people’s ability to spend more (so a recession might be a great time to stay lean), and the utility of the product in their lives (can they live with a broken dishwasher for a while without needing to replace it or not). Keep these factors in mind when making predictions on your annual sales and units to manufacture.
Automating your curtains or windows from scratch requires considerable investment that many people think about only on 2 occasions: when they are building a house from scratch, or renovating a big portion of the house, which happens when they have surplus capital to spend, or its medically necessary for them (which is the market that stairlifts, compact 1-person inhouse elevators, and accessibility modifications for vehicles depend on).
I invested in Ryse because it’s an addon that lets anyone automate their window shades at any time of the year and is something that can serve as a holiday gift. People will buy this product as soon as they have a problem. They will know whether they want it as soon as they see it on social media or in the store and don’t have to wait to allocate a budget for the perfect time to buy it.
If you are creating a new device that replaces people’s existing one instead of retrofitting, make sure you get into physical stores or Amazon quickly where people go to buy those products and focus on improving your SEO, instead of relying on selling online through your own website or spending massive marketing budgets on social media.
2. Medical. Now medical robots have been around for quite a while. First, it was surgery, and then during the pandemic robots that carry stuff or help keep environments sterile were huge only during that period. But there are parts of the medical field that I have not yet seen roboticists focus on which are equally massive opportunities. The one that I came across from a random conversation in Nov 2018 was, specifically in India, for robots to perform autopsies. I am so excited by this that I even have a name for a robot that would do this if I ever decide to build it myself. It’s a morbid field but here’s why it makes sense and should be focused on.
The problem is such: In India, autopsies at night are NOT performed by doctors. Yes. It is true. The non-medical staff may actually be the one that does the dissection before the doctor finally drops by to look around and give their results. In government hospitals, it is usually the janitor who has to get drunk and high in the middle of the night to be able to perform a dissection of a human corpse. He gets high because he is not mentally trained for it, but does anything for money as janitors in India live pretty close to the poverty line and have maybe a middle-school education. Now most people do not know this as it happens behind closed doors, but if they did, imagine their reaction to their loved one’s body being handled in such a way. Robots can fill this role of dissection quite nicely. How? I will leave that up to you to decide instead of sharing the details of the idea I have and corrupting your approach, but it is certainly a market that I have yet to see anyone target in the last 5 years that I have been sitting on the idea. This problem could be an issue in other developing countries too, though I admit I haven’t done enough global market research on it to know for sure.
There are parts of the medical field that roboticists have a cursory understanding of from watching TV shows but to get a deeper understanding we need to have more medical teams involved in the development phase. I am confident that even more use cases will arise from this that cobots and machines will be able to fill.
3. If the “Alexa on wheels” household butler robot starts to get adopted at a fast rate, adoption rates of robots like Gita are not far behind which have additional features including, storage, arms, vacuuming / mopping, that allow them to do multiple chores WHILE retaining the conversational features. Depending on how a company approaches its design, marketing, and pricing, and uses AI LLMS for conversations there might be a possibility of skipping the Alexa-on-wheels phase altogether and making it deployable quicker than ever before.
Starting in 2008, BEFORE I got into robotics, I created a series of 9 robot designs called JJ (Junior Jasmeet, egotistically named after me): JJ0, JJ0.5, JJ1, JJ1.5, JJ2, JJ3, JJ4, JJ5, JJ6. Each of these designs is meant to be similar to a butler robot for households and improves in looks or features. Some of these designs could not exist at the time because of unavailable technology (Kinect was barely a thing at this point), but I think it is safe to say that in the next 5 to 10 years, it is possible to get to a point where JJ5 can be achieved and sold in houses. Early version of JJ0 is the one that most people can see on Robot JJ‘s website. If I had to restart the project now in 2023 I would skip a few designs and jump right into JJ2 and include all features of its previous versions as best as possible.
A feature that is currently missing but with the help of AI has the potential of being implemented in these robots is one that I was working on pre-pandemic for Robot JJ to have that I don’t think even Astro has. I like to call that feature Object Remembrance. Now that the robot is capable of moving around your house, it can recognize and differentiate between you and the objects it sees, can successfully navigate around them, and talk to you. It can remember you with facial recognition. The next step is its ability to remember where it saw an object last. You will not lose your wallet ever again, you won’t need to go searching for your child’s bag in the morning when they are running late for the school bus, and you’ll know which room your child is in. You ask where the object or person is and the robot will be able to tell you exactly where it last saw it, and might even be able to go retrieve it if it is accessible within its range. It is based on human understanding of object permanence. This is the one aspect of robotics that can help us achieve the elusive Artificial General Intelligence (AGI) or its closest approximation by allowing the robot to explore and learn from its environment. Talk to me if you are working on it or achieve this, but I think it is only a matter of time before this becomes common in all house robots.
A feature of JJ series is that it does not have a single design with touch screen. Displays can work but having to touch a robot to interact with it outside of airports, is not a relevant feature. If its in your home, it needs to be without a screen and interact in a way you would with a person.
Another feature that Robot JJ was supposed to have would mean that you would never need another phone app to control your smart devices. The robot itself would be the central hub. With the conversation feature in place you would be able to tell the robot to control those devices in every room and it would connect to all your IoT devices and smart switches to control everything automatically. Now you’re probably starting to see the scale of the problem I was trying to solve with RobotJJ and what the future could potentially hold. If not, more details on this are in the section below. This is the reason I took a back seat from a lot of media feature opportunities. Every waking moment was spent either planning or developing this with no time for anything that was not directly or indirectly involved.
JJ6 is a totally different breed and parts of it require MagLev technology to be more powerful, easily deployable, and compact and so i’ll say it might take another 12 -15 more years to be commercially viable. If you’re in these fields, I absolutely want to talk to you and need to do more research into it myself. I tried a maglev project in 2007 and it wasn’t as impressive as I had hoped but has been on my mind since. Once magnetic levitation is close to that point I’ll tell you more about the predictions of how it will impact the hardware field and the development of robots.
4. Smart Household Devices. Houses have other chores besides the kitchen. Go out there and ask your potential customers what they want.
Here’s a few hints to get you started: In a survey I did before I started working on Robot JJ, the biggest task women wanted help within the house was not taking out the trash, or cooking, or washing dishes/ clothes, but instead with sorting their closet space. In other words, other tasks are good to solve, but women are more likely to consider buying a robot that cleans their closet. Housewives especially because they spend all day taking care of everything else, but the biggest task that they procrastinate on is their own closet as it is not a problem for the collective household and hence lower priority but very desirable as a feature for a robot to have so you could potentially sell the robot to houses by highlighting this one feature. Single women want the feature of security monitoring and alerting system in their robots.
Single men or people who are bad at cooking are your main target market if you are thinking of selling robots that make food from ingredients. Robots that assist in automating cooking, by making parts of a meal, like bread, you’ll probably sell to people who are good at cooking already IF it matches or exceeds the quality of the way they do it. You want to focus on giving people more time with their families. In some families, cooking can be a joint activity, in which case, you focus on solving some other problem to make sure that they do have that time to cook together. But your robot needs to be priced according to the perceived value of this family time to your customer. If you’ve built a $1400 appliance that lasts at least 5 years and gives them 700+ hours of time in return per year, is this worth it to them? Do they see the value in it? Is your marketing done right to convey that value? Does it fit easily in most small kitchens?
IoT devices were never meant to be run with smartphones, or created so you could monitor data through the cloud. That’s not the future or right focus for those devices. Why would I want to constantly look at data graphs all day? Remember the Data→ Information→ Knowledge pyramid? You’re aiming for knowledge, not data gathering. Using AI, IoT devices are meant to remember your patterns and habits. The aim of the smart device isn’t for you to be able to turn on your coffee maker from your phone while in bed in the morning or monitor your daily energy spending for that machine. It’s for you to never have to turn it on or off after 1 week of learning. With Robot JJ this was the problem I wanted to solve. The device needs to learn that you wake up at 8am in the morning and drink your coffee by 8:15am, and after learning this it should turn the machine on and off automatically. If it detects changes in your patterns, it needs to adapt accordingly. Started living with a partner who drinks their coffee at 6:30am before you even wake up? It needs to know that the coffee needs to be made by 6:30 and again by 8:15. Many times people have to provide this data and set a fixed time on their smartphones, but schedules can change, there can be anomalies and outliers in the data, like you waking up earlier than usual one day, and smart devices need to adapt with them. Your coffee machine should ask you if you want coffee that morning at 6am if you are in the kitchen earlier than expected, or alert your phone if you are outside the house that it will make coffee if you don’t stop it by 8:10am because you left the house early that morning and are getting your coffee on the way. Smartphone apps for these devices need to exist as a failsafe emergency switch rather than a remote control.
Now it would be hard for each smart device to constantly monitor that you’re in the kitchen, right? So in Robot JJ’s case, the central control would not even be the smartphone, but rather the robot which would monitor your movements and follow you around, as well as control all the smart devices. I’m sure by now you’re seeing the magnitude of the problem I wanted to tackle was multi-fold.
5. Eldercare. My prediction for robots that provide social interactions in elder care involving robots will not work out as well as most people think. Specifically in the case where we try and build robots that interact with elderly patients. Most investors might think it is a growing market to be in, and roboticists might think it is a worthy application to build to care for the elderly, but the honest answer is that humans are not wired that way.
Robots or hardware devices that can detect and prevent problems before they occur, such as people falling, and alerting authorities are wonderful applications in eldercare. Want to build a device that prevents people from getting fatal and preventable injuries, good! Want to build something that predicts heart attacks or seizures minutes before they happen and alerts ambulances, great! Want to have a smart mirror that will use cameras to detect whether or not you have visible symptoms of an affliction before it manifests, or monitor the progression of it, wonderful! Want to build an autonomous bed that wakes you up, puts you in an upright position, and maybe even dresses you without you getting out of bed, AMAZING!
But a robot that provides the level of care and social interaction that elders need, will not work. Or rather is unnecessary if all it is doing is providing interaction. Is it providing them with medication on time? Perfect! Focus on that problem which may not need a robot! Solving a problem that exists for the staff providing the care? Excellent!
As humans, we need social interaction. Not a robot that interacts with humans or one they can “cuddle”. People can survive alone, but a lack of social interaction causes loneliness. Why does an aging person need a robot pet to cuddle with when they can have an actual pet that the care home staff can bring in? Robot cuddle pets would make sense if we were suddenly seeing the extinction of cats and dogs. It is solving the wrong problem. If you are building in this field focus on freeing up time and effort of the staff taking care of the elderly so they can be more present, more involved, have to do less paperwork, have to spend less time in the other parts of their activities which do not involve interacting with or taking care of the elderly. Create things that help them deliver that care more safely and efficiently. Think robots can be used to take care of elders at home? Do the same thing, free up time for the people in that house to take care of and spend time with their elderly family members.
Remember the 6D’s of Robotics I shared before:
Robotics specifically is a field that aims to be prevalent in fields where the tasks fall under one of the 6D’s.
Dirty: Robots used for dirty or unclean tasks
Dull: Robots used for repetitive or monotonous tasks
Dangerous: Robots for hazardous tasks that pose a risk to human safety
Domestic: Robots used for tasks in the home
Dextrous: Robots used for fine motor skills or precision
Dear: Robots that perform otherwise expensive or highly skilled tasks
It applies to almost all applications of hardware. So use these as your guide to solve problems that don’t take away from human interaction, but instead amplify and improve it by taking care of such tasks and freeing up time.
6. Productivity. Speaking of freeing up time, this is a big sector where all XR, 3D printing, Industrial IoT, come together and can change the game in homes, commutes, and at work.
AR devices that help you interact on the move without needing to look at your phone, XR devices that allow you to explore and be productive remotely as part of the metaverse or remote control robots and physical world through haptic feedback, 3D printing that speeds up the time for making houses or manufacturing products, Industrial IoT that automates and tracks production without human intervention, are all together going to impact this part of human life. As we automate more and more of the 6D tasks around us, humanity will have more free time to spend on things that matter: art, social interaction, travel, caregiving, and fixing other worldly problems around us.
If you know Maslow’s hierarchy of needs, the bottom half of psychological and safety are the things that robots and smart machines can help us achieve at a global scale, freeing up time for belonging and love as a collective. This is another reason that I am optimistic about automation. While people mostly worry if robotics and AI will take up jobs, and the truth is that they will, it will also free up time so we can focus on the upper half of that pyramid not as individuals but as a collective so it is in humanity’s best interest to have it scale.
This is also where applications in agriculture can be focused on because we need more food to sustain more people and less transport of food from one country to another to solve the global food crisis and the waste generated from it. Agtech is a field that has yet to see its full potential envisioned and mushroom harvesting is only the beginning. Along with productivity, innovation in food-growing technologies will be a major factor in the push toward automation. So if you love this sector keep an eye on not only how to pluck tomatoes from a field, but also how technology can assist in growing more tomatoes in a given sq ft space. That is an equally big problem to solve.
7. Hospitality is a major application for robots. Parts of the world have aging populations and wage issues that are causing these applications to rise, especially post the Covid pandemic. Room service delivery is a great application for this, but I am yet to see a “robot restaurant” do well simply because it has robots.
If restaurants lack servers, replace the entire server experience with robots, not have it work alongside a server. Cobots work well in industrial applications because there is specially trained staff in those environments who may be familiar with working with machinery or have slight technical expertise. In the restaurant industry, however, robots get in the way of the servers who are usually faster. Having a server walk slowly to the table in a chaotic restaurant during the lunch or dinner rush because the robot can’t get to the table as quickly is slowing down the server. It serves as a gimmick more than it serves a purpose. But if the whole order-taking, kitchen meal prep, delivery to table, extra assistance, and payment processing (without tips) experience was automated, at comparable speeds, with minimal supervision by 1 or 2 people, it would be a great well-oiled machine of a restaurant.
A cobot that makes a process safer but is slower is still perceived as worse by most people. If you are building something that makes things safer, aim to make it as fast as a human in its ability to execute that task otherwise you are facing an uphill battle in adoption.
You could make a robot that cleans dishes because it doesn’t get in the way of anyone else’s work in the kitchen or in the front-of-the-house staff. Robots that do the heavy lifting of boxes in the back before opening hours, interacting with the limited staff of restaurants only, and then going to sleep during operational hours are a better solution to the problem.
Robot bartenders are great, though you might not need big industrial robot arms for that. They don’t need supervision if done properly, they aren’t moving around among the crowd. Children don’t walk up to them and start kicking them.
The future is bright for robotics, if you are building things for solving problems, not simply for the desire to create cool things. But if you make a cool thing that doesn’t quite solve a purpose, share it with me, I still wanna see it.