Just last week, I had the opportunity to discuss the future of deep tech with AI legend Peter Norvig, Director of Research at Google and co-author of Artificial Intelligence: A Modern Approach amongst his other fifty over publications in various areas of computer science. His book is like the Bible for me and it’s something that I’d recommend everyone to get a hold of.
As I reflected upon 2020 and the challenges we faced as an organisation such as sales negotiation over Zoom, Peter highlighted a point which was a good reminder for all businesses.
“It’s easier to do sales if you’re asking for less.”
While simple and questionable at first for a business, he went on to recount an experience he had in the past selling enterprise software where the sales cycle was long but they shifted gears and adopted a revenue sharing model. Customers did not have to pay for the software until they were making revenue, then they would share it. It’s akin to Google’s strategy where anyone can buy ads starting at a low cost and if they reap some growth from their small investment, they can then increase their investment.
“Allowing people to get in and try things out easily is good both because it’s a shorter sales cycle and it also means you have more people you can connect with”.
Now back to our deep tech discussion.
Singapore the next Silicon Valley
In the field of software, there are a few companies that I noted have interesting business solutions. One such example is ViSenze that does visual search, serving the e-commerce and retail businesses. Their product is built upon deep learning and computer vision that is fully-automated and scalable.
Deep tech is ultimately based on research and there is much research work happening in this part of the world. And while Silicon Valley where Peter is at, is the heart of technology innovation, we in Singapore are poised to be the Silicon Valley of Asia thanks to the government’s effort in creating that ecosystem and being a firm supporter for innovation.
You don’t need to be in Silicon Valley to succeed
Peter chimed in and shared that despite being in the heart of Silicon Valley which has many advantages for a tech company, there are it’s disadvantages too such as the high cost of living and competition for hiring. And with his time with Google for Startups which runs programs around the world, he has seen many companies around the world doing things that would be hard to do coming out of Silicon Valley. Reasons being there will be a lack of understanding of the local marketplace and difficulty in reaching them. At the end of the day it’s about coming up with a product that makes sense for the users and someone sitting in Silicon Valley may not understand or won’t be interested in working on such a project for a marketplace outside of Silicon Valley.
“You can be successful from anywhere as long if you have the right drive, right idea and the ability to execute on it.”
5 deep tech predictions
From my discussion with Peter, I’d like to highlight five deep tech industries that we both feel will experience leaps in development in the coming years.
Passionate about the applications of artificial intelligence in robotics, Peter is a strong believer that this field of AI has much room for growth. Coming out from the pandemic, his predictions for the first few applications we can expect to see would be medical robotics and delivery applications given social distancing regulations. In an earlier deep dive into robotics, Peter highlighted that while we see a few applications of robotics now, these are mainly in controlled environments such as in warehouses. The most successful examples would be Amazon’s small cart size robots and flying drones that deliver packages to their customers. However, the challenge of robots to operate in the real world with many changes still persists.
“In the next few years, we’ll see specific applications such as delivery vehicles and worksite vehicles before they come to the general public.”
2. Computer Vision
The ability for computers to gain high-level understanding from digital images or videos, computer vision is a big part of what makes robotics possible. According to Peter, computer vision touches almost every single one of Google’s machine learning projects. Peter added that we are at a stage where we are looking for the right applications for these technologies and there are new uses everyday. Personally, I’m looking forward to the growth of quantum computing which would propel the capabilities of computer vision and many other innovations we see now. However, that development might only take place in the next five to 10 years.
A large frontier that is emerging, Peter highlighted the advancement in protein folding which presents the opportunity to design new types of drugs and new types of therapy. Another example is Singapore startup Biofourmis that’s making its waves in the industry with its recent US$100m series C investment led by Softbank. Biofourmis, a portfolio company of SGInnovate as well, leverages vast amounts of data to provide a clearer diagnosis of a real-time patient. Biotechnology is a large industry with potential for a variety of innovative applications such as these three that Northeastern University recently published.
4. Food Tech
With the global pandemic pulling the brakes on supply chains around the world, food sustainability has become a focus for many countries and especially so for land-scarce Singapore. To achieve our ‘30 by 30’ food production aim, Singapore needs to look to food tech and the Singapore government has rolled out multiple initiatives and support for startups venturing into this industry. Just last week, Singapore was called a food tech hub.
5. Natural Language Understanding (NLU)
This field isn’t new to many of us, we’ve got devices such as Google Home, Amazon Alexa, Apple Siri and even conversational chatbots like TAIGER’s Omnitive Converse that can understand the meaning and context of a user’s question, and not to mention slang such as Singlish. But where are we exactly with this technology and where is it headed towards? Peter feels that we are right about the stage where we are to see a big jump in capabilities for NLU.
“Right now you can talk to your devices and you can say ‘tell me the time’ or ‘tell me the weather’ but you start to run out of things you know are going to work when you have a conversation with your device. I think it will really be transformative if we can get to the point where you feel like no matter what I ask my device, it’s going to come through and going to have a good answer and going to be as good as a human assistant.”
No matter the prediction, the sociology question remains the core
Despite sharing predictions of deep tech, one thing Peter stressed to which I echo, is that we need to be grounded in why we innovate new technologies. It’s not just to produce the best technology but we need to connect back to the users, to humans. In our strive for innovation, we must not forget how to be human because that is where robots cannot win. He said that moving forward, we need to figure out what we really want our technology to be doing. We already see questions about privacy and whether we are making the right thing?
“Are we building a system that’s working for them who are using it rather than the company that built it?”
Technology at the end of the day is meant to augment what we do as humans and make our lives easier and better. We often emphasise to our customers that buying our technology should not just be about automating their processes but it should also be about helping their employees be more fulfilled in their jobs. This means, freeing up their time from mundane tasks, so that they can engage in higher value tasks that they enjoy. The reason why many of us often dread work is because we face a lot of friction with the things we have to get done. Intelligent automation removes that friction. Intelligent automation frees the human.
Hear more on my conversation with Peter here.