What a Year
Everyone must have gone through a change they never had gone before in their life. When a major challenge is thrown to the world, it’s also a window for changemakers to make their way out from the ordinary.
The pandemic challenge, I must say, was a good challenge to our journey – we were probably one of the very first companies confidently rushing into remote working mode before the government even announced the orange alarm. It is our habit to initiate change before change happens. In this process, we have made a lot of contingent but memorable initiatives, including an indie short film, a few new products, and some remarkable projects. In the end, not only did we keep the whole team, we managed to double its size. This is only possible because of the trust from our clients, partners, investors and the technologies that everyone had tremendous faith in.
2020 is the year that I witnessed a lot of disruptive technologies that managed to stay. At the end of 2020, I’d like to share some of my observations and thoughts on what we should expect for the upcoming year of 2021. I hope these insights will inspire your journey as a changemaker.
This year many lives have been taken, but their digital assets may not have been well managed. The digital legacy is growing the entropy of the cyber world.
Early this year, I myself had to take medical leave, and the situation forced me to practise how to hand over every single access to the person in charge. It’s a unique and fruitful experience when you start to take yourselves out of the things you control all the time. When we look at a larger scale, it’s a practice of building the vision that carries on without you.
Collective intelligence is the concept that we started to establish using digital technology. We have established some meaningful practices and managed to digitize the whole MNC organization’s long SOP into a live interactive data storytelling experience. In this process, we digitized leaders of great companies into life-like virtual avatars that can speak and interact. We even digitized more than 3000 university students for them to celebrate their best graduation moments in 2020.
Meanwhile, whatever has been digitized before, can also be reproduced with AI. There are a few noticeable videos about restoring the color and resolution from archive footage. Developers online made a full HD, 60fps colored video from a 1910 San Francisco black-white footage. Our team also made a successful test to restore 1950 Singapore election footage using similar approach.
The magic of recreation is powered by synthetic data that is used to fill missing information from learning the patterns from the given information. But one thing we must be aware of is that AI can only produce based on the info we provided to them; it is not able to imagine something beyond the boundaries of a predictable data set. For example, an AI cannot generate more accurate facial features of someone you already know if those details only exist in your memories and not in the data sample.
Given the limitations of this technology, our eyes have a tendency to accept the results and never really question the accuracy of these restored footage as there is no way to compare. When we start to accept this, we should be prepared to see a lot more creations. We should also realize synthetic data has started to help us decide what our past should look like.
You might be asking: can we use such technology?
When you create an animated emoji on your iPhone that uses your voice and mirrors your facial expressions, the same depth detection technology can be used to create a virtual 3D avatar of yourself or the entire team. From saving a few photo frames in your room just 10 years ago, to keeping millions of photos and videos in the mobile and cloud, today, we are able to store spatial information like 3D scanned faces, environments, and moments of your life in specific devices like light-field projector, and display them around you.
Technologies like volumetric capturing, facial capturing as well as mobile-based 3D scanning are getting a lot more accessible this year with devices like the latest iPhone with Lidar scanner, or consumer-level 360 cameras. I encourage you to create more of such spatial assets as these will better preserve your past.
As remote work becomes the new normal, communication, collaboration, and training are going to get more powered by digital technology. Every company is moving towards being a tech company this year. When Singapore, where our HQ is based, starts to attract global tech firms, one’s future boss could be anyone, but will most likely be someone that knows tech better than your previous boss. How will your organization adapt to keep the top talent for their future roadmap? If you can’t even meet people in the office, how would you manage their demands like how you used to do?
It all starts from a simple question: are you moving towards a digital organization? When you start to manage things using a data-driven approach instead of a people-driven approach, you’ll identify more accurate patterns with minimum human influence.
Here’s a very simple example: As a site manager, when you replace your site visit check-in/ check-out form with an online form connected to a backend database, you are using a consistent logical program to eliminate any scope of human error that may arise from manual checking. In addition to that, you can even generate insights by tracking traffic and peak activities, to decide how to optimize onsite man hours.
You can also start developing a series of connected data reports across different departments. Open API is one new interface that you should definitely be aware of. It allows almost any software to talk to each other, and customize the end solutions for you. For example, your Microsoft team calendar can be integrated with your sales report from the finance software, to generate insights on every business developer’s efficiency. Programming languages like Python will help your team to automate processes. For example, you can keep eyes on marketing information and auto-generate reports, without having to find someone to prepare a powerpoint for you.
For the most part, the abstract of human input can be encoded if it happens on a digital platform. Digital meeting minutes tool can be used to save the meeting record to texts automatically. But the text record has no value until you link it to something else, like your task management platform, or NLP (Natural language processing) database that will pick up the key message that is relevant to your organization.
One of the disruptive concepts is to take yourself as a filter of data that you have access to. Data is a different creature, you need to observe, elaborate, adjust the variables, test and deploy, instead of trying to control it. When the entire operation data of your organization is becoming transparent, your vision and mission can also become clear enough to every teammate. It’ll help you fill the gaps caused by people management.
Besides contextual data, we are also seeing synthetic data getting closer to describe the world we are living in today. A game engine has been used to create realistic scenarios for filming. For the first time in history, a new Starwar Series managed to be produced from a single indoor studio.
How can we take advantage of this trend?
This year, we have successfully deployed a good number of immersive training programs with dynamic data reports of user behavior. Work like calibration of a machine and assembling complicated components, is being fully replicated in digital format. We are able to trace the completed user behavior data and provide insight for managers. Meanwhile, the same digital assets can also be used for the machine learning process to observe the site operation. These solutions that replicate the element and dynamic nature of any physical entity are called digital twin technology.
We believe digital twin technology will stay and scale across diverse organizations, from simulation planning, product design and testing, observing company’s live operation details, to making predictive scenarios for decision making. But digital twins can’t be completely established by external parties, it must be you who should look into the infrastructure. Is your teammate comfortable to use Excel, or a cloud based CRM, or any mobile app to provide data? How’s the security protocol? What are the test and deploy environments? Build it, piece by piece, it’s never too early to build the digital twin for your organization.
We have seen a lot of insane machines this year. After a multi-generation of updates, AlphaGo has finally met it’s real challenge, to find a challenging game to play. It’s recent version beats the best human player in any type of games without knowing the game rule in advance. Hyundai has acquired Boston Dynamic’s star product “Spot”(The big dog) and “Pick”, while the famous human-like robots which can do backflips are left over to decide it’s development roadmap.
Meanwhile, massive user applications like Grab, TikTok and Echo (amazon speaker) are receiving, processing, and inventing user data within their ecosystem. AI is deciding the best price for your trips, the best videos to kill your time, or the best ways to teach your kids via it’s unlimited online story libraries. Our dependency on these technologies made them a permanent part of our life.
Further down, are we making more human-like robots or making more robot-like human beings? I believe each party should naturally respect the advantages from the other side.
When we look at distributing universal knowledge across organizations, data-driven intelligence wins easily, as code takes seconds to reproduce, but human beings take years.
On the other hand, humans are very good in general descriptive intelligence, which can help machine learning to direct its computing priority. That’s why most machine learning projects require heavy manpower in calibrating data to improve the training model’s accuracy. This whole process of training-calibration-deployment is happening almost real time like how humans function today.
How can we benefit from this technology?
We should start to accept that dynamic human input is vital to improve any machine’s prescript performance. For example, When you turn on google meets auto caption function, the program auto converts your audio to any language in real time for better communication; meanwhile, it’s also learning your accent to improve the quality of speech-to-text result. You can improve it by purposely testing some pronunciations for benchmark learning.
Keep an eye on all technologies you are using. Do they have any protocol to further improve its performance by learning human’s behavior? Does your vacuum robot track each manual cleaning task for you to decide the most effective auto cleaning schedule? Does your CCTV have motion detection for you to set some event based video recording policy? Does your own IOT device gather information from your input and further adjust it’s own behavior? I had a lot of fun teaching my 4-years old kid to learn robotic programming, and we managed to develop a customized personality for our ROBOMASTER (a robot product by DJI). There are a number of visual coding programs like SCRATCH that allow you to write functions without knowing the code.
So, observe and identify a scenario and start to experiment with programmable technologies around you. It’ll be your first step towards building a more comprehensive human-machine collaboration workflow.
Thanks to Mixed Reality devices like Hololens2, our caveman brain has met it’s 1.0 version plugin to improve our spatial intelligence. Spatial computing devices help us navigate and visualize information by projecting 3D elements in world space. Although this plugin still runs at only 10% of our retina resolution, it’s convenience has been recognized. A worker can easily identify physical objects, read barcodes, or navigate themselves around a full warehouse without any intense training in advance. Relevant digital information is loaded and projected in real-time into their retina wherever they go. On-site orientation will be replaced by on the job spatial assistant solutions.
Further down, a series of technologies like beacon, spatial anchor, and hybrid algorithm that combines machine vision, location based service and cloud computing, will help this device to connect with a master brain behind the control room. Edge computing and rendering technology will allow any lightweight portable device to process it as good as a super computer when dealing with massive low latency data.
How can we benefit from this technology?
Spatial data is generated in the environment that has enough bandwidth to maintain consistent computing when the user moves. Your first step is to look at infrastructure updates.
Are you running the service on premise or online? Can you deploy mixed cloud structure to separate your secured operation data and remote/co-location/universal mapping data? Do you need to consider applying Container concepts to distribute different computing services? How do different sensors and machine vision enhance data collection? Lastly, dive into the immersive solution like AR and MR when you have a good understanding of the spatial data structure. A dedicated solution architect will be a great help to plan out your exploration.
We believe in the power of spatial data computing in the long run, starting from growing the future workforce for industries like logistics, manufacturing and transportation. It’ll scale up to smart city level to improve national level collaboration efficiency and instant decision making.
Last but not least, we are used to attending different types of online meetings and events. When we resume our daily work in the office, we might still lean towards digital connections, instead of standing up and reaching out to the person. This new phenomenon will continue to stay.
In 2011, I made a low budget short film on holograms being used for remote communication. It’s been almost 10 years, with so many generations of OS, UI/UX and hardware advancements. But the nature of media still hasn’t changed. Video call just scaled and stayed, but not moving. When was the last time you got goosebumps from a video call?
Sadly speaking, there isn’t any other accessible digital technology that can fully upgrade us at this moment, at least not in 2021. So no matter what fancy idea you’ve got from a virtual event, do set your expectations carefully. The priorities are still about delivering a linear, engaging presentation experience, instead of getting your virtual booth more virtual visitors, or making more virtual friends.
Hence, it’s wise to invest in a proper studio for your own organization. The trends of virtual events will stay and most of software will be open sourced. It’ll only cost you a few thousands of dollars to set up the studio. After that, you can run it by yourselves. Considering the reusability, it’ll be good to own the setup for basic inquiries for the next few years.
Meanwhile, there are a handful technologies that can make your virtual events more engaging – like AR broadcasting or hologram panel discussion. We believe those powerup technologies will increase in popularity in the next 5 years.
As we enter 2021, we are starting to accept the concept – disruptive changes are here to stay. Remote work is here to stay – how you manage the organization matters more than how you manage people. Digital organization is here to stay – how you build your digital infrastructure matters more than which consultant can give you a better vision.
The good news is, the world is here to stay, so you will always have a chance to win. We are here to help those who are at the frontier embracing the changes to achieve their success. It’s a wonderful journey that we have embarked on with everyone ever since we started our journey 7 years ago. We wish you a great new year ahead, and looking forward to exploring more exciting new changes together.
Ender Jiang with team Hiverlab