The Future Stack AI Newsletter #4
Welcome to the fourth edition of the newsletter that brings Australians the latest developments in AI from all over the world.
If you were not able to attend the NVIDIA GPU Technology Conference, find the notes here.
⚡New this week
The biggest trends for AI in the next 10 years
Do you want to be a consumer or a producer and why it is going to matter more than ever?
AI power centers in Australia
How my friend successfully changed careers from software engineering to AI tech lead?
💗Mental Health Check-in
This is the toughest section for me on days I am struggling with my own mental health. But I like to remind myself that we are all in this together and life is not a zero-sum game. Uplift someone else today and feel your vibrations change.
Before moving on to the next section, think of something that made you smile today and then smile at yourself in the mirror.
🌍AI in the world
Applied AI trends for the next 10 years!!
I know you’re tired of hearing about trends, but they are a cliche for a reason. Humans like to predict the future even though we are terrible at it. But, we are pretty good at recognizing patterns.
And do you know what all of these trends have in common? They are all based on observing the world around us, making sense of it, and then recommending decisions based on data (autonomously in some cases). Apply this anywhere you want, and you will see how AI is immediately applicable to the industry of your choice. My current list is based on the relative number of people in the AI world working in these domains, because,
People == Momentum == Progress
Arts
AI for content creation is already beginning to pick up. Instories raised $200 million, Scalenut raised $3 million, Aussie startup Outwrite has over a million users, and many many others. Follow the money, and you can see the trends for yourself.
AI art generators are taking the world by storm. Dall-E 2 (not open-sourced yet), Midjourney, and Hugging Face are some of the places you can go to generate your own art. There are some legal issues when it comes to AI-generated art, but it’s only a matter of time before artists would start incorporating AI into their workflows. Another avenue is photo/video editing, and I see that Adobe research is putting quite a lot of effort into augmenting it with AI.
Healthcare
The way I see it, AI in healthcare is still in its nascent stages. The good news is that we are collecting more data than ever, we have better techniques to protect that data, and we also have better AI techniques that are directly applicable to this domain.
AI for medical diagnosis and prognosis will get much better with more data. Identifying epigenetic factors from a person’s DNA will become the cornerstone of personalized healthcare and it is now ready to reach the masses. Freenome is one such company.
We will have to be careful though, not to overplay the role of AI, and build sensible human-in-the-loop solutions for healthcare.
Autonomous Systems
Everyone knows about self-driving cars and the race to create fully autonomous driving systems. But the less known side of the same coin is the immense effort being put into defence towards autonomous submarines, fighter jets, and drones. Just think of a situation that endangers the life of a human driver/pilot, and that is perfect for a fully autonomous system. Again, as in healthcare, it would be dangerous, not to create human-in-the-loop systems. But it is much more difficult to do that if the system cannot communicate with a human without exposing itself which is often the case for submarines.
Education
The next front in education is the incorporation of chatbots, online learning, and virtual classrooms, all powered by VR, AR, and AI. Stanford professor Younes Bensouda Mourri has created LiveTech.AI which has affordable project-based CS courses. For educators, LiveTech offers personalized help by training chatbots on the curriculum and creating an AI Teaching Assistant. Auto-graders are another feature they are working on.
Virtual worlds
Gaming: Let’s face it, it’s not just about cool graphics anymore. How awesome would it be if our video game opponents can think intelligently? Apart from pathfinding and finite state machines that are already used in video games, how about using deep reinforcement learning, computer vision, and NLP techniques to decide the behavior of characters? Well, it’s happening and although such AI would be less predictable, this is where game development is heading according to NYU researcher Julian Togelius.
Metaverse: When used in the right way, virtual worlds can supercharge learning, improve productivity, and enhance the human experience. We will take a while to get there, but we will.
The opportunity here is to create environments that are tailored for the success of an individual which will ensure humans keep pushing boundaries. It is not about escaping reality to live in a virtual world, but rather about leveraging technology to upgrade our reality.
Climate Tech
NVIDIA is building Earth 2.0, a supercomputer designed to run AI models for climate science. Australian National University’s ACCESS is another earth system simulator for climate research. Such projects hope to leverage modeling and simulation, high-performance computing, and cutting-edge research in physics-based deep learning to help understand the consequences and mitigate climate change.
Want to know how to see the trends and even create trends in your industry?
Think about data first. Where can you find opportunities to get more data? Which industries have untapped data that you can use and leverage to produce insights? All cool AI starts with this - an untapped data potential.
The most important decision you will make in your life.
Maybe this decision is not as important as choosing your life partner, but it is a close second, especially in this increasingly digital world.
Are you going to be a producer or a consumer?
Technology only becomes exponential when it reaches more people than ever and becomes easier for them to use and create using it. Think of the first browser, nobody except computer scientists could access the internet. Think about what happened next. Browsers became commonplace, the internet blew up, and so many more people became part of the internet revolution, not just computer scientists. This is going to happen with AI now (it’s already happening).
This poses a question before all of us. Are we going to only consume the wonders of cool technology or be creators in this new world order?
🦘AI in Australia
If you come to Australia, do keep them on your bucket list of AI adventures 😉
Australian Institute for Machine Learning, University of Adelaide, Adelaide
With research strength in robotic vision, medical machine learning, trusted autonomous systems, surveillance, and tracking, 3D modeling, and augmented reasoning, AIML is one of the world’s best facilities if you are into Computer Vision. They are situated on Lot Fourteen, which is one of the biggest innovation hubs in South Australia.
CSIRO, Australia-wide
With researchers in every field you can name, CSIRO is Australia’s national science organization. Check out their research here to find out what interests you. They also have industry SME programs to help companies develop prototypes, consulting services, as well as research translation and commercialization programs. Some of the best AI researchers I know work or have worked at CSIRO.
Australian National University, Canberra
The College of Engineering and Computer Science at ANU houses many AI programs with world-leading researchers as guides throughout the journey. ANU is a research-intensive university, and that was my reason to pursue my master’s here. Data61 (CSIRO) also has a presence on the campus. The entrepreneurial spirit though is added by Canberra Innovation Network as well as the Centre for Entrepreneurial and Agriculture Technology which are a stone throw’s away from the campus. ANU has a special focus on Computer Vision with great faculty and courses. ANU’s 3A Institute which hosts the Master of Applied Cybernetics (the first and only master’s in applied cybernetics in the world) is an interdisciplinary research body that is focusing on designing safe and responsible cyber-physical systems.
UTS, Sydney
The Australian Artificial Intelligence Institute is a dream for any applied ML scientist or engineer. Their focus on healthcare, marketing, autonomous systems, finance, and education makes it a great place for anyone looking to test their skills on industry problems. Sydney is also the ideal place to be if you are interested in FinTech or healthcare due to the ecosystem enabling these industries.
UQ, Queensland
I love this one! First of all, my favorite applied AI researcher - Jeremy Howard just started teaching here. Secondly, I love their message about creating an integrated AI ecosystem to expand human potential. Find the link here.
RMIT, Melbourne
If you want to learn more about robotics and autonomous systems as well as find deep expertise in natural language processing, RMIT’s AI innovation lab is the place to be. Although they do much more, their expertise in robotics is unparalleled.
🔧Upskilling
A friend approached me asking for some advice on his career switch to an ML position. He is my mentor and I am in awe of his software development skills, but he was not sure if he wanted to take a leadership position in an AI company. Here, he would have to not only manage software developers, but also data scientists and ML engineers.
His concern was that he will be transitioning to a field he is not an expert in and will be managing experts in ML. If you are in the same boat, I will tell you what I told him.
Focus on the business problem and not the technology.
Define clear success metrics for your team and trust them to figure out the rest.
At the individual contributor level, you need more courage, but when you start managing people, trust comes first.
The “why” and “what” comes before the “how”. The job of a team lead is to articulate them as clearly as they can, and the “how” will take care of itself.
Find someone in your new job to support you as you switch careers and learn as much as you can from them. It’s much faster than figuring things out yourself.
I am happy to report that he decided to take the leap and is now very happy learning new things and pushing the boundaries of his own career, as well as running an awesome team of data scientists and ML engineers.
👩💻 About me
I am an ML engineer in Canberra and my most intense obsession is to help companies and individuals become competitive by investing in data and AI. Subscribe to my newsletter if you have similar ideas on making a dent in the universe. All the content that I create is free and I intend to keep it so!
Follow me on LinkedIn here!
Love the post about Upskilling. It gave me a new perspective to look at things.