One of my hobbies is to watch key lectures from experts that are freely available on Youtube – there really is a great wealth of knowledge in those lectures that we all can access for free.
Renowned institutions invite world-class experts from hot /promising areas for keynote lectures and often upload them on their YouTube channel – institutions like MIT, Berkeley, Stanford GSB, among many others really serve the public by sharing what had traditionally been available exclusively to those that belong to those institutions.
As a growth investor, I am always on the lookout to learn more about for the secular, mega trend so that I can understand the mega innovation that will have transformative power – AI, big data, cloud computing.
Today, I was watching a lecture by Tom Siebel at MIT. In enterprise technology, he has unparalleled credentials AND track record – there is so much to learn from his message, but more importantly, he is someone who has ears of key decision makers around the world, influences those decision markets, and therefore drives the industry trend.
He is a successful computer scientist, entrepreneur (founded Siebel Systems and C3.Ai), capitalist, (sold his CRM company, Siebel Systems to Oracle for $5.8bn in 2005), and visionary (essentially created CRM industry).
As I was watching his lecture, two things really got stuck with me – which I wanted to share here – it has implications for EVERYTHING and for next-gen enterprise SaaS company.
BIG DATA MEANS NO MORE SAMPLING ERROR BECAUSE YOU HAVE EVERY DATA POINT
As a former scientist, this was an extremely powerful notion for me – no sampling error means enormous upside to power/accuracy of the predictive analytics – which has HUGE implications on EVERYTHING around us.
This is all driven essentially by unlimited computing power and data storage that has become enabled through adoption of cloud computing.
Constant stream of data allows constant adjustment of extrapolation-based predictive analytics and this has enormous implications in so many industries –
What NO SAMPLING ERROR MEANS FOR HEALTHCARE
Clinical trials:
Failure of clinical trials is largely driven by underpowering that is due to not enriching study population enough for the effect size of the drug (unless the drug simply does not work) – this is just intrinsic volatility that comes from sampling error when the study was designed. With more complete data.
Increase probability of success for drug program leads to higher valuation for the whole sector.
Patient care for chronic diseases:
Chronic diseases (diabetes, hypercholesterolemia, or hypertension) are becoming larger problem for our society as our life expectancy continues to increase. Quality of life is significantly declining for these patients and financial burden is becoming increasingly unbearable. This situation calls for innovative solutions. What can big data-based AI do to solve these problems? For one, continued monitoring of vital signs or parameter of concern (such as blood glucose level) can provide powerful preventive signal for patients or healthcare professionals before at-risk patient goes to serious condition that can cause irreversible damage or expensive treatment at hospitals.
Or patients no longer need to be bound to hospital beds and can get discharged earlier – as long as there is constant monitoring system that is AI-powered such that there is 24 hour surveillance on patients that is just as good as, if not better, being in hospital.
OTHER IMPLICATIONS
There is WAY MORE applications of big data – it can be used for 3D design/printing for implants (you can bake in for all unique movements of the patients’ body), more precise / accurate diagnosis (fewer false positives or negatives), or better drug design. It really is upto imagination – our bodies are always constantly sending out signals that can be stored and analyzed for practical use.
HOW ABOUT IMPLICATIONS FOR ENTERPRISE SaaS?
For all the above, the data collection and analytics are always constant iteration as continuous data stream constantly corrects and refines the predictive model to continuously reduce or eliminate sampling error. What this means is the importance of platform for smooth interconnectedness.
Tom Siebel compared buying AI platform to buying a car. At MIT, there is a lot of smart people who are capable of building a new car by assembling parts from various manufacturers, but they all resort to buy a car (he mentioned Ford, but I think he should have mentioned Tesla) because the parts are constantly interacting with each other and there needs to be one company that deals with the parts as a WHOLE.
He saw the same approach with enterprise AI – the smooth interconnectedness is vital for not only creation, but also maintenance of the whole program and it is much better to buy as a whole than to try to build its own (by hiring a bunch of consultants from Accenture or other consulting firms that stitch together a bunch of different SaaS vendors and walk away when done).
CONCLUSION
“No sampling error” and “platform approach” to AI were mind-boggling for me. For this, I am bullish on the future of:
Teladoc ($TDOC): absolute leader in telemedicine, but I think their chronic disease management platform is actually the most valuable business that is just the beginning.
C3.AI ($AI) and Palantir ($PLTR): these two companies are the only ones that are really pushing for platform approach to ML and AI for big data analytics. Needless to say, both have strong positioning in industries that have heavy data generation – manufacturing, sales, security and healthcare among others.
Link to Tom Siebel’s lecture at MIT is here: https://www.youtube.com/watch?v=-W10Wxoh_uA&t=780s&ab_channel=MITSloanCIOSymposiumVideos
What is ur favorite theme/trend that will be transformed by AI/ML/big data? Please share in the comment below!
* not investment advice
