What was fiction in the 60s became reality in the 2000s. Remember how this became this?
In 2011, America witnessed Watson – a computer developed in IBM’s DeepQA project compete on Jeopardy (television game show) against Ken Jennings – arguably one of the top data rich human beings on the planet, who could retrieve facts and answers out of the blue under the pressure of competition for 74 straight games, winning $2.5 million over the course of those consecutive games. .
Of course, give me a computer with access to all documents of law, science and history at my fingertips and I can find you an answer to any question in a few clicks. But does that mean the computer is smarter than human? And if computers can think, where are the investment opportunities going to show up in Artificial Intelligence, aka AI?
Over 100 years ago, Napoleon Hill was challenged by Andrew Carnegie to research and write Think and Grow Rich. This sums up the entire book:
What the mind can conceive typically starts out accepted as science fiction. That’s the only way people believed flight was possible before the Wright Brothers made it happen. Before we put a man on the moon, or scanned the heavens with a telescope located in orbit it was all fantasy.
The imagination of yesterday became the quest that drives startups to make it so.
Today, conversations in practical AI center around self-driving cars. But did you know that in 1986 Navlab – the first autonomous car was built by Carnegie Melon. AI impacts banking, marketing, entertainment, and even conversation! Imagine having a fluent, intelligent conversation in two languages as that conversation is translated real time! [source: Harvard]
The science behind computers that think is referred to as “deep learning”. Early in the development of deep learning, computers were being tested in games and challenges where decisions had to be made. The first notable contest was between Deep Blue – an IBM computer you might consider as Watson’s grandfather played against Garry Kasparov. Chess is a complex game involving strategy that anticipates the opponent’s moves and counters effectively in order to win. In a six game contest in 1996, Kasparov won three games and drew two, effectively winning the match. But in 1997, after a significant upgrade to Deep Blue, the rematch had Deep Blue winning 3 ½ to 2 ½ – becoming the first computer system to defeat a reigning world champion under standard chess tournament time controls, but Kasparov accused IBM of cheating. He felt that human chess players were intervening on behalf of the machine. There was human intervention between games to shore up weaknesses in the computer’s play during the match. But IBM denied interventions within the game play itself. [source: Wikipedia]
Popular Science Magazine posed the question about the outcome of a competition between Watson and Deep Blue. Watson’s program was compared to an English Major while Deep Blue was a math guy. It’s expected that computers of the future will likely merge data retrieval and computational power – broad and deep. [source: Popular Science]
The major premise is that machines can not only find answers to our questions but find the questions we don’t yet know to ask.
Here are a few startup companies in the artificial intelligence space.
AI. Reverie refers to itself as a synthetic data company. It serves the needs of businesses and other non-government entities (NGOs). One of the biggest problems they solve is taking an idea – a dream through the bottleneck of simulation data into sensor feeds and needing to have humans annotate them in order to train deep learning perception models – in effect hundreds and even thousands of humans performing these tasks.
Just as an example of this one industry impacted by this company’s work: the critical barrier to fully autonomous self-driving cars. (Can I get a fully autonomous house cleaner, anyone?)
Think about it. When you were 16, your hands were at 10 and 2, unless your friends were watching, and you were trying to appear experienced and nonchalant. Think of every single data point that gave you experience to anticipate children darting into the street, avoiding debris on the road or the semi drifting into your lane, or how to respond to a sudden slow down with no way to avoid a traffic jam.
How many times have you driven a car? Consider how many micro-decisions have you learned to make, and built into your central nervous system so that you drive and hardly recall what you encountered on the road by the time you get home. It happened naturally and automatically.
Add that to all of the cumulative decisions and experiences of other drivers. That’s an impossibly large amount of organic data to collect before you can plug in to simulations and teach a machine how to drive.
But, tie that outcome into data monitoring traffic flows, and you have the makings for fewer traffic jams, fewer accidents with far less death and personal physical damage and property damage.
But this is the challenge addressed and answered by AI. Reverie: how do you collect all of that data in a short period of time without the costs and barriers to plugging into the human driver’s nervous system.If you can take simulation technology and plug that data into a simulation platform that scales algorithmically to train machine learning models. And this can happen across markets – not just automobiles.
Founded in 2016, AI. Reverie has raised $5.8 Mil with it’s last seed round in April, 2020 raising $5.6 Mil of that total. Top investors include TechNexus Venture Collaborate and Vulcan Capital. [Source: Crunchbase]
Anodot is taking on the task of leveraging AI to detect and diagnose high-impact problems faster than is humanly possible. Just a sampling of industries and types of support include monitoring payments platform, business to business (B2B) customer success, gaming monetization, streaming services, trading, data quality and customer experience. The outcome focuses on safeguarding revenues and costs, digital partners and audience journey, experience and engagement as their client companies grow.
The company now has headquarters in the US, the UK, Australia, and Israel. David Drai, CEO and co-founder of Anodot blogged about their beginnings with a problem for a Russian taxi service trying to diagnose the cause for a drop in the number of rides. He said the challenge in gaining insight as to the cause pushed them to find ways to collect, visualize and track the metrics, and then innovate. He met Ira Cohen who was a data scientist and Shay Lang was an R&D manager with a security software firm. With their combined expertise, they were able to create the service that is now Anodot. [Source: Anodot]
Anodot has raised $64.5 Mil so far in 7 rounds of funding, $37 Mil of which came in their Series C round. Top investors included Intel Capital, Alicorn, Redline Capital, Disruptive Technologies Venture Capital, Aleph, Samsung NEXT and La Maison Compagnie ‘Investissement. [Source: Crunchbase]
I’ve addressed how AI is using data analytics to make autonomous vehicles safe, and data analytics to monitor and solve business problems. But what about talent management? HR departments have been using technology referred to as “the algorithm to sort through data and keywords that might be contained in a resume or application. As a result, skilled and talented candidates never make it to the interview stage, let alone an offer.
At the Silicon Slopes Summit in 2019 in Salt Lake City, Utah, I heard a speaker say that the algorithm isn’t going to find a John Stockton. If you didn’t follow basketball in the 1990s, you might not understand his point.
John Stockton was one of 74 college players invited to tryout for the 1984 US Olympic team and made the initial cut to 20, but was one of four released before the final team was selected. He was a relative unknown during his college career and when the Utah Jazz selected him as their pick during the 1st round of the NBA draft, the crowd in the arena were silent.
John Stockton would go on to become the legendary Hall of Fame NBA point guard for the Utah Jazz, and holds the NBA records for most career assists and steals. And he was a driving force with Karl Malone that took the long-shot, often underestimated small-market team to two NBA Championship Finals against Michael Jordan and the Chicago Bulls.
John Stockton was nearly an HR miss. And too often today, HR misses include heavy reliance on technology to find their ideal candidates.
Eightfold created what they call a “Talent Intelligence Platform” to break through the barrier for the talent as well as the companies who hire them. This way employers and recruiters can specifically define the skills they are looking for, and match skills and past experiences. And candidates can find the jobs they most want.
Eightfold’s CEO Ashutosh Garg studied at University of Illinois-Urbana Champagne, then worked in research for Google before starting his own company in the e-commerce personalization space. Eightfold was developed out of his desire to address the importance of employment for people, and how it defines and changes their lives. “In today’s time, you get a job based on who you know and not based on what you can do. The mission was to enable the right career path for a billion plus people.” [source: One Million by One Million]
In their most recent round of funding was in June, 2021, Eightfold raised $220 million of their total $396.8 Mil. Top investors include Foundation Capital, Lightspeed Venture Partners, Softbank Vision Fund, Institutional Venture Partners, General Catalyst and Capital One Growth Ventures. [Source: Crunchbase]
Viz.ai is approaching healthcare emergencies with artificial intelligence. After co-founder David Golan suffered a stroke, he and his co-founder Dr. Chris Mansi started the company, with the goal to reduce treatment and intervention time. The technology enables doctors and other medical professionals to upload CT angiogram images and within minutes determine whether there has been a stroke, and once confirmed alert the treating physician and surgical teams. During a stroke, 2 million brain cells die every minute and strokes are the #1 cause of disability in the US, making the urgency of the appropriate intervention response critical. [Source: Bloomberg, ]
Viz.ai’s goal is to reduce time to treatment and improve access to care. Every minute counts, as the billboards say.
Viz.ai technology was already established in 300 hospitals across the US in July 2019. The goal is to place the software in every US stroke center.
Top investors include Google Ventures, Grenoaks, Threshold Ventures, Insight Partners, CRV and Kleiner Perkins. The most recent round of funding (Series C) was announced in March of 2021 and brings the total funding amount to $151.6 Million. [Source: Crunchbase]
Mixmode is a cybersecurity startup company leveraging AI that learns and adapts to dynamically changing environments; MixMode calls it “Self-learning for autonomous cybersecurity” that can be deployed via cloud or onsite in minutes. [Source: mixmode.ai]
MixMode establishes a baseline over 7 days of normal network behavior; other platforms typically need 6 to 18 months to establish a full baseline.
MixMode points to the failure of cybersecurity platforms that use a patchwork of SOAR (Security Orchestration Automation and Response) SIEM (Security Information and Event Management), NTA (Network Traffic Analysis) and UBA (User Behavior Analytics) that fail to address the requirements they were intended to solve.
The big problem MixMode’s use of AI solves is proactive protection instead of just reactive protection. MixMode points out that advanced adversaries are also utilizing AI technology, and “the only option we have is to understand what distinguishes the good from the bad, and which type of AI or machine learning is truly helpful to the advancement of cybersecurity protections.” Mixmode goes on to say that current technology can only detect that an intruder is on the network but it’s already too late; that the intruder has already accomplished whatever he came to do. On the other hand, predictive AI works from a normal baseline to detect disturbances, making the system very difficult to trick. [MixMode Blog]
So far, MixMode has raised $17.3 Million, with the most recent influx of capital in April, 2020 coming from Entrada Ventures and Keshif Ventures.
The Bottom Line
Chances are, you’re interacting with artificial intelligence in some form or another every time you click on a website, take a drive, or have a medical procedure. Data analytics are getting smarter.
Forecasts for fully autonomous vehicles have been optimistic; in 2015 Elon Musk said self driving cars that could drive anywhere would be here within 2-3 years. But experts aren’t sure when, if ever that will happen. [Source: Wall Street Journal] AI is still learning, but even an intelligent assist could make a difference whether it’s your car, your home, your career, or your medical care.