Thanksgiving travel is seldom fun. It seems every year I come closer to the edge of missing the flight that will disappoint someone I care about. As I rushed to catch my flight the other week, however, my smartphone began tilting the field back in my favor – by making decisions for me. Rather than speeding blindly into a traffic jam, my phone guided me off the Mass Pike, over streets I hadn’t seen in ages, delivering me to my flight on time as if traffic didn’t exist. Its actions not only helped me catch my flight, but also helped clear the way for emergency crews to assist those impacted. When the radio eventually reported the accident, I realized this combination of events was yet another example of how autonomic intelligence is revolutionizing our lives.
Of all the technology trends I’ve observed over time at Bain Capital Ventures, Ai holds the greatest promise.
While the term autonomic originates in biology – referring to unconscious systems controlling basic functions like your heart rate – it takes on new meaning when applied to technology. Indeed, autonomic intelligence has enabled software capabilities to evolve from basic data collection, reporting and alerting to actually making decisions, enhancing human cognitive function. It enables humans to manage increasingly complex scenarios at new levels of scale - consider how a market operates autonomically based on supply and demand - in ways we can’t even imagine. It will change the way we work, but also promises to elevate those willing to adapt. Of all the technology trends I’ve observed over time at Bain Capital Ventures, autonomic intelligence holds the greatest promise.
Autonomic Intelligence as Cognitive Augmentation
Interestingly, the drive for such cognitive augmentation predates software. In 1912, Sperry invented the first autopilot. Its design outcome solved problems similar to those addressed by autonomic intelligence. By relieving pilots from making minute adjustments to a plane’s elevators and rudders – a repetitive, highly-fatiguing task – autopilot enabled them to focus on higher-order considerations, like “are we flying in the right direction?” Today, software-based autopilot controls much of the in-flight experience, radically improving air safety.
Another type of intelligence in the news more recently is machine learning, enabling computers to deliver inferences without being explicitly programmed. Recall IBM’s Watson beating human champions in chess or on Jeopardy – and more recently at Go, once thought impossible. Such are examples of deep intelligence that is machine learning – machines studying specific datasets to discern patterns to more accurately predict outcomes within those datasets. Often, machine learning can be used to more accurately predict failure based on anomalies – not necessarily predicting the future, but spot potential issues based on prior behavior.
In the case of machine learning, “artificial intelligence” (AI) is doing things that humans can do (chess or Go) but at far greater speed and efficiency. Conversely, autonomic intelligence (Ai) makes decisions at a scale impossible for the human mind alone, achieving outcomes that were previously impossible.
Consider how turn-by-turn navigation like Google maps revolutionized my trip to the airport. It uses Ai to help a driver like me get somewhere via the safest and fastest route. Google doesn’t know what your safest and fastest route is at any point in time; the decisions guiding you to your preferred outcome are generated by the system itself – the product of millions of data points on speed and position gathered from other users. A system organizing itself.
Google doesn’t know what your safest and fastest route is at any point in time; the decisions ... are generated by the system itself – the product of millions of data points on speed and position gathered from other users. A system organizing itself.
In this way, autonomic intelligence not only can re-route cars around accidents and traffic jams to shorten commutes today, but ultimately with self-driving cars, may enable humans to travel closer together at higher speeds – more efficiently – than humans driving alone. This combination of higher precision with a broader understanding of traffic will enable more intelligently routing packets of cars through a network of roads.
Autonomic Intelligence in Data Centers and Clouds
Data centers and clouds powering applications like Google maps represent another opportunity for the Ai revolution. For decades, data centers required humans to make all application deployment, placement and sizing decisions. Unfortunately, human oversight can no longer deliver the performance necessary for a web-scale world where “slow” now means “down.” Today, a type of autonomic intelligence guarantees application performance across private, hybrid and multi-cloud data centers by enabling apps themselves to select the best infrastructure on which to run – effectively self-organizing: An autonomic platform controlling the data center.
An autonomic platform controlling the data center.
A recent Deloitte Tech Trends report argued that autonomic platforms in IT “can remove dependencies between business outcomes and underlying solutions, while also redeploying IT talent from rote low-value work to the higher-order capabilities needed to deliver right-speed IT.”
That argument is spot on. Autonomic systems enhance human cognition, enabling humans to focus on the higher-order: administrators can move from lower-value, labor-intensive tasks, elevating to higher-level architecture and planning functions. And like Google Maps, autonomic systems don’t omnisciently know the best home for any application workload – those decisions are the result of the system self-organizing. Enabling IT to run at the speed of business.
That’s what we’ve built here at Turbonomic: autonomic intelligence that can manage the most complex, mission-critical bit of real estate on the planet – the cloud (or data center, depending on your perspective). Just as economic markets self-organize to determine price based on supply and demand, Turbonomic leverages those same economic principles in your cloud or data center. As my old friend (and now Turbonomic Executive Chairman) Bill Veghte wrote recently in Pulse:
"As a customer, you will come to see [Turbonomic] as a platform that dramatically lowers the cost and improves the quality of experience for the applications and web services you rely on to run your business. It is the product that you use to understand and optimize your infrastructure and to run your applications and web services optimally, whether they are cloud native in a public cloud or traditional applications behind your firewall. You rely on it to optimize your traditional environments. You look to it to become your hybrid cloud management platform, providing the elastic capacity, private or public. Over 1600 businesses have adopted the platform, with some of the largest and best run IT shops in the world relying on it – and lots more signing on every day.
"As a technologist, you see it as a platform with an elegant architecture that effectively matches application demand with infrastructure supply in real-time, regardless of the environment, on-premises or off, cloud native or traditional, VM or container. It accomplishes this by creating a data model (abstraction) that is a semantically rich representation for introspecting and controlling the environment. There is an intelligent decision analysis engine that takes this representation and makes continuous real time decisions to enable the desired state. Finally it is automation… an orchestrated set of actions driven by the analytics engine to control any workload (current or new) on any infrastructure. It accomplishes this not through probes and agents but through APIs giving partners and customers the ability to build to it, customize and optimize it on their own as well.
"These three elements come together as an autonomic software platform enabling applications to self manage, matching application demand with infrastructure supply in real-time."
I love Bill’s perspective through the lens of both the customer and the technologist, but the key to both is “self management”. That’s what Turbonomic’s autonomic system does, on a moment-by-moment basis in your data center or cloud – it enables your own environment to self organize using budget and price to manage itself, just as your body’s autonomic system adjusts your heart rate and breathing during your morning run without a single thought by you. Freeing you up to think about other things, to imagine the extraordinary, with all the extra bandwidth your autonomic system gives you back.
Autonomic intelligence is even more critical as enterprises move at scale to public and hybrid clouds
Autonomic intelligence is even more critical as enterprises move at scale to public and hybrid clouds. Once you leave the control you have in your on-premises data center for the exponential complexity of a public cloud, autonomic intelligence becomes more vital. Like Sperry’s autopilot guiding pilots through the clouds, enterprises can leverage the power of Ai to guide them to the right cloud – and flying blindly to the public cloud can lead to very expensive mistakes.
Autonomic for the People
So where does Ai leave us, beyond helping manage clouds at scale or me miss fewer flights? Like other types of AI, autonomic intelligence won’t replace humans – after all, no matter how advanced Sperry’s autopilot, I doubt any of us would board a plane without a human pilot – but it will both elevate and change human work as much as our lives.
For those who do embrace it, autonomic intelligence will not only augment but unleash human intelligence...
There are data center operators who may never trust software to make decisions, just as some developers are disinterested in anything other than Java monoliths. And that’s fine. For those who do embrace it, Ai will not only augment but unleash human intelligence, as the creative power of the mind – that software will never replicate – focuses on, for example, architecting clouds to deliver faster-processing Hadoop clusters, unleashing insights into the human genome, our changing climate, perhaps even into our Universe itself.
Driven by the explosion in today’s smartphones, their computing power (132,000x more powerful than Steve Jobs’ original “bicycle for the mind” – the Macintosh), and the clouds they leverage (more powerful than billions of smartphones), autonomic intelligence empowers humans in ways Sperry could never have imagined. A power we’re helpless without (just imagine me at TSA if I had forgotten my smartphone).
Autonomic intelligence is the power to extend human cognition, rather than replace it – freeing humans to do what we do best: create, design, and collaborate at a scale and complexity beyond which the human mind alone cannot operate.
And also ensure more Thanksgivings go off without a hitch.
Originally published on LinkedIn Pulse here.