The Future Of Computing – HPE’s Machine

Ok, I admit this is a little indulgent, but stick with me for a while.

Consider how technology has revolutionised life. How it has changed your life recently. I’ll wager you didn’t video conference family overseas from your phone until recently. That you rarely researched information about a purchase before going to the shops. That you only photographed special moments, holidays, birthdays, even with a digital camera, rather than your breakfast.

Now consider how this scales when we add sensors and actuators to everything with the Internet of Things.

Again, consider the amount of compute power, storage capacity, and network bandwidth you need for this amount of growth.

Even taking Moore’s Law into account, demand will massively outstrip supply. We simply do not have the processing power, let alone the data centre capacity and electricity to meet the exploding needs of data. Especially with IoT.

Computer Re-Architecture

If you were to reinvent computers today, with the technologies that are available, would you happen upon the same architecture?

HPE’s answer to that is, ‘No.’ The outcome of that answer is a research project called The Machine. Quite literally a rearchitecture of a 50 year old paradigm that drives all computing we know today.

The Machine In Summary

In short, the idea is to replace the expensive, quick, but volatile DRAM, and cheap, slow, but persistent Storage, with Non-Volatile Memory, or NVM.

Ions

HPE is working on a new form of NVM called the “memristor” that uses ions, rather than electrons, to store data.

Once you have massive amounts of cheap, energy efficient, and persistent but quick memory, you can dispense with 80% of code today. The code that deals with the ‘volatility hierarchy.’ Essentially swapping data between layers of memory and storage.

Now we can make a computer that is ‘memory centric’ rather than ‘processor centric.’ In fact you could theoretically have limitless memory, and allow any number of processors to act on the data in this byte addressable fabric.

Photons

Finally, to shift the data between the processers and NVM, we use optical fibre rather than copper wire. Photons rather than electrons. This massively increases bandwidth, and dramatically decreases energy requirements.

Happy to Talk

I’ve been working with the Machine Team at HPE Labs for well over a year now, and would be happy to deliver a presentation to your organisation if you’re interested in the future of computing.

The Unspoken Challenges Becoming An Uber Driver

One Smiling Uber Driver
What You Don’t Consider Becoming an Uber Driver Credt: PCMag.com

As an avid Uber passenger in over 16 cities around the world, I continually interview drivers. I like to understand their motivations, business model, success, challenges, previous occupations (over half were taxi drivers) and what they love about the career choice.

As you’d expect there are many that do this as a part-time, or transition, gig. A way to keep the lights on between jobs, or to supplement another “not-quite-enough” income. Equally there are an increasing number who view it as an alternative to “working for the man” in a corporate job. Something they don’t need training for, or to undergo an interview.

The Cool

Whilst ‘flexibility’ scores highest in my not-quite-empirical research, ‘meeting interesting people’ comes a very close second. (Although admittedly they may be simply flattering me)

The Not So Cool

What is interesting though are the challenges or difficulties.

For many there’s the sudden need to learn business accounting. Tax implications and GST responsibilities.

But the biggest surprise to me, by far, is loneliness. For better or for worse, whether you do this well or not, working in an office is a corporate affair. From formal meetings to the ‘watercooler.’

Sure you meet interesting people driving an Uber. But different people. Every time.

For most people, that takes getting used to.

Why Your Job Safe From Artificial Intelligence (For Now)

The Rumour Of Artificial Intelligence Is Greatly Exaggerated

It’s easy to both over and under estimate the effect Artificial Intelligence, indeed automation in general, is going to have on our economy.

It’s also interesting to note the emotive reaction to this technology driven progress. Recently I was chatting to a work colleague, a fellow Chief Technologist, and he was vehement in his denial that driverless cars will replace cars any time soon, if at all. His rationale is that people ‘like driving’ too much.

My response is whilst there are many who enjoy driving (myself included, not to mention motorbike riding) very few enjoy congestion, commuting, RTA’s, traffic enforcement, let alone the expense of owning a car. We put up with all of that for convenience. There were probably an equal ratio of people that enjoyed riding horses a hundred years ago. You can still ride horses for leisure, but no longer do they make sense for transport.

Autonomous vehicles are coming. The financial, social, and healthcare benefits are far too compelling for this not to happen, and all of the automotive manufacturers have this in their plans.

So what does this mean for people that drive for a living? Truck drivers (the job with the highest employment in the US), bus drivers, taxi drivers absolutely face disruption. Well over 50% of these jobs will ‘disappear’ in the next 10-15 years.

Augmentation Not Replacement

AI is augmenting our cognitive skills in the same way steam, then electricity, augmented our physical skills. And it’s already happening. If you use a smartphone, you already benefit from specific AI in your life. Even if it’s simply researching a purchase, or navigating.

But speak to any airline pilot and they’ll tell you why pilotless aircraft aren’t coming any time soon. Of course they’re biaised, most pilots are pilots because of their romantic dream of flying since childhood. Of all career choices there are very few pilots who don’t want to fly.

However, they do work for organisations that would drop the cost of payroll in an instant if they could.

Image of the book Passengers
Passengers centres on a plot about distrust in technological automation

Technological automation already plays a large role in a commercial jet airliner. Fly-by-wire systems use software to translate control inputs to flight surfaces. The F16 couldn’t fly without that. (Read my father’s novel, Passengers, written in 1983 about this very technology) And there’s the autopilot, increasingly becoming more sophisticated.

But autopilots don’t replace a human pilot. Although they have reduced crews of 3-5 down to 2. Autopilots simply automate a number of manual tasks, freeing the pilots up to apply their brain elsewhere. So they may not manually turn the yoke, or adjust the throttle, but they still decide when these events need to happen, and program the system accordingly.

If you never learned to program the autopilot and other systems, you couldn’t progress as a commercial pilot.

In the same way as technology reduces our cognitive burdens, we can apply our capabilities to higher order problems. Think of a taxi driver purchasing one or more autonomous vehicles, then determining the best routes and times to grow their business. Or a nurse spending more time listening to and caring for a patient because they no longer have to read charts and invasively take BP readings across the ward.

Despite the coming disruption, if you adopt new technologies, and learn digital skills, you’ll be best placed to benefit.

 

Interfaces Are The New Black

Image ofg hands manipulating data
How do we seamlessly interface to information from devices, in applications, and for humans? Credit: Deviant Art

If the Internet of Things is going to get off the ground, it’s clear that interfaces between devices and systems is key. Yet we have a legacy of proprietary systems that create complex interdependencies today.

The Past- Content

In a Knowledge Economy, “Content is King.”

But the underpinnings of this Economy, for the last 6 centuries knowledge, has evolved. Not that knowledge is no longer important. Far from it. It’s just that ‘owning knowledge’ is no longer an economic differentiator, because everyone, no matter their geography, demography, or intellect, has near real-time access to planetary knowledge. I say everyone, which of course is hyperbolic to make a point. But the truth is more people have access to the mobile Internet than ever before, closing on 3 billion, and that is unprecedented. Staggering in its implication.

The Present – Connection

So todays economy has shifted to an ‘Amplifier Economy.’ In this world it’s about execution. Being able to take knowledge and apply it. Build something, implement a process, develop a new market. Scale rapidly.

In this economy “Connection is King.” Because of the abundance of content, and the complexity of solutions, the agility in engaging experts and content from multiple domains trumps trying to own everything.

The Future – Context

But even as we move from competition to collaboration, from Systems of Record to Systems of Engagement, we build on both Content and Connection (or Collaboration). The next wave of technologies such as the Internet of Things, Wearable Tech, AI & Machine Learning, Automation of physical work like manufacture (robotics & 3D printing), transport (drones & autonomous vehicles), are ‘Systems of Foresight’ (Actionable Insight).

This drives an ‘Innovation Economy.’ It’s no longer about executing on knowledge faster, quicker, or cheaper than anyone else, but coming up with a new channel to market, a new solution. Reinventing the tried and true in a novel way. True mass customisation is not only possible, but imperative.

In this economy, ‘Context is King.’ Literally the context of given content – access to knowledge – and collaboration – access to expertise – change the game.

Those who have the most accurate context for a given citizen, voter, customer, patient in real time will forge ahead.

Interfaces – User, Application, and Device

Which brings me to interfaces. If a system is going to benefit from, or provide context, it needs to share information with other systems and people.

So whatever system you’re developing, it needs to be “Interface Centric.” Consider not only how users need to interface with the system, but other applications and devices.

Interfaces are the new black. User and Application Programming Interfaces.