Don’t Forget How to Fly The Plane – Managing Technology Transitions Safely

A video of this recent impressive cross-wind landing at Schiphol airport found its way to my Facebook feed. Great work from the pilot, but in the comments I saw a friend of a friend wondering if that was the auto-landing software or a pilot at work.

Well, I can assure you it was the pilot; the wind speed at the Schiphol airport averaged 40kts at the time the video was taken, with gusts to 55kts. However, B777’s maximum wind limit for auto-landing is 25kts.

However, for people to even consider that it could’ve been an autopilot got me thinking on one topic that is currently making our world temporarily more dangerous: we are over-reliant on technology that is not ready to take over from us.

With all the – completely valid, I might add – concerns about artificial intelligence, there is a more imminent risk that we are facing, and living with every single day without paying much attention to it. Ironically, the culprit is technological progress; it’s a problem because there’s enough of it, but not enough of it.

Forgetting how to fly the plane

As technology becomes good enough to do some work on our behalf, we tend to be a bit lazy and let it do that work on our behalf.

At first, it breaks down frequently enough so that we retain the skills necessary to take over when needed. But as it improves, we use those skills less and less, until we begin to forget how to take over in the first place – but at the same time, the technology is not yet so developed as to be superior to humans in handling all situations, nor is it infallible or often even particularly resilient. We enter a dangerous chasm of being over-reliant on immature technology.

That’s the dangerous zone we currently live in, and it’s happening everywhere. Like the pilots on Air France flight 447, we happily rely on the computers to run the show for us when it’s smooth sailing. But when something goes wrong, we, like the pilots, suddenly find that we lack the experience to deal with the situation.

Under pressure, we’ve forgotten how to fly the plane; the pilots’ very first responsibility.

The tragedy is that this is unlikely to have happened if we had never had the autopilot taking care of things for us in the first place.

Progress and resilience

And that is how things go, on scales small and large. Over the past centuries and increasingly rapidly over the past decades, humanity has piled on technology upon technology, and as they have grown to be good enough most of the time, we forget how to live without them and rely on them, somehow wrongly assuming they are better than us all of the time.

Whether or not that has disastrous effects like on AF 447 is often a matter of luck.

It’s a matter of luck we haven’t had another Carrington event or a well-coordinated terrorist attack that could disable a large electricity grid for weeks or months.

It’s a matter of luck we’ve avoided a globally devastating pandemic for so many years.

It’s a matter of luck we haven’t hit any climate change tipping points yet. Maybe.

This is good, because we’re not ready for any of them. The electricity grid is remarkably fragile, and we are totally dependent on it. We are systematically unable to deal with a major pandemic – just visit an ER on a good day. And on climate change, the lack of action in face of near-certain string of escalating disasters is deafening.

But luck only lasts so long.

Building Resilience

At some point, the technology surpasses our capacity to deal with issues. At some point, autopilots on airplanes become statistically far superior to humans in dealing with any situation thrown at them.

It is only when that happens – if even then – that we can forget to fly and let the machine take over; when we’re confident it will handle anything better than we ever could have.

We are not at that point yet.

And until we reach that point, we must maintain our capacity to fly – to fly that aircraft, to manage this planet, and to get ready for the turbulent times ahead we so clearly see are ahead of us.

We need to build resilience and backups.

We need to be ready to change and adapt.

We need to harness technology as it becomes available – and yes, drive it forward.

But we must not forget how to survive without it too soon.

Posted in General, Innovation | Tagged , , , , , , , | Leave a comment

Cutting Through the Fluff

Sometimes, the world of technology is not much better than reading a tabloid; a lot of wasted enthusiasm and energy, but zero substance. In a world where every vendor needs to be the “global leader”, every startup a “disruptor” and every product “revolutionary”, it’s useful to have a set of quick techniques for performing elementary reality checks.

Luckily, there are two exceedingly simple tools/techniques that I’ve found invaluable; from performing due diligence to dissecting news articles to debunking outrageous analyst forecasts, they provide a quick and easy “first-pass” filter for most claims, technology-related or not.

1) Ask Why

The Five Whys is one of many variations of root cause analysis which can be used for much more than just troubleshooting. Asking why is incredibly powerful, because it forces the other party to justify the claims made in a logical manner – which strikingly often appears to be impossible for them. Granted, sometimes it can also be hazardous to one’s health – in a typical corporate environment, people are often ready to assault you by the third why 😉

A simple “Why” can quickly uncover basic, fundamental problems with an offering or activity. And if you prepare to answer such questions yourself, you’ll be forced to develop a much more compelling, logical story that is better based on data.

2) Do The Math

This is something so many companies, journalists, analysts and others fail to do that it’s not funny. We’re not talking about complex climate model-type of math either, but really basic fundamental reality checks. It’s more about answering some simple questions about any of presented numbers:
– do they make sense?
– are they realistic?
– are they backed up by some evidence/data?

Let’s take Cisco’s “IOT is a $19 trillion market“-spiel as an example that still gets repeated, completely uncritically, ad nauseam. For that to happen, it would require the global GDP growth to double, starting last year. Unless, of course, they think IOT is a zero-sum game and all the “opportunity” is someone else’s loss – which would in turn mean a rebalancing to the tune of 25% of global GDP. Unsurprisingly, Cisco isn’t keen to talk about such ‘details’.

Of course, it’s not about outrageously big numbers either; many more mundane claims can be debunked by basic math, realistic adoption curve modeling etc.

First-pass filter rate…resembles Sturgeon’s Law

Frankly, I find it alternatingly sad, infuriating and depressing that the vast majority of … everything, really … fail even such elementary tests. It is another proof point for Sturgeon’s Law (“90% of everything is crap”).

Speaking of laws, there’s another law that I see someone falling prey to on a daily basis – Amara’s Law; “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” That’s something to think about before you pitch your next revolutionary overnight-world-domination idea.

But, but… I have to exaggerate!

This is a common retort to the above criticism; that one simply has to exaggerate and make ludicrously unrealistic claims to get any airtime these days. And there would be some truth to that – it is indeed difficult to be noticed by the media, VCs or whoever if you’re a realist. Extreme messages sell.

I might be naive in that way, but I don’t accept that one would somehow be essentially forced to lie. It reminds me of a quote from Jiddu Krishnamurti, which is a good place to wrap this rant up;

It is no measure of health to be well adjusted to a profoundly sick society.

Posted in Business | Tagged , , , | Leave a comment

Electric vehicles & utilities’ death spiral

Discussion on the potential energy utilities’ death spiral has moved from industry articles to mainstream media. It’s a fascinating topic that has the potential to rework one of the fundamental corner blocks of modern society, electricity.

Cycle started by solar PV

But what exactly is a death spiral in this context? For those unfamiliar with the term, a death spiral is when lower demand for electricity from the grid leads to higher prices (because the same infrastructure still needs to be maintained), further leading to lower demand, higher prices etc. Increasing prices will obviously hit the poor hardest, especially with the up-front capital requirements of solar PV (though some schemes are helping consumers with those).

In Australia, electricity demand has gone down for five consecutive years. There are many reasons for this (manufacturing recession, increased efficiency etc) but the most interesting and perhaps the most important one is increased distributed generation in the form of consumer solar PV panels – approximately 10% of the households now have a solar PV installation.

That amount of panels generates a significant amount of electricity, even if it is only when the sun shines. The aggregate impact is reduced grid demand – the distributors are however left with the same costs to cover for maintaining the grid, but with less revenue to cover it from. Their “natural” reaction to this is to raise network charges and lobby for policy changes like reducing feed-in-tariffs (FIT), both of which have successfully been done – see, e.g.Grattan Institute report Fair Pricing for Power on the former.

Enter storage

It seems inevitable that the cost of solar PV will continue to come down, as it has for decades. There is however another, even more important, part of the puzzle with also reducing costs: storage. While energy storage is still relatively clumsy, expensive and resource-intensive, this is slowly changing and household-scale batteries are already available, albeit at a high price.

When – not if – that price comes down sufficiently, it will become economically feasible for households to do something en masse that is today mostly a special circumstances or an ideological solution: disconnect from the electricity grid altogether. If households start disconnecting in significant numbers, the risk of a death spiral suddenly becomes much greater – and with rapid cost increases for grid-supplied electricity, the emotive rationale of disconnecting becomes increasingly compelling as well.

Will that happen? It remains to be seen, and there is another interesting development that may change the course of this development:

Electric cars

What do electric vehicles have to do with a death spiral? Perhaps ironically, they may delay it at first, and hasten it in long-term. Forbes reports on an EEI study that electric vehicles will save the utilities from the death spiral because of their high energy & power demands.

By proxy, however, success of electric vehicles may enable households to disconnect. Morgan Stanley highlights this angle in their report Solar Power & Energy Storage Policy Factors vs. Improving Economics where it is pointed out that the scale – Tesla’s Gigafactories in particular – will come to help:

“Given the relatively high cost of the power grid, we think that customers in parts of the US and Europe may seek to avoid utility grid fees by going “off-grid” through a combination of solar power and energy storage. [..] We believe there is not sufficient appreciation of the magnitude of energy storage cost reduction that Tesla has already achieved, nor of the further cost reduction magnitude that Tesla might be able to achieve. once the company has constructed its “Gigafactory,” targeted for completion later in the decade.“

Morgan Stanley is predicting battery costs will come down by more than 50% from today’s prices by 2020, with equal full system price reductions. Cheaper batteries of course mean not only cheaper electric cars, but cheaper household-scale energy storage. Barclays joins the sentiment and has downgraded the US utilities as it sees solar as viable competition, stating “We believe that solar + storage could reconfigure the organization and regulation of the electric power business over the coming decade”.

Tesla is already, through SolarCity, trialling stationary storage systems in up to 20kWh range; if all goes as planned, Tesla’s Gigafactories can in some years’ time feasibly produce battery packs at one fifth the price of currently offered solutions.

What do I think will happen?

So, interesting times. I can’t say where this will end, but I can say where I would like it to end: part of me, the resilience-seeking, decentralization/relocalization-driving part, would like to see the grid completely reconfigured, microgrids in place, with distributed generation commonplace and with a healthy dose of completely disconnected entities – with possibly the grid serving in reserve capacity for at least some time. On a system level, this would be much more resilient.

Transition to something so fundamentally different would, however, be extremely difficult and costly. If we wanted to maintain the current basic idea of centralized generation and the style of grid we have today, I would vote for the bulk of generation moving from coal to nuclear, with renewables playing a role as well.

Whatever the future brings, I hope we get rid of coal – it is by far the most environmentally disastrous and dangerous. In terms of deaths per TWh generated, coal is king at 161 deaths, followed by oil at 36, biofuels at 12, gas at 4 and hydro, solar, wind and nuclear all about 1 or less (data from here).

I’d welcome any thoughts or guesses on how this will play out and what the inevitable geographical differences will be.

Posted in Business, Energy, Environment | Tagged , , , , , , , , | Leave a comment

Why my new home won’t be “smart”

(And with this cross-post from LinkedIn, the blog returns from its break.. I tried LinkedIn as a publishing platform for a while, but with more than half of the posts there experiencing issues of not being visible or comments not showing up, I’m giving up on them and returning here.)

I’ve always considered myself as a technologist and a bit of a geek. Given that, I thought that if I ever build a house, I’ll have smart this and smart that, everything remotely controllable and all that. Now that I would have the opportunity to do that, I am surprising myself with my decisions on smart home gear: I am essentially getting none of it.

Why? Two main reasons:

1) Lack of net benefits.

2) Security and privacy issues.

Let’s tackle the security first; the more I have dug into security over the years, particularly in the “Internet of Things” domain, the less convinced I am that anyone, anywhere – exaggerating only slightly – knows what they are doing. It’s gotten to the point where I predict the whole IOT/IOE visions to result in a variety of security-related tears unless the priorities change quite fundamentally – and it’s even worse in the consumer space, where short product life cycles and a generally blasé attitude to security aren’t exactly helping.

What it comes down to is that I don’t trust the providers to keep things secure – because, to a large degree, they cannot. On that, this is a great introduction to why it’s next to impossible: Everything is Broken by Quinn Norton.

Loosely related to security is privacy. What I trust even less is that the providers of all that smart home gear would keep my data private and not abuse it. So I won’t even give them the data to begin with.

Another big problem are the net benefits, or lack of them. It’s not so much that all the smart home gear is useless, but it imposes additional actions on the users I don’t want to deal with. In other words, I don’t see significant net benefits from the available smart home stuff. The costs, not just monetary but in terms of time, outweigh the benefits. One of the main issues here is that much of it is designed to give the users more control over something. Very few products are designed to work in the background, or do it sufficiently well for me to trust them to do their things how I want it done. They’re just not smart enough yet.

Take the state-of-the-art smart thermostat, Nest. It’s not designed for you to constantly play with, it’s designed to learn your habits and self-adapt, work in the background. Sounds good, right? It does, but I don’t trust it enough to try it. The reasons? First, most of the time I won’t need any heating or cooling in the house. Sometimes I want to cool it when the temperature hits +30C (e.g. when coming in from a run), but at other times I’m happy to let it hit +35C. In winter, I may want to turn on heating at +15C, but when the daytime high will be over +30C, I’d prefer to have a cold night and morning. I obviously haven’t lived with it, but would Nest be smart enough to do all that and more? I doubt it.

There are a host of other issues as well – such as the poor availability of some products in Australia (vs USA in particular) and lack of faith in continued support in what is a fast-developing field and/or small players. Much of smart home-stuff is integrated to the building to a greater or lesser extent, and I don’t want to have to change it every two years when someone goes out of business. A bit of a chicken and egg-problem, I know.

So my new home will be a Luddite. It’ll be environmentally sustainable, comfortable all year, energy-efficient, practical and above all liveable.

But it will not be “smart”.

I’ve done a fair bit of research into what’s available and reached the above conclusions, but I’m always happy to be convinced otherwise and hear suggestions if you think something is a “must have”.

Posted in Building | Tagged , , , , , , | Leave a comment