Exponential priorities

The great attraction of fossil fuels is their huge energy density.  Burning a kilogram of oil can produce 40 to 50 megajoules of energy to propel a car, truck or aircraft.  By comparison, the best batteries can barely produce one megajoule of energy for each kilogram before a recharge.  This wouldn’t matter if battery energy density was improving in a way that was analogous to Moore’s Law for computing, but it simply isn’t.

The car makers are battling this out at the front line with one group, largely led out of the US, pushing for battery powered electric cars. The other, largely led out of Japan, are arguing for alternatives including a hydrogen economy.  Today, hydrogen in a fuel cell can deliver over 10 times the energy density of the best batteries and is still not even close to its theoretical limit of 120 megajoules per kilogram, creating an exponential opportunity.

The debate matters because transport requires infrastructure.  If there aren’t readily available places to refuel or recharge your vehicle, people won’t take-up either technology.  It’s unlikely that any country can leave it to the market to get to scale in time, meaning we need to decide as a society on the technology to back and scale the supporting infrastructure.

One of the great privileges of my role is that I get to talk to a lot of business leaders, particularly across Asia Pacific.  There are some consistent themes that they raise, each of which is exponential in impact, risk and opportunity.  The three at the top of most lists are climate change, artificial intelligence (AI) and cybersecurity.  Each of these priorities are not only today’s media headlines but they are also exponential disruptions in search of exponential solutions.

We are desperately looking for an alternative to fossil fuels because, by any definition, runaway climate change is an exponential problem including feedback loops ranging from ocean temperatures to melting permafrost.  Trying to solve such a problem with inherently linear tools such as chemical batteries isn’t a fair fight.

While climate change is existential, AI makes it to near the top of most disruption lists today.  I was recently asked whether AI was another fad.  After all, the technology sector hypes a new solution at least every other year.

Last year it was all about blockchain and the metaverse, neither is dead but nor are they top of mind as priorities.  After all, whose metaverse real estate has gone up in value in the last twelve months?  I’ve argued before that blockchain, and the crypto industry it supports, is inherently limited whereas AI is an exponential technology with commensurate impacts and opportunities.

The rate of growth of large language models (LLMs) underpinning generative AI illustrates this as well as just about anything else.  The capability of current generation of AI is best measured in terms of “tokens” of the prompt and response.  LLMs simplify natural language into tokens, with 100 tokens representing roughly 75 words.  In 2018 OpenAI released GPT-1 with limit of about 500 tokens while each subsequent release has roughly doubled this capability.  Similarly, the parameters that underpin the learning of these models have reportedly grown from roughly 100 million to one trillion during the same time.

While it is relatively easy to grasp the scale of the opportunity, most responses have to-date been limited to draft regulation and fairly simple pilot services.  Regulation and legislation is seldom able to scale to meet any exponential challenge and pilots need to quickly move to the mainstream if society is going to learn quickly enough to keep up with emerging threats from this technology.

Similarly, cybersecurity is an exponential challenge in that the number and sophistication of attacks is escalated by the network effect of increasingly sophisticated and connected bad actors.  Some estimates put the global impact of cybercrime at more than $10 trillion by 2025 up from $3 trillion in 2015.  Worse, new technologies from AI through to cryptocurrencies are creating extensive new opportunities for new, innovative and connected attacks on business, governments and individuals.

Like AI, the threat of cybersecurity is being met with emerging regulation, ineffective international agreements and technology that responds to each individual threat as it emerges.  However, it is entirely possible that the two disruptions of AI and cybersecurity can be combined putting exponential tools in the hands of those that seek to protect rather than disrupt society for the first time.

Climate change, AI and cybersecurity are exponential issues but the responses of business, government and individuals for each, including regulatory, technical, engineering, commercial and social, are still far too linear.  Until we close the gap these issues must stay at the top of any list of business threats.

Leave a Reply

Your email address will not be published. Required fields are marked *