By 2040, AI will be a zillion times more powerful than it is now following Moore’s Law. Moore’s Law addresses a doubling up of computer processing capability every two years. Moore makes this observation noticing that the number of transistors on an integrated circuit doubles about every two years. The law is credited to Gordon Moore, the former CEO and co-founder of Intel Corporation.

While the ability to scale a single chip may be constrained by materials science, the explosion of AI has increased the pace of innovation to maintain this exponential growth. The new Nvidia AI chip, the H200, will roll out next year with Amazon Web Services, Google Cloud, Microsoft Azure, CoreWeave, Lambda, Vultr, and Oracle Cloud.

The H200 delivers greater high-bandwidth memory. This parameter of the design defines how much data the H200 can process quickly. The H200 has 141-gigabytes of high-bandwidth memory. The supplier of the memory on the H200 remains undisclosed.

In the coming months, AI will create new supply chains, deliver new drug discoveries, change conflict and combat, and accelerate progress in farming and climate mitigation. It is already the case that African glaciers distributed along the equator in Uganda, Kenya, Tanzania, and the Democratic Republic of Congo are under threat from the climate emergency. The disappearance of these African glaciers will result in the critical loss of access to drinking water. The Kilimanjaro threat is also associated with deforestation.

The marine heatwave this June increased UK water temperatures by about 3-4C. Along the west coasts of the UK, especially in Cornwall and Wales, warmer waters invited the Portuguese man o’ war, Lion’s mane, barrel and crystal jellyfish into a fresh habitat.

This is a new tropicalisation of oceans and seas. No one can state if this is a longer-term natural trend, or a punctuated equilibrium event, linked to the recent ocean heat wave. Perhaps the unravelling of these changes pivots on data from ocean surveillance that may use non-human forms of logic that AIs bring.

In the Caribbean, German U-boats operating in formations known as “wolfpacks” sunk many freighters and oil tankers, during the period 1941-1945. At 05.32 hours on 19 February 1942, U-161 fired two torpedoes into the harbour of Port of Spain. The torpedoes struck the Mokihana and the British Consul lying at anchor in the port. Both ships sank.

Military history correlates a nation’s political influence with its strategic capabilities and military power through implicit threats to inflict damage on infrastructure. But stability based on a calculus of power is not self-maintaining nor is it motionless. It requires recalibration either through countervailing force or the assimilation of a new reality. When the calculations of relative power become uncertain, the risk of conflict peaks.

In the Age of AI, these recalibrations have reached a new level of abstraction. Cyber weapons that have civilian and military capabilities have blurred the status of many devices. As AI transformative capabilities grow and bloom in the absence of verifiable global restraints, meshworks will strive to achieve superior positions. Moreover, AI’s fundamentals and key advances remain largely in the public domain.

This may account for why old methods of intelligence gathering may encounter puzzling voids and silences. In situations where AI is under human control, we will discover its permeability. Users will adapt underlying algorithms for very different aims. But the most unpredictable and revolutionary effects will occur at the interface of Human Thinking and Artificial Intelligence.

Enemy combatants may use AI to plot or target – or even assist dynamically during an assault. But the interaction will become newly strange as it will entail interpretation of, and communication with, unfamiliar non-human forms of logic.

The paradigm shifts to AI-assisted defence systems and dynamic combat operations will involve a measure of reliance on and delegation to – an intelligence of considerable analytic capability. That battlefield will be operating inside a fundamentally different experiential paradigm.

For now, the risks involved in such situations remain poorly understood. If we are to avoid all errors, then AIs must be under human control if we aspire to guarantee moral agency and accountability.

Should AI be cabined in a free society? Or, should it blossom on its own terms? Or perhaps we may need to partner with it – to govern the future. Truth buried by lies can damage, distort, divide and incite. Freedom is not free. It is the opportunity to accomplish what we value. Capability is the set of vectors that give us the freedom to lead one type of life or another and to choose from possible livings. How then do we plan a new beginning?

The aim of development is thus the expansion and promotion of valuable capabilities, and the freedom to use our reasoned agency to live the life we value. Shortly, we will have “AI twins” of ourselves that function as co-creators of digital content.

But how will humans account for what “AI twins” may write or utter publicly? How will the courts distinguish human from machine expression? What is certain is that our children will not be walking into the past we lived, and the futures we invent will be far from the present.