Advanced Analytics and

The fundamental premise of Kurzweil's view is based on Moore's law, expressed over 50 years ago. This suggests that the number of transistors doubles every two years, leading to exponential growth. AI researcher Ray Solomonoff, who invented the concept of algorithmic probability, applied the same rule to self‐learning machines (which double their speed every two years), and as a result he suggests that singularity will occur with a definite period.

Worried? A singularity symposium already exists, in the form of an informal community, which takes part in discussion ‘about technology, AI and exponential growth’. Students of the subject should be careful not to mix prediction or speculation with science fiction, or to accept only one viewpoint as the definitive premonition of the future, as it may be wrong in terms of both nature and time frame. The history of technology is already sprinkled with so‐called laws – some of which are more realistic than others.

Some of Arthur C. Clarke's favourite laws include:

When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.

The only way of discovering the limits of the possible is to venture a little way past them into the impossible.

Any sufficiently advanced technology is indistinguishable from magic.

For every expert, there is an equal and opposite expert.

Despite Kurzweil's reassurances that there is no likelihood of another AI winter, it's worth spending a moment to reflect on the causes of that period of stagnation and to consider whether they might occur again. If so, what mitigating action might be available to us?

The causes of an AI winter have been variously suggested as being due to the following factors:

Institutional factors. Major institutions tend not to collaborate as budgets get tighter, resulting in less knowledge transfer and a reduced ability to leverage shared development.

Economic factors. During a downturn in the economy, programmes that are viewed as being speculative and without hard benefit are at increased risk when cuts need to be made.

Empty research pipeline. Unless organisations and institutions see tangible research output from a programme, it is at risk of being shelved.

Failure of existing systems to adapt. Existing technology becomes redundant in the face of new systems and processes, and users of existing technology resist moving from tried‐and‐tested systems to new models.