Books of Interest
Website: chetyarbrough.blog
The Infinity Machine (Demis Hassabis, DeepMind, and the Quest for Superintelligence)
Author: Sebastian Mallaby
Narration by: Vidish Athavale

Sebastian Mallaby (English journalist and author, senior fellow for international economics at the Council on Foreign Relations, two-time Pulitzer Prize finalist.
Mallaby has written a biography of Demis Hassabis, the co-founder and CEO of Google DeepMind. This is a somewhat frightening exploration of AGI (artificial general intelligence). The author argues AGI is being grown at a faster rate than it can be understood or managed by humans. Though science may wish to develop AGI to aid society, it may be creating a system that will end it.

Sir Demis Hassabis (British A.I. researcher, neuroscientist, video game designer, and entrepreneur.)
Mallaby reflects on the early childhood of Demis Hassabis who becomes a chess master at age 13, a video-game designer at 17, and who founded his own company in his early twenties. His life experience leads him to study cognitive neuroscience at University College London. He explores how humans construct mental models to think and act in the world. With that view of the building blocks of human thought, he founds DeepMind with help of Legg and Suleyman. Their thought is that in solving the mystery of the way human intelligence works; they could create a computer program to more rapidly and successfully improve human society.


The two other principals in the creation of Deep Mind, Shane Legg and Mustafa Suleyman, have different roles in the company. Legg is a computer scientist who focuses on a framework for Deep Mind that would make Deep Mind more intelligent with a recognition of human reliance and its potential risk of minimizing the importance of human judgement. Mustafa Suleyman focuses on fundraising for the Deep Mind organization and the operational function of the company. The ultimate acquisition of Deep Mind by Google resolves fundraising concerns of Deep Mind and assures its future development.
Mallaby believes the speed of AGI development exceeds the ability of society to either understand or govern itself without the influence of AGI. The author does not believe most of its developers or the public clearly understand the risks of AGI development and its influence on society. The author argues AGI concentrates power in ways governments cannot fully comprehend and, without that understanding, AGI may lead society beyond the fundamental goal of human survival. The consequences of too much reliance on AGI may escalate conflicts in the world.
Shane Legg is alleged to believe “…human extinction is not necessarily the worst possible outcome…” for the advancement of AGI because of an evolutionary melding of human and machine intelligence.

The question becomes whether human beings can program AGI with a preeminent goal of human life’s value and preservation. Mallaby argues at the heart of the issue is human fallibility. This is a subtle argument because it is saying humans are not capable of encoding human values in A.I. to assure and enforce human preservation. He implies human programmers are too conflicted over how to program A.G.I. to guarantee human preservation. This is the author’s “Catch 22”; i.e. a no-win paradox because of circular logic. Shane Legg argues there is no paradox and that the future may mean the melding of human beings with AGI superintelligence.

As one ponders Legg’s belief, one is reminded of Karl Marx who wrote “History repeats itself first as tragedy, second as farce.”
Malaby compares AGI to the nuclear bomb. Proven by the nuclear bomb E = mc2 has a tremendous upside. It is the proof of Einstein’s energy formula which is the basis for belief in an inexhaustible source of energy. However, Mallaby implies AGI is like Einstein’s discovery of the equivalence of mass and energy. Like nuclear power foretells an inexhaustible source of energy it can be a destroyer of worlds. AGI foretells the possibility of a revolution in human knowledge, but with the accessibility of the internet, AGI carries the danger of human misuse and societal conflagration. Both AGI and nuclear power are only constrained by human judgement. If judgement of AGI seems to exceed human judgement, the path of AGI may mean the end of humanity because human judgement is viewed as inferior. The jury seems out on use of nuclear weapons and the same seems true for AGI.

The second atomic bomb test at Bikini Atoll on July 25, 1946. The Marshall Islands, where Bikini is located, is suing the U.S. for what it calls a violation of the Nuclear Non-Proliferation Treaty. There seems a correlation with the advance of AGI. Who is at risk, if judgement is left in the hands of AGI rather than human beings?
When AGI reaches the test of nuclear weapons, the only constraint seems the nature of human beings and their judgement. The “Catch 22” of AGI is the potential for knowledge beyond the comprehension of the brightest human beings. Many will choose to forgo human judgement and action because of the presumed value of knowledge offered by AGI. That is the threat being explained in “The Infinity Machine”.






































































