.jpg)
Max
Bennett’s A Brief History of Intelligence explains how
intelligence evolved through five breakthroughs, and why today’s AI may be
repeating that same journey in digital form.
Introduction: Intelligence as an Evolutionary
Story
Max
Bennett’s A Brief History of Intelligence: Evolution, AI, and the Five
Breakthroughs That Made Our Brains is not just a book about brains. It
is a theory of how intelligence itself emerged, layer by layer, over hundreds
of millions of years.
Bennett
argues that intelligence is not one invention but a sequence of five major
breakthroughs. Each one solved a different problem, and each one became the
foundation for the next.
That
framework matters today because AI is advancing in a similarly layered way.
Bloomberg Intelligence has projected that generative AI could become a $1.3
trillion market by 2032, showing that this is not only a scientific story but
an economic one as well.
Breakthrough
One: Steering
The first
breakthrough in Bennett’s model is steering. This is the ability of
early organisms to move toward good outcomes and away from bad ones, creating
the most basic form of purposeful behavior.
Steering
is primitive, but it is powerful. It gives life direction before it has memory,
planning, or language, and it explains why basic response systems matter so
much in evolution.
In AI, the
parallel is simple: systems still need objectives, reward signals, and
optimization rules. Without steering, a machine can process information, but it
cannot learn what matters.
Bloomberg
Intelligence chart idea: A
stacked bar chart showing how AI spending is distributed across hardware,
software, and services.
Breakthrough
Two: Reinforcing
The second
breakthrough is reinforcing, which means learning from
consequences. Bennett traces this to vertebrates, where organisms began
repeating behaviors that led to positive results and avoiding those that led to
harm.
This stage
added memory, adaptation, and emotion-like responses to intelligence. It helped
create curiosity, fear, anticipation, disappointment, and relief — all of which
shape behavior in much richer ways than simple steering alone
Modern AI
reflects this breakthrough through reinforcement learning. Many systems improve
by testing actions, measuring feedback, and adjusting behavior over time, which
is one of the clearest biological parallels in the book.
Bloomberg
Intelligence chart idea: A
performance chart of an AI index or AI-related market basket to show how
reinforcement systems are rewarded in both biology and markets.
Breakthrough
Three: Simulating
The third
breakthrough is simulating, and it may be the most important bridge
to modern AI. Bennett links this to the rise of the neocortex in mammals, which
allowed brains to imagine possible futures before acting in the real world.
Simulation
made intelligence more efficient. Instead of learning everything through direct
trial and error, mammals could rehearse options internally, compare outcomes,
and choose better strategies.
That is
exactly why this chapter feels so relevant to AI. Today’s models increasingly
generate internal reasoning steps, candidate responses, and multi-stage outputs
that resemble digital simulation.
Bloomberg
Intelligence chart idea: A
trend chart showing the growth of enterprise AI use cases that depend on
reasoning, planning, and workflow simulation.
Breakthrough
Four: Mentalizing
The fourth
breakthrough is mentalizing, or understanding the minds of others.
Bennett connects this to primates, whose larger and more complex brains made it
possible to infer beliefs, intentions, and goals in social settings.
Mentalizing
changed intelligence from private problem-solving into social intelligence. It
enabled cooperation, competition, trust, deception, teaching, and politics —
all of which depend on modeling other minds.
This is
one reason human intelligence is so unusual. Much of what we call smart
behavior is actually social prediction, not just individual
Bloomberg Intelligence chart idea: A chart showing projected AI agent adoption in the enterprise software market, since agents are increasingly expected to infer user intent and act on behalf of humans.
Fifth Breakthrough:
Speech
The fifth
breakthrough is speech, which Bennett treats as the key to
cumulative culture. Language allowed knowledge to pass from person to person
and from generation to generation, making intelligence collective instead of
isolated.
Speech
changed everything. It allowed people to learn without direct experience,
coordinate in groups, and build institutions that stored knowledge outside the
brain.
In AI,
language is the gateway that made advanced systems usable at scale. Large
language models turned machine intelligence into something people could
interact with naturally, and that helped drive explosive adoption.
Bloomberg
Intelligence chart idea: A
market curve showing generative AI’s projected rise to $1.3 trillion by 2032,
tied to language-driven adoption.
What
the Book Says About AI
Bennett’s
deeper argument is that AI is not separate from biology in spirit. It is a new
medium for repeating the same intelligence-building pattern that evolution used
in brains.
That does
not mean today’s models are human-like in a complete sense. It means they
already reflect some of the same architectural ideas: steering, reinforcement,
simulation, and language-based coordination.
The
important lesson is that intelligence evolves in layers. AI is likely to become
more capable not through one giant leap, but through the gradual addition of
memory, agency, social reasoning, and better world models.
Bloomberg
Intelligence chart idea: A
layered ecosystem chart showing chips, infrastructure, software, and agents as
distinct stages of AI commercialization.
Why
Bloomberg’s View Matters
Bloomberg
Intelligence’s AI research helps place Bennett’s ideas in a market context. Its
projections show that AI is expanding across the tech stack, from hardware
demand to software spending and enterprise workflow automation.
That
matters because the book is about more than history. It helps explain why the
market is reorganizing around intelligence, and why every improvement in AI
capability can create a new commercial layer.
Bloomberg
has also highlighted AI agents as a major force in enterprise software,
suggesting that the next phase of AI will not just answer questions but take
actions. That aligns closely with Bennett’s story about intelligence becoming
more active, social, and cumulative over time.
Bloomberg
Intelligence chart idea: A
comparison chart of AI spending growth versus AI agent adoption in enterprise
software.
The Big Takeaway
The
biggest value of Bennett’s book is that it makes intelligence legible. Instead
of treating the brain as a mystery, he breaks it into evolutionary steps that
can be studied, compared, and applied to AI.
That gives
readers a better way to think about progress. A system that can speak may still
be weak at planning, and a system that can plan may still be weak at
understanding other minds.
This is
why the book feels so useful in the AI era. It gives a vocabulary for
identifying what kind of intelligence a system has, not just whether it appears
smart.
Bloomberg
Intelligence chart idea: A
maturity chart showing the move from basic AI tools to agents and autonomous
workflows.
Conclusion:
The Next Breakthrough
A Brief
History of Intelligence suggests
that the story of intelligence is still unfolding. The five breakthroughs
explain how brains got here, but they also hint that future AI may develop a
sixth major leap of its own.
That
possibility is what makes the book exciting. It does not just explain
evolution; it helps readers see where digital intelligence may be headed next.
Bloomberg
Intelligence’s forecasts reinforce the same message from the market side: AI is
growing fast, spreading across industries, and becoming a central economic
force. In other words, the next chapter of intelligence is already being
written.
Comments