Interview with Inventor of Neural Nets Warren McCulloch, neurologist who helped start it way back.
What this covers
About life and how neural nets can replace humans when the sun burns out.
Source description (no synthesized summary yet).
McCulloch argues that the human nervous system operates as a distributed, parallel, fault-tolerant computing machine fundamentally different from sequential computers, and that machines could eventually replicate human cognitive functions including emotion and purpose if the underlying logical mechanisms can be formally specified.
- The brain uses anastomotic (interconnected) parallel processing where information from all sources mingles before output, unlike sequential machines where errors propagate
- Biological systems are designed for degradation—neurons die constantly but the system continues functioning, requiring fundamentally different architecture than conventional computers
- Emotion and purpose are mechanistic properties that could theoretically be replicated in machines if expressed in finite, unambiguous logical form
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
The human nervous system is designed to tolerate continuous neuronal death—thousands of neurons die per day throughout adult life—and the system must be architected so that dying neurons degrade gracefully without catastrophic system failure.
“Neurons die at the order of thousands of neurons per day. We're built to run well to the 16th year of age, roughly. From that time on, you begin to be able to count the holes where there were cells and there are no more.”
McCulloch believes it is possible that machines created by humans could eventually survive humanity and continue into a distant future when humans are no longer present.
“Will we one day in the distant future, perhaps even create beings superior to ourselves? Beings who will survive us on this planet? Dr. McCulloch thinks it possible.”
McCulloch argues that emotion and affection are mechanizable—if he can formally specify the mechanism of emotion in finite, unambiguous logical terms, then it can be implemented in a machine, and he sees no logical reason why machines couldn't eventually develop emotional capacities equivalent to humans.
“Well, I think I could set it up for you because I am certain that if I do it, there is a mechanism that can do it. If I can manage to state that in a finite and unambiguous manner, then uh I think it can be done. And I see no reason why we can't develop a logical relation sometime to come.”
In ordinary sequential machines, errors propagate inevitably through the entire computation once a mistake occurs, whereas the nervous system uses parallel processing across multiple channels that compare results before proceeding, preventing error propagation.
“In an ordinary machine, operations are carried out sequentially. And a mistake entails mistakes which follow inevitably from it all the way through the computation.”
McCulloch's single guiding scientific question throughout his entire career has been: 'What is a number that a man may know it and a man that he may know a number?'—essentially asking how cognition and numerical abstraction relate to each other.
“I have never had but one question in the whole of my scientific life. I've only wanted to know what is a number that a man may know it and a man that he may know a number.”
The brain uses an 'anastomotic' architecture where information from all sources mixes together before the final output, similar to how tributary streams merge in river systems like the Danube, Nile, or Suwannee River, creating integrated signals rather than parallel but separate channels.
“They're braided with an intermixture so that the information coming out at any one point in the mouth is combined from the information coming in from all of the sources of that river.”
McCulloch acknowledges he has made progress on understanding the first half of his question (what a number is) but remains uncertain about the second half (how a man may know a number), and has instead studied simplified biological systems like the frog's eye to approximate answers about neural coding.
“I think I am fairly clear as to the first half of it. But as the second half, a man that he may know a number, mhm, not yet. I've had to content myself with what a frog's eye tells a frog's brain.”
McCulloch argues that if machines do outlive humanity, they would continue in the same general direction that humanity would have continued, effectively 'standing on the shoulders' of human achievement and extending human purpose rather than abandoning it.
“In so far as machines might survive man, they might. They would only carry on in a sense the same direction same general direction that man would have carried on if he could have. Am I clear? So to speak, the machines would be standing on our shoulders.”
McCulloch and his colleagues believe the human brain operates as a machine with principles fundamentally different from conventional computing machines, specifically in how it processes information in parallel rather than sequentially.
“Dr. McCulloch and his colleagues believe they are beginning to understand how the higher nervous system, a man's brain, might work as a machine. A machine basically different in principle from anything we can now build.”
McCulloch came into neuroscience and mathematics through theology, specifically because he recognized that mathematical and logical principles underlie divine ideas, making mathematics a more fundamental expression of theological concepts than textual theology.
“I got seduced by mathematics. Surely because there was more fun in it, and because if you know theology at all well, you'll realize that the ideas in the mind of God are mathematics and logic.”
Interviewer challenges McCulloch by asserting that machines will never feel toward their creations the way humans feel toward their grandchildren, suggesting an unbridgeable gap between machine and human emotion.
“I can't see that a machine will ever feel toward any of its creatures the way you feel about them. That's the one thing that I can't get.”
The human brain is an organ of such extraordinary complexity that even optimistic neuroscientists believe it will take centuries to fully understand all its mysteries.
“An organ so complex that even the most optimistic believe that it will be centuries before we unravel all its mysteries.”
McCulloch rejects the premise that machines, no matter how complicated or clever, can be equated with human consciousness or purpose; machines lack something essential about human purpose that cannot be mechanized.
“Somehow the purpose is that a machine, no matter how complicated and how clever, could have, to my mind, isn't equatable.”
McCulloch's theory of neural computation accounts for and explains the continuous degeneration of neurons throughout adult life while maintaining overall system function.
“The theory that we're working on takes care of all of these things. That's the important point about it.”
McCulloch acknowledges that his mechanistic theory is currently defective and incomplete in its ability to formally specify emotion and consciousness.
“Granted, it's defective at the present time.”