“Intel” is the category under which The Bob Zadek Show’s episodes on artificial intelligence and on constitutional design are grouped. The excerpts span two episodes: a 2017 conversation with MIT professor Luis Perez-Breva on artificial intelligence myths, and a 2021 conversation with Timothy Sandefur on what the show frames as the libertarian constitution. Neither excerpt uses the word “Intel” as a corporate or organizational name; the material concerns intelligence as a concept and the institutional arrangements that govern it.
Artificial intelligence as a set of tools
Bob Zadek opened the 2017 episode by naming the fears he intended to dispel: that robots would take over the world, that the Terminator would control everyone, that machines would take over jobs and leave humans servicing them. He introduced Luis Perez-Breva as a professor at MIT, a serial innovator working at the forefront of artificial intelligence, and said Perez-Breva would dispel the myths of the field Luis Perez-Breva on Artificial Intelligence Myths (2017).
Perez-Breva’s central claim was that artificial intelligence has become an enormous field containing many separate tools — data science, big data, machine learning, logic, computers that play games — each the product of enormous work, each narrow, and none working for absolutely everything. The robots of modern movies, which appear able to use all those tools seamlessly, do not exist; they exist in the movies. He contrasted the present era of evil robots in film with the 1970s, when audiences thought of robots as good, citing Data in Star Trek and R2-D2 in Star Wars, and said the fear has to do with the movies rather than the field.
Asked whether a machine processing data to make decisions functions like a human, Perez-Breva answered that it functions very differently. He recalled that shortly after Turing’s work people thought beating a human at chess would constitute intelligence, and said the field has since discovered how much more intelligent humans are and how much more path remains. The objective, he said, is not a machine that reasons like a human, because humans already exist, but a machine that helps people do things they cannot. He offered a concrete contrast: a human needs about two examples of anything to figure out roughly what it is, while the machines being built today require an enormous number of examples to address a very simple task and then cannot explain back what they did. They are good at spotting patterns, he said, but nowhere close to emulating humans, and how humans do it is not fully understood — a question he placed in neuroscience rather than AI.
AI winters and the failure of replacement
Perez-Breva described recurring “artificial intelligence winters”: every 10 years since the 1950s, AI has been claimed to have failed, interest has fallen away, and it has returned a decade later. Each failure, he said, came from the assumption that the goal was to replace humans, and each attempt to do so was a miserable failure that backfired on both researchers and the companies that tried it. He added that he cannot speak for every AI researcher, but as far as he can tell none of his colleagues wants to replace human reasoning; they want better reasoning machines that help understand human intelligence without replacing it.
Bob Zadek drew the implication that the field does not aspire to build better humans than humans, but to take repetitive jobs and do them so the human becomes more productive. Perez-Breva agreed and went further, tying the winters to that same mistaken premise.
What is artificial about it
Returning after a break, Bob Zadek asked what is artificial about artificial intelligence and what the intelligence component is. Perez-Breva said the artificial part is getting computers, robots, mechanics and systems to do work beyond merely programmed movement, and that what exists now is an enormous set of tools — machine learning, which extracts patterns from large pools of data to build models for more informed decisions, big data, data analytics. He noted that listeners have been using these tools for 20 years without knowing it.
Intelligence, he said, has been defined and redefined for decades, each attempt revealing more complexity. His working description was conversational: two people can talk without consulting Wikipedia for everything they say, derive a theme, and walk away smarter, able to accomplish more because of the conversation. The aspiration for AI, he said, is a companion that on very specific topics can engage in conversation and make a person smarter. Today’s examples are Google and Netflix, where a conversation of sorts develops but the system is hardly intelligent. The change will look incremental, he said, but will change life because of how much further a person can go.
The motion detector analogy
Bob Zadek proposed a motion detector as a simple model: it receives an external stimulus, processes the information, and warns the user. He asked whether artificial intelligence is that a billion times over. Perez-Breva answered that it relates zero — the motion detector is a system programmed to perform a single function, as dumb as it comes, a simple line of code, implementable mechanically with sensors before computers existed. He said it is the root of today’s confusion: if you think that is the beginning of artificial intelligence, you see AI as much closer and much more threatening than it is. He then flipped the example: a motion sensor combined with the ability to close a door is all anyone needs to lock a person out of their home, so evil needs far less than AI, and AI is not evil.
Bob Zadek pressed further, describing intelligence as receiving external information and reacting. Perez-Breva replied that this describes computing: every computer takes information in and is programmed to produce an outcome. The incremental factor in AI, he said, is that it is programmed not for one task but to learn, which means engaging in conversation. He defined intelligence as taking in what one sees and hears, adding it to one’s thoughts, and producing a response the other party can bring back — the moment in a class when a student recognizes a new idea in what a professor is saying and engages with it afterward. Without the conversation aspect, he said, it is just a computer.
Intelligence as a different way of thinking
Bob Zadek suggested that the word intelligence in the phrase is almost a perversion, since a machine will never do anything very intelligent but will do things well, efficiently and with a low error rate. Perez-Breva returned to Turing, one of the founders of the field, who is credited with saying the best way to think of a computer is not as something competing with humans in intelligence but as something with a different way of thinking: it can take enormous amounts of data at once and look at many dimensions at once, while humans tend to be limited to three. The best use of computers, he said, is to make them more capable in that realm so humans can be more capable in theirs. Bringing computers into the human domain would require limiting them to three dimensions; nature is smart, evolution has produced efficient machines, and building computers that do the same things humans do is not obviously a good objective, especially since humans are far less expensive to build and can be trained in myriad ways.
Bob Zadek closed the hour by framing the question as whether AI is a threat or a boon, then went to a break.
Constitutional repair and intellectual property
The 2021 episode with Timothy Sandefur concerns a project of revising the Constitution. Bob Zadek asked whether the project corrected mistakes the Founders made or adjusted principles to changing circumstances. Sandefur said the initial impulse was to hand back the existing Constitution and call it libertarian, but that a number of things needed remedy, and that existing amendments make the document in some ways more and in some ways less libertarian than it originally was Timothy Sandefur on The Libertarian Constitution (2021).
The project took the existing Constitution and incorporated the existing amendments, leaving out the income tax, which Sandefur said libertarians strongly oppose. The 13th Amendment, prohibiting slavery, was written into the body of the document rather than left at the end. After weaving in the amendments, the drafters turned to what to tinker with. On intellectual property, Sandefur said many libertarians, including himself, have real problems with the idea of intellectual property at all, and that the US Supreme Court has said Congress can renew the length of a copyright indefinitely, which he called a real problem. The proposed constitution therefore allows copyrights and patents but limits them to 28 years — the period patents lasted at the time of the Founding Fathers — so that Congress cannot keep renewing copyrights whenever Mickey Mouse is threatened with becoming public domain.
Bob Zadek asked what is wrong with owning a creation for a long time, framing the question as a surrogate for the libertarian approach to property rights generally. Sandefur answered that the libertarian critique is that intellectual property is not really property but a government-created monopoly making imitation or copying illegal, which initiates force against someone who has not initiated force, violating a fundamental principle of libertarian political theory. Copyright and patent, he conceded, reward and incentivize creation, and the compromise is a limited period of exclusive use as a subsidy to creators, after which no one can indefinitely prohibit others from profiting. His example was the 1980s Men at Work song “Down Under,” which he said was found to be a violation of copyright because it uses a few notes from “Kookaburra Sits in the Old Gum Tree” in the background, requiring the band to pay a large sum to the rights holders — which he called ridiculous and a real injustice.
Across episodes
The two episodes do not argue the same question. The 2017 episode treats artificial intelligence as a misunderstood tool whose danger is overstated; the 2021 episode treats the Constitution as a document requiring libertarian repair, with intellectual property as one site of that repair. The excerpts show no development between them on a shared topic, and no speaker appears in both.
What the sources do not cover
The excerpts do not state what became of the constitutional project Sandefur describes, nor how the proposed 28-year term would be enforced. They do not give the outcome of the “Down Under” litigation beyond Sandefur’s account, and they do not identify the court or the date. The 2017 episode ends mid-sentence at a section break, and the 2021 excerpt breaks off before the discussion of the Commerce Clause that its heading announces.