Ask A Genius 1577: Mel Brooks, Meta-Primes, and the Future of AI
Author(s): Rick Rosner and Scott Douglas Jacobsen
Publication (Outlet/Website): Ask A Genius
Publication Date (yyyy/mm/dd): 2025/11/28
How does Rick Rosner tie Mel Brooks, television’s evolution, AI’s future, and meta-primes into one Thanksgiving meditation?
In this Thanksgiving conversation, Rick Rosner talks with Scott Douglas Jacobsen about the enduring genius of Mel Brooks, from Young Frankenstein to Get Smart, and the changing sophistication of television from Hill Street Blues to today’s streaming era. Rosner laments no longer working for Kimmel, where legends like Norman Lear once appeared, and reflects on how creative legacies still shape culture. He riffs on AI’s multimodal future, humanoid robots, and the risks of systems with agency. He revisits his “meta-primes” idea on twin primes and information in the number line, and recalls favourite reading like Neal Stephenson’s The Diamond Age.
Rick Rosner: I am not thankful for much. It’s thanksgiving in America. If I list everything I’m not grateful for, then I’mjust complaining about the same things I usually complain about. I always do that. Obviously, I’m not thankful for our political situation. So let’s move on to something else.
Scott Douglas Jacobsen: What’s your opinion on Mel Brooks?
Rosner: I’m glad he’s still alive. He’s in his late nineties now—born in 1926 and still creating. He’s an intelligent, funny man.
Since it’s Thanksgiving, I’m thankful for him and for the fact that he managed to get so much made. Some of his works are among the greatest comedies of the past century, particularly Young Frankenstein. Get Smart, which he created with Buck Henry, was an excellent idea. Very few shows had spoofed the James Bond–style spy genre directly on television. He’s had a lot of great ideas.
And his batting average has been very high. It is still high. He’s still with us.
Some of his work drags a bit—The Producers and The Twelve Chairs—but that may be due more to the era and the format in which they were made than to him. Anyway, Rotten Tomatoes…
It makes me sad that I’m no longer part of Kimmel because Jimmy gets to work with, interview, and meet some of the greatest creators in entertainment. He was working with Norman Lear—sorry, I just woke up—who lived to 101. Not Neil Simon, but the man who created All in the Family and The Jeffersons.
Jacobsen: What do you think the most lasting impact of that generation was? What show from your generation had the most influence?
Rosner: People often say Hill Street Blues was one of the most impactful scripted TV shows of the 1980s. That makes it part of my generation. It was one of the first major network dramas to juggle multiple ongoing storylines without tying everything up neatly at the end of each episode.
Some people even point to that shift in television—audiences following more complex plots—as loosely paralleling discussions of the Flynn effect, the idea that average performance on certain kinds of IQ tests has risen over time. I’m not sure how strong that connection really is, but Hill Street Blues certainly marked a change in what TV assumed viewers could handle.
A writer published a short book comparing an episode of Starsky and Hutchfrom the 1970s with an episode of Hill Street Blues from the 1980s.
Now, network television is dwindling.
As we’ve discussed many times, network TV is mainly for people who are either too old or too poor to bother with cable or streaming. As a result, it tends to be overly explanatory and overly explicit, and audiences don’t like that. Since everyone has seen everything by now, we expect at least a minimal level of sophistication.
Rosner: In our entertainment.
Jacobsen: All right, what else do we have? What do you think the subsequent development will be after large language models, once we better understand their internal architecture? Just the multimodal systems that are already being sold?
Rosner: Companies like Google are already selling AI applications that more easily understand what we want. To do that, you need multimodal understanding.
You need AI that can understand both visual and auditory content. We already have systems you can yell at, and they recognize what you’re saying and follow your verbal commands.
The next generation is already here—AI that understands many of the same inputs from the world that we do.
Part of Musk’s compensation package—where he stands to be paid close to a trillion dollars over the next ten years—has productivity milestones. For Musk to receive that immense salary, roughly a hundred billion dollars per year for a decade, he must sell a million humanoid robots by 2030.
Obviously, among the following forms of AI will be AI with agency, even if that is a terrible idea to release into the world without far more caution than simply letting it proceed unchecked.
Jacobsen: What was the last math problem you worked on where you felt you made significant progress?
Rosner: That would have been… I worked on the twin primes conjecture decades ago, and I felt I had a framework that might offer more insight.
But I never went anywhere with it. Many of my attempts at math end that way.
We’ve talked about how counting numbers incorporate an infinite amount of information because they are infinitely precise. We use them without acknowledging the precision or what goes into it. I came up with, I think, a way of looking at how…
Rosner: Where maybe some of the information—is where the information is.
Information is added to numbers by a series of choices in determining where the primes appear along the number line.
I came up with something I call “meta-primes,” which are systems of prime numbers where the primes appear in a different order.
When you lay out the counting numbers, you start with 1. The following number must be prime. After that, you have a choice: the number after the first prime can be either a new prime or the first prime squared.
On the number line we’re used to, it goes prime, prime, then the first prime squared, then another prime, then the first prime times the second prime. But that order—if you think of it as a sequence of choices rather than something determined strictly by moving forward one integer at a time—creates many possible number lines.
You get many different number lines, and they are all subsets of the number line we use, the familiar counting numbers.
For example, if you choose to have the following number after the first prime, be it the prime squared rather than another prime, you’ve left out the multiples of three. You get 1, 2, then the first prime squared, which is 4. If you keep the sequence as compact as possible, you’ve just omitted three and all its multiples. But this is still derived from the ordinary counting numbers, which contain an infinite amount of information.
If you instead view the numbers we use as not having infinite information “built in,” but instead having that information added by the sequence of choices—at some point you must choose between several possibilities, such as A×B or something else—then the structure looks different.
Once you reach five on the number line—the third prime—you again face a choice between yet another prime or A×B. You have multiple possible paths. Thinking this way, what I hoped to develop (but never fully did) was a way to show that the number line is the most compact product of an infinity of choices as you move forward, always choosing the most compact sequence of numbers.
My argument would be that the amount of information that goes into this, and that is contained in the number line precludes determinate results such as there being a finite number of twin primes. The twin prime conjecture says there is an infinite number of twin primes—pairs of primes that differ by two. The first pair is 3 and 5, then 5 and 7, then 11 and 13, 17 and 19, 29 and 31, and so on—an apparent infinity of them.
Statistical analysis, along with number theory, strongly suggests there should be infinitely many of these pairs, but no one has been able to prove it. I was hoping there might be a way to prove it by showing that precluding all twin primes after some enormous number—for example, saying the last pair exists somewhere around 150 digits and then never again—would require too much information.
Twin primes become increasingly rare along the number line, just as many prime-related structures become rarer. But eliminating all twin primes beyond some huge cutoff would require more information than the number line actually contains. The determinateness of the order of numbers already requires a tremendous amount of information, and there’snot enough left over to do something special like forbidding an infinity of twin primes.
But I never did anything with it. I never learned number theory formally, and I never worked with anyone on it. Dean Anatta almost immediately recognized that the order of primes along the number line can be imagined as a series of slices at a particular angle into a corner of an n-cube—a multidimensional structure in which, as n goes to infinity, each axis corresponds to a prime. You slice into this infinite-dimensional corner at an angle that intersects each prime axis in a certain way. That insight was different from anything I’d seen.
So there you go—Rotten Tomatoes.
Maybe it wouldn’t have led anywhere. If it were going to lead somewhere, perhaps someone else might have found it sooner. I don’t know.
Jacobsen: You want to reconvene tomorrow?
Rosner: Let’s reconvene tomorrow. On the third, yes. I’m much clearer in my discussion of this when I haven’t been sleeping off a massive amount of Thanksgiving turkey. I wrote articles about this for Noesis in the nineties.
So there are Noesis articles on meta-primes—at least one or two.
Jacobsen: There were later ones about road rage.
Rosner: Road rage. That one was in Esquire, one of my two published magazine articles.
Jacobsen: What’s your favourite thing you’ve written or read—doesn’t have to be comedy broadly.
Rosner: I don’t know, probably The Diamond Age by Neal Stephenson at the time I read it. I don’t know—the thing I’m working on now, which I’ll discuss in more detail as it gets further along.
Last updated May 3, 2025. These terms govern all In Sight Publishing content—past, present, and future—and supersede any prior notices. In Sight Publishing by Scott Douglas Jacobsen is licensed under a Creative Commons BY‑NC‑ND 4.0; © In Sight Publishing by Scott Douglas Jacobsen 2012–Present. All trademarks, performances, databases & branding are owned by their rights holders; no use without permission. Unauthorized copying, modification, framing or public communication is prohibited. External links are not endorsed. Cookies & tracking require consent, and data processing complies with PIPEDA & GDPR; no data from children < 13 (COPPA). Content meets WCAG 2.1 AA under the Accessible Canada Act & is preserved in open archival formats with backups. Excerpts & links require full credit & hyperlink; limited quoting under fair-dealing & fair-use. All content is informational; no liability for errors or omissions: Feedback welcome, and verified errors corrected promptly. For permissions or DMCA notices, email: scott.jacobsen2025@gmail.com. Site use is governed by BC laws; content is “as‑is,” liability limited, users indemnify us; moral, performers’ & database sui generis rights reserved.
In-Sight Publishing 