Exploring AI, Knowledge, and Ethics with John Helm (Bonus Episode)
In this bonus episode of "How Many CTOs Does It Take?" podcast, hosts Scott Porad and Brad Hefta-Gaub sit down once again with technologist and philosopher John Helm, this time for a wide-ranging conversation about artificial intelligence, epistemology, and what it means to be human in an age of machines. John draws on his background in statistics and philosophy to frame AI through the lens of knowledge, reasoning, and moral agency.
The episode touches on the hype and promise of modern AI systems, explores the difference between correlation and understanding, and raises deep questions about consciousness, ethics, and whether general intelligence can exist without emotion. John even challenges the language we use to describe AI by grounding the conversation in ancient philosophical questions, the structure of scientific truth, and how humans conceptualize value and meaning.
- 00:00 Introduction: What Makes Us, Us
- 00:19 Diving into AI: Hype and Reality
- 01:03 Philosophical Insights: Knowledge and Understanding
- 03:00 AI and Human Understanding: Patterns and Predictions
- 05:19 Ethical Concerns and Human Morality
- 08:14 Concluding Thoughts: Ethics, Knowledge, and Governance
- 10:28 Outro: Follow and Subscribe
Resources:
- How Many CTOs Pod: https://howmanyctospod.com
- Scott Porad: https://www.linkedin.com/in/scottporad/
- Brad Hefta-Gaub: https://www.linkedin.com/in/bradheftagaub/
- John Helm: https://www.linkedin.com/in/johnlhelm/
#TechPodcast #EngineeringPodcast #PodcastForDevs #HowManyCtos #Podcast
#Ctos #Technology #Leadership #EngineeringLeadership #CtoPodcast
#ArtificialIntelligence #AGI #MachineLearning #AIHypeCycle #AIPhilosophy
#EthicsInAI #MoralMachines #ReasoningModels #Epistemology #KnowledgeAndTruth
#HumanConsciousness #AIAndHumanity #GeneralIntelligence #CorrelationVsCausation
#AIAlignment #PredictiveModels #PhilosophyOfAI
Brad: It boils down to what makes us us.
Scott: Let's talk about AI for a minute.
Brad: You are talking to an AI.
Kidding.
Scott: I'm going to start with a very open-ended
Brad: question Look,
Scott: before I have more specific ones.
What do you think?
Brad: we're in a hype cycle.
We all know what that is, right?
It's a huge...
advance in the technology and it will make some material improvements that will have lasting benefit.
I think it's just like other hype cycles. There are a bunch of things that we think it might be able to do that it may, you know, we're maybe oversetting expectations and that will set up for some disappointment.
And it actually,
I think,
instigate some very deep questions because essentially,
you know, I do a lot of statistics as at least Scott knows and statistics is essentially acquiring knowledge and everyone,
you know, people hate statistics because you can prove anything with them and it is true.
So that leads to sort of a philosophical question that I think Scott knows where we're going here and it's the branch of philosophy that deals with the
to the question of knowledge,
which is how do we know what we know and how do we know it's true?
The fancy name is epistemology, and there are kind of two poles in the domain of epistemology, and one is sort of, it's called modern empiricism, and the other one's called modern rationalism, and modern empiricism is knowledge is acquired completely by experience, and causality is mostly correlation. Big philosophers are like Locke.
Locke and Hume that talk about this.
And then the other is modern rationalism, which is kind of platonic, but Descartes kind of gets associated with it.
And it's this perfectionism nature and it's out there and our job and we have better and better understanding of nature's perfection through experience.
And so there's this platonic form like the perfect icosahedron.
you know we'll never get it but as we get we study it more and more we'll understand that and you know and neither extreme is completely true um there was a philosopher named emmanuel kant that you know he's considered more an empiricist but he actually in my view did a synthesis and i'm not going to go there well very quickly um he had these things called the category of understanding and it was from experience but one of the categories of understanding is the idea of number and there
John: was
Brad: was one many infinity but infinity can never be experienced so it can't be a pure empiricist but that's a footnote i digress so it gets back to what are these ais well you know what they do is essentially they're trained on a giant amount of linguistic experience and their patterns just like scott probably knows how what i was going to say when i started a few sentences because i've known them since 2000 and
and four or 2003 and we've talked a lot so they're patterns to my speech so he can predict the next token coming out of my mouth because he's been trained by a whole bunch of long lunches and dinners and walks in the sculpture garden where I babbled endlessly and he and trained his neural network so he can predict my next tokens that's what they're doing they're doing my golden retrievers do when I do certain patterns
They predict that they're going to get fed or they're going to get treats or there's a ball in the air and they're going to retreat.
And I don't want to depreciate that. That is purely human understanding. And the fact that these
systems work so well and then you can apply reasoning models on top of them which are introduced exogenously by us so it's kind of like what Kant did when he says okay we have this framework but then we have this notion of infinity which you can never get from experience because we won't live long enough but we can conceptualize it he's exogenously introducing kind of a reasoning model into
experience driven knowledge so what are we are we are we human or are we Cartesian or are we a synthesis and if there were a synthesis where did the reasoning models get exogenously introduced you know is there is there a way to create knowledge without having knowledge to create it I don't know so back to these models they're
human pure correlation engines and there's a heck of a lot of getting through life which works just fine with correlation and sometimes we got serious ones and you know experience teaches us that things were subtler but but they're pretty good so i think it's going to be pretty significant um are they gai that's a philosophical question
and I've already introduced enough so that you guys both know both that I don't know what I'm talking about and I can't answer it even if I did.
Scott: Do you have any ethical concerns about it?
Brad: Of course I do.
Scott: There
Brad: I mean,
Scott: you
Brad: look,
Scott: go.
Brad: you can teach reasoning and it boils down to what makes us us.
Right.
And,
you know, we have a moral fabric and, you know, whether you believe in Darwinism or not, there are behaviors that, you know, like like self-sacrifice for the greater good that make no Cartesian sense at all that evolved because.
Because instead of, you know, Darwinism being survival of the fittest, which is not what it is, it's reproductive success.
And what's what's being successful in the beginning, it was individual species, but later on species became packs,
herds, schools,
whatever,
murders of crows and for the collective to sort of survive these individual behaviors evolved that made the collective.
collective more successful and um and you know and and we have this idea of consciousness and there's a debate in the philosophical community about where does consciousness come from and if you're kind of aristotelian you you might be comfortable with the argument that it began with how do you survive well when you're a completely non-sentient thing like an amoeba you have no idea you don't you survive by
by reproducing faster than whatever is around to eat you but eventually you want to get out of the way well how do you get out of the way you kind of have to know where your body ends and the rest of the world begins so you start actually through survival maybe not knowing consciously with a theory of mind that there are other amoeba out there and they have feelings too you can sort of see how our neural network evolved and its purpose is to help us survive and be
and be successful as a species.
So I don't know if you can be a GAI without emotion.
That's really kind of the, I've now had too many beers at the bar question so I can say this and think it's safe,
but I don't know.
So I don't know what these things are vis-a-vis the human condition.
I don't know what the role of ethics and morality.
might i don't know if they're nice to have or must haves in order to be a gai is so if you if you put in a reasoning model that's like you know things that move kill it right and it has no feeling and the ai can predict what the next token is going to be when i jump around behind a tree it's going to be bad so but but but that you don't need to have general intelligence to do that and
And you might be able to make machines that have no moral intention.
And, you know, there are lots of Star Trek episodes and things. And then the flip side,
right,
remember in HAL, the 2001 computer,
Scott: Mm
Brad: in 2010,
Scott: -hmm.
Brad: we find out the reason why HAL was killing Dave was because its programmers lied to it and had this existential conflict.
Oh, I remember the point I made.
In philosophy,
the problem of ethics is not the root one. There's three tiers of philosophy.
There's the problem of knowledge,
the problem of behavior,
and the problem of governance.
The problem of knowledge is epistemology. The problem of behavior is ethics.
Given what I know,
what is the proper way to behave in this world?
Well, if people can't agree on what we know and there are axioms at the bottom of what we know and those are leaps of faith, which leap of faith is the right leap of faith,
right? Is the geometry of nature?
Do parallel lines meet at infinity or not? That's Euclid's fifth postulate. Well, it turns out that if you buy special relativity,
it turns out that they meet at infinity because that creates hyperbolic coordinates and hyperbolic coordinates are the coordinates that relativity is described in. But prior to that,
we thought it was Euclidean, right?
Where X,
Y, and Z and the parallel lines never met.
And there was a big crisis because this mathematician...
A mathematician named Riemann said,
I'm going to throw away the fifth postulate and see what I get.
And he got a completely self-consistent geometry.
So now we have Euclidean geometry,
which is completely self-consistent, and we've got Riemann geometry or hyperbolic geometry that's completely self-consistent. Which geometry is the true geometry?
So that gets back to the question of ethics is subordinate to the question of knowledge. Oh, and the last one is.
is the problem of governance,
which is political science.
Given how we individually behave,
how should we behave as a collective?
All depends on what you believe is true.
How do we know what we know and how do we believe it's true?
So ethics may not be necessary for a thinking machine to function very well in need to sadness.
And how did nature fix that?
Well,
lots of trials.
That's back to Darwinism.
organism we're only one we're only one meta organism so we get one trial we only can we have to do it right the first time and if we don't um the rock will start over again
Scott: Yep.
Brad: sorry you asked a heavy question you got a heavy answer and of course i'm i'm full of bunk
Scott: I love it.
John: This episode of How Many CTOs Does It Take podcast was edited by Saffron Epigal. Our theme song was AI Generated by Brad.
Please follow us on all the social media platforms at howmanyCTOsPod.
You can find us on YouTube,
LinkedIn,
Instagram,
Facebook,
Threads, X,
TikTok,
and Patreon,
or at our website at howmanyCTOsPod.com.
And remember to rate, review,
and subscribe on the podcast platform of your choice.
Thanks for listening.