11: Drones, AI, and Media: Understanding the New Frontiers of Warfare and Reporting

Hi, I’m Jesse Hirsh, and welcome to another episode of Red Tory, recorded live in front of an automated audience. And today we’ve got a special treat.

Alan is away, unfortunately, but we’re joined by Alex Fink, who I suspect is going to provide us with some excellent insights, in particular into the role of AI in journalism, which is.

It has been a recurring subject for us, partly because AI is so occurrent and in particular when it comes to media, both a positive and a negative, depending on how it’s used, depending on the context.

But in researching you, of course, I did see that you’re involved in another area of AI, which is around drones and drone software management, in particular with regard to the ongoing conflict in Ukraine. So that’s kind of our rough agenda for today. But as we get started, Alex, what drew you to AI in the first place?

Why don’t we do that as a quick establishment before we get a little deeper into Ukraine? Because I suspect I know why you’re interested in Ukraine, but why AI? What is it about AI that excites you in particular?

I mean, I came to this from the technology side. I spent about 15 years building perception systems, video cameras, computer vision, that sort of thing.

Over time, I drifted from the camera hardware towards computer vision, which is a type of AI.

Then I discovered there are new interesting things like natural language processing, which is what Most people consider AI after ChatGPT became popular. Now I’m seeing decision making and other kinds that are also interesting. It’s not that I get the right code anymore.

At this point, I’m just directing the orchestra for the most part. But the technology is still really interesting, and it’s all just one step removed from what I did previously, and I’m evolving in that direction.

Right on, right on. Now, we’ve talked about Ukraine here on the podcast for a number of different reasons.

Partly because as Canadians, there are a lot of Ukrainian Canadians who. This has been a very important conflict that we’ve all been watching. But from a technological level, we’ve seen a revolution in drone warfare.

I mean, it’s been really quite incredible, the rapid pace of IT innovation and the role that drones have played, especially for Ukrainian forces. So why don’t we start with that kind of big picture, but also personal level, how did you get involved with drone technology?

Was it before or during the conflict? And how did you see this as an opportunity to play an important role, to do your part, as it were?

So. So I guess the short answer is I didn’t.

A friend reached out to me and he was the head of computer vision in Ring, which was based in Ukraine, at least most of their engineering was. And I’ve known him for 14, 15 years. We’ve worked on previous projects together.

And he just asked me, here’s what I want to create, here’s the company I want to register, how would you do this? And I wrote him a four page memo of how I would set everything up and he replied with a one liner saying, why don’t you do it?

And that’s how I ended up doing it. So that’s how I ended up involved in it. I am not Ukrainian myself. Obviously there’s one side on this conflict that is right and the other is wrong.

But that is to say, from my perspective, I don’t have an ethnic reason to be involved. It’s just cool technology. I like what we’re building and I am on the side of the people who are right. Double whammy. Why not?

Although your latter point, and I do agree with your assessment of right or wrong, unfortunately not everyone agrees with us and not everyone sees it in such simple terms. But to my point about a revolution, in terms of how drones have been used, can you give us a big picture here?

Because it does seem like we’ve seen a rapid pace of innovation and maybe you disagree. Maybe from a technology perspective you’re like, no, no, no. All these pieces were there. People are just putting.

Or am I right in thinking this has been a huge pace of innovation, a huge technological leap when it comes to at least the way that drones are being used?

Well, I think it’s both at the same time. As with everything AI, it’s moving too fast and too slow at the same time.

So on the one hand, when I look at what is being used at the front line, it’s just 95% Mavics. Why are they using this Chinese piece of shit? I don’t know, but it actually takes a while to propagate through the system.

On the other hand, yes, we’re seeing a lot of really cool stuff. And by the way, this is not the only conflict where we’ve seen this right?

If you saw Saudi Arabia versus Yemen, drones played a really big role in that.

If you saw Nagorno Karabakh, Armenia versus Azerbaijan, then I tend to jokingly say that war was won by elbit because they sold their drones to Azerbaijan and basically that won the war. So I think drones are playing a bigger and bigger role.

I think that they are going to play an even bigger role in the future because my personal Definition of a drone is anything that can be automated remotely. Nobody said it has to fly. It can swim, it can swim underwater, it can drive on dry land, or it can stand and just coordinate where it’s shooting at.

If you basically turn the high Mars into a drone and all of these things can be coordinated in a joint force operation. And I think the bigger question to me is, let’s talk about the flying stuff, right? We used to have an airplane with a pilot in it.

We removed the pilot from the airplane, which is definitely progress, but we still have a pilot piloting the airplane. Why? That makes no sense to me. You should have one person saying what needs to happen, click a button, and then the button does what it’s supposed to.

Just like when you see a text box on your PC, you click a button. You don’t know which process is going to run on which core of the cpu. That’s none of your business.

You just want the button to do what it’s supposed to. So from my perspective, figuring out which drone flies where, that’s the same level of detail humans shouldn’t concern themselves with.

Humans should just decide what is the objective. And that obviously includes a lot of life or death decisions that we don’t want to let AI decide for itself. Right.

But once humans decided that everything else is an implementation detail and if something changed in real time, then the drones can adjust much faster than a human pilot would because they have much better 3D awareness than humans generally. And I think most importantly, drones can communicate among themselves much better than humans communicate because I speak pretty fast.

Pretty fast is like maybe 200 words per minute. That’s 100 bits per second in computer speak.

You don’t want that to be the bottleneck of the cutting edge AI system that the entire world standardizes itself on.

So, so let me ask you what may come across as a stupid question, but I think for me and our listeners is, is kind of straightforward, which is like you mentioned how they’re using Mavics in Ukraine and you mentioned what was the other manufacturer in, in the other conflict you mentioned?

Oh, Nagorno Karabakh. Yeah, that’s Elba. That’s an actual dedicated, nice military drone. So the Hermes 450, Herman’s 900, they’re pretty good.

So in both cases, is the software separate from the hardware?

Like when you, if you’re a military and you’re procuring these, let’s say, with the Mavics, are they using DJI software or are they using different software to both control and to your point, potentially program the drone so it could go on an autonomous mission?

Or are these complete packages in which the vendor is giving you kind of the whole deal and you’re limited to what functionality that they have baked in.

I think it actually differs between real military grade stuff and consumer stuff. Home baked consumer stuff converted to military. Right.

So on the home baked side of things, it’s a wild west, especially in places like Ukraine, you have a lot of stuff flying with open source software that is just not encrypted. So if it falls in the wrong hands, the enemy gets all the data. Right.

And that’s one of the things that we try to remedy even before doing all the fancy AI stuff. Just secure firmware, upgrade, secure video streaming, secure communication so that you’re not actually jeopardizing yourself by flying this thing.

Five manufacturers, and every one of them has their own operating system, their own processor, their own disk drive, their own memory, their own everything. Right.

And I think it’s bound to evolve in the direction of a more diversified supply chain where you have five people who make this component and 10 people who make this component, and maybe 200 different manufacturers of hardware that mostly looks the same with interchangeable parts, and then hopefully one operating system to rule them all. Us.

Well, and tell me more about that.

In terms of the importance of the operating system, the importance you alluded earlier in terms of swarm management of not just one pilot, one device, but multiple devices talking to each other. So one assumes the ability to defend or counter attack becomes harder and harder.

Yeah, So I need to be careful about what I say because not everything is public, obviously, but in general you can think about the human on one side and the hardware on the other side. Everything between that I view as something that software can automate pretty well.

In this case, it’s not an operating system that runs on one device, it’s actually distributed.

Because you could have multiple edge devices, let’s say flying drones, and you can have a home based device that sends them additional information from the outside world.

And obviously the edge devices need to be able to operate even if they lose connectivity to home base, even if they lose gps, because let’s say it is being jammed by the enemy.

And even in other kind of conditions, they could in theory also lose connection to each other, though that is less likely because they’re pretty close together. So any jammer is going to be farther than they are from each other. But that entire thing together can be called a single operating system.

It’s just distributed across multiple devices.

And to what extent you mentioned the countermeasures, it’s been interesting to see now fiber optic based drones, where they have this really tiny fiber optic cable that kind of flies behind them so they can’t be jammed. To what extent is that a stopgap measure?

It sounds to me like some of the software potentials from a counter jamming is having multiple drones relaying these communication signals and having other solutions rather than going exclusively to like a contained system, a fiber optic wire. I’d love to hear your thoughts on that.

Yeah. I would say that jamming as a way to counter drones has a very limited shelf life. It’s relevant now. It will be useless in a year in my opinion.

And so countermeasures to jamming like fiber optic have a limited shelf life. Right. It’s a stopgap to counter somebody else’s stop gap, in my opinion. Right.

I think basically visual data and other things that you get from a bunch of other sensors is enough to fly without gps. Right. We do it quite well. I don’t even think we’re by ourselves here. Right. And so I think that countermeasures to drone have to be more kinetic.

They cannot just be radio waves. And this thing will stop working. Right. Somebody needs to drop some shrapnel on this thing if they want to stop it.

Although to what extent would it be drone versus drone that if your swarm is greater than their swarm, that that could be a competitive advantage?

I mean, obviously it’s going to be that at some point, because you’re not. I have seen some startups develop interesting guns to shoot at drones. Right. But again, that sounds like a stopgap measure.

Even if it works on the particular drone that is being used right now, it will stop working very soon. So y it will be kind of an arms race of drone versus drone.

But I do want to kind of point out that I think the defense side of this is much more complicated than the offense in some sense because the offense tends to shoot at simple targets, either stationary or moving in a very predictable pattern. Whereas the defense has to shoot an evading target that flies around in 3D space.

So people will figure out how to shoot a supply truck before they figure out how to shoot the drone that is trying to shoot the supply truck.

Now, you know, I started by suggesting that we’ve seen kind of a revolution in military tactics and technology outside of Ukraine. Are you seeing other countries, other military organizations kind of waking up to the potential of drone technology?

And is this, again, I assume you can’t get into too much details, but it strikes me this must be a growing marketplace, not just of buyers, but of vendors as well, who are eager to be part of this new ecosystem.

I mean, yeah, the entire world is obviously waking up. The question is, do they have enough impetus to wake up quickly, or can they be in a slumber for a few years and then be elated out?

So, obviously, countries that are in an active conflict right now, they have to modernize, right? They can’t afford to fight the old way because the old way doesn’t work. Countries that don’t have an active conflict, they take a bit longer.

But still, everybody knows that’s where they need to go. Every doctrine of any Western military right now says this is what they’re building, right? Drones warming, all of these things.

Autonomy, collaborative autonomy, they tend to call it, which is like autonomy plus warming. So you see this everywhere. The only question is, how fast is this transition going to be, and is anything going to force it to move faster?

And do you think there’s any, for lack of a better phrase, regulatory element here? And I say this only because governments do like to restrict the technology, the weaponry that their citizenry can get access to.

And all of a sudden, we’re starting to look at drones different in terms of how they can be improvised and adapted. Do you think that there’ll be greater control over drone technology, or is the proverbial cat already out of the bag?

Well, there already is to some extent, because you have export controls in all the countries that are developing new stuff, Right.

But when it comes to regulation of the AI element of this, that is a much tougher question, because the people regulating it don’t quite understand the thing they’re regulating yet. Right? And so if we are able to help them and at least explain to them what could be regulated, what couldn’t, I think that would be really good.

If they listen to somebody who is not from the industry, then I think it might be worse. And if they overregulate, then the countries that will develop it first are the ones who absolutely don’t care about ethics. Right?

And so we kind of have to try help our governments regulate it in a proper way, because otherwise China and Iran and Russia are developing autonomy first.

And I think that’s a really good frame to look at the media and journalism side.

But one last question, which again, you can decide to answer how you wish, but it would be a mistake if I didn’t ask for your thoughts on the Ukrainian conflict and where you think it’s going to head in the weeks, if not months to come, especially given all the politics and so called negotiating that’s supposed to be happening right now.

Yeah, well, my thoughts on the conflict don’t start with where I think it’s going to end. Right. I think it’s pretty clear where it began. Right.

And it’s pretty clear that you have a bunch of borders that everybody pretty much agreed to and then one country decided that doesn’t like those borders and likes different ones. Right. And they came up with a whole bunch of nonsensical reasons which again, I was born in the Soviet Union.

And so you can just go down the list of these nonsensical reasons. I don’t know how many there have been so far. It’s a different one every day.

I think in Putin’s interview with Tucker Carlson it was something about Hungarians wearing cylinder hats. I’m not sure, but it’s that level of argumentation.

And unfortunately the news ecosystem is such that people end up thinking that is true and repeating it very often. So that’s on the causes side. Now where it’s going to end.

I think we’re kind of converging into where it’s going to end, even though we don’t think it’s a just ending. Right. But I think it is pretty clear that both sides can’t really sustain this for much longer.

And I don’t think either side is sure which one will be unable to sustain it first. And so they have a reason to come up with some sort of a way to at least take a break. Right.

Or so are we getting a four week truce or are we getting more sustainable ceasefire? I don’t know, but we’re getting something because I don’t think they can really sustain this. And again, it’s on Both sides, really.

If you look at data from the Russian economy and how much they have to pay soldiers to enlist, it’s exponential. And their economy is not doing quite well. In Ukraine, the demographic problem might actually be worse. Right.

Especially since they don’t want to have the youngest part of the population, sir. Right. Basically they are only recruiting people from 25 and older. So I don’t know. I think both sides can’t really afford to continue.

We’ve seen crazier things happen. But it’s been three years. Most wars don’t last that long. Like world wars tend to last four years.

It’s hard to sustain the space for much longer.

Yeah. And that was a perfect segue to the Other Web, which is another company that you run, which I’d like to learn about.

And you sort of alluded to it in terms of. I find the difficulty of following the Ukraine conflict is deciphering what’s true and what’s not.

That, you know, certainly the Russian side is so active in putting out their own one sided information, their own disinformation, reframing of information to your point, about even how the conflict started on repeated sides. Sorry, over repeated dates.

Tell me a bit about OtherWeb and the role that it plays in trying to clean up our overwhelmed, our polluted information ecosystem.

So that was a big part of my concern. And news was always very important to me because again, I was born in the Soviet Union.

We left because my parents were listening to Voice of America at night while nobody was hearing to get real news. And so we knew this thing is about to fall apart. And we were on the periphery where it was kind of dangerous. So we left.

There were tanks on the street where I grew up three months after we left. So listening to news and getting real news is pretty important.

And being misinformed, as most people were in the Soviet Union is probably not leading to good decisions. So that’s why it was always very important to me.

And when I had this crisis of conscience of I don’t know why I’m making more cameras, the world doesn’t need more cameras, I decided I’m going to work on this thing. And so that’s how OtherWeb was born. And the basic idea was, let’s just initially help people select higher quality information.

We didn’t want to do source based filtering. Just saying this outlet’s good, this outlet’s bad because every outlet has a Gaussian somewhere in there.

So we just wanted to figure out what is a good story. And by the way, we also didn’t want to say this is true or this is false because sometimes you don’t even know.

We just wanted to define what is a good story supposed to look like. Let’s say you make a crazy claim, there’s a line in your kitchen right now.

Okay, I’m not going to break into our house to check whether it’s true or not. Right. But do you have references? Do you have eyewitness accounts? What kind of language did you use to describe it? Did you say biggest line ever?

Right. Or did you just describe the lion and its location?

We just analyze all these form based things and decide which stories are the best version of that particular story. Then we try to just sample far and wide, 900 plus sources right and left and center and let people basically configure everything themselves.

That was the general idea. We have recently added a lot more on the side of the content creators. So we have an AI tool for journalists called Press Hub.

We have our own newswire, technically, I guess, competing with AP and Reuters, though ours is free and under creative comments.

But the general idea is people who can’t afford a subscription to one of the big guys, we still want them to have access to what happened in the world written in purely factual, simple language that they can reuse in their own articles.

Because you have a lot of ghost newspapers in the US where they just don’t have the money to buy a reuter subscription and they don’t have journalists to send to Iraq to report on what happened there. Right. So we’re adding a whole bunch of other things.

We recently launched a tool for bloggers, or rather for corporations that want a corporate blog so that at least that is in better quality. So we’re doing everything that we possibly can to help people consume higher quality information, whatever that might mean.

And the creator tools you mentioned, the corporate one and what was the name of the first one you mentioned? Press Hub.

Yes, PressHub. AI is the tool for journalists.

So is this like an LLM to help journalists? Give me more details on.

Yeah, it has a few LLMs built in. So it has a few elements. The first one is the built in newswire.

So you have a constant feed of new things that happen in the world that is already curated and written by our LLM. Right. Then they can pick any one of those or start a new article by themselves.

They have a nice editor with AI tools that help the journalists write faster. Right. So let’s say that they see a particular claim, they want to fact check that claim against Wikipedia and other sources.

One click you have a fact checker running on it. Let’s say they see somebody make a really dubious claim, they want to generate good counterclaims for an opinion piece. Right.

There is a counterpoint generator that just comes up with the best arguments against something, whether it’s true or not. Right. It just argues against whatever point of view. Basically being like a proper Jewish guy, just arguing against whatever has been said before.

Sorry, I’m Jewish, so I’m allowed to make that joke.

And I’m Jewish, so I’m allowed to laugh. Please continue.

So it has a bunch of these tools.

The final tool is once the article itself is written, it actually has an SEO tool that generates multiple headlines for you to try, multiple metadata. Right. Sort of description, headline and header. And it even generates social media posts of how you might post about this. So it just tries to.

I want to think how to phrase this, probably the best way to phrase it, as the journalist does the first 10% and the last 10%, we want to automate the 80% in the middle because that means one journalist can be much more productive.

And we know for sure that there is an element of quality to what they do because at some point in there they used R tools that we are pretty certain of. It’s also worth noting we opened the source code of all our models so that they can be audited.

You can’t use them for your own stuff, but you can read the code, you can look at the data sets, you can be sure that there’s no hidden right wing or left wing bias in there or hidden Russian propaganda in there. Which is, I think, much better than the transparency we get with a human editor. Because you don’t quite know what’s in the human’s head. Right.

Maybe they were bought. Maybe they’re having a bad day. Maybe they’re just upset at something.

Maybe they’re just hungry.

That is true as well. I’ve seen a lot of research that judges give out much harsher sentences right before lunch than right after lunch.

Oh, it’s absolutely true. I think blood sugar is a huge bias.

Now let me ask you from the perspective of both an entrepreneur and a researcher, you described kind of a set of ideas, a desired kind of framing.

I’M curious then, based on these resources, based on kind of the logic of finding quality content, how are people using the site, how are people using the platform? Both on the creator end, but also on the user end.

Because I’m assuming you’re able to kind of see the analytics, you’re able to see the usage data, and it both influences how you think of the product. But what I’m more interested on a kind of cultural, anthropological level is how are people using it? Right.

Like how are you kind of impacting how your parents would stay up late to listen to Voice of America, the consumption habits? Right. Because this is a challenge in our kind of overwhelmed, polluted environment.

So on the consumer side, I have a lot more data because there’s more than 18 million users. Right.

On.

On the producer side, well, it’s a paid resource, so it’s like hundreds of people now, maybe growing towards thousands, but.

That’S still a decent sample size.

It is, but let’s start on the consumer side. So I would say we broadly have two archetypes of user.

The first one is, like me, it’s basically an information junkie that was already spending a lot of time reading the news. And so this gives them a more efficient way to go through stuff. And this is why if you go through our site, you will see a lot of customizations.

Basically everything about your feed is customizable up to and including what kind of emotions you like to receive. And you can say less depressing, more infuriating, whatever that might be.

You can configure everything because that is a big part of our user audience.

Also, if you ask that audience what’s the most important feature on the site, they will typically tell you summaries because they don’t read the full articles. They skim the summaries and we have a bullet point summary for every article. Right.

And they only click on maybe every 20th article roughly to actually read the entire 800 word piece. So that’s one archetype. The other archetype is people who have checked out from the news a long time ago.

They have been avoiding this thing because it is infuriating, it is depressing, it annoys them. That is actually a very large group. It’s more than 42% of all adults in the United States right now that are consciously avoiding the news.

So the version of the news that we present to them is much more informative and much less distressing. Right. Like a lot of them, tell me I know what happened and I’m not upset about it.

Well, and to your earlier point, a lot of the headlines, a lot of the framing are meant for emotional triggers, not for transferring information. And that strikes me as a key design principle to the way in which you’re creating this user experience.

Yeah, so we actually have a separate filter for what we call attention grabbing headline, but we really mean clickbait, right? We just don’t want to offend anyone. So it’s attention grabbing headline and if something scores too high on that filter, we replace the headline.

So we just do a one line summary of what the article actually contains and we put that as the headline within other web. Of course, if you click through to the article you will see the original, but we don’t want any clickbait on our platform.

So we’ve gone as far as to say, even AI can do this better than you, Mr. Writer. We will replace what you did.

Well, especially when it’s manipulative, right? Especially when it’s sort of digressing from the ultimate goal, which is to inform the reader. Right. To help them understand the information.

And that’s the other thing. One of the filters we have is for propaganda techniques.

There was a nice data set published by I think QCRI a While back with 19 different propaganda techniques that they’ve identified. So we created the filter for each one and we’re trying to detect them as much as possible.

Obviously you can’t always tell the difference between propaganda and just a politician speaking. Sometimes they look the same, but at least statistically there is a difference. So sometimes you can filter out the obviously bad stuff.

That’s another thing I wanted to point out where our entire philosophy is. We don’t want to tell you this is true, this is false. We don’t want to tell you here’s the only right version of this particular article.

What we want to do is to filter out the bottom 90% and that increases your signal to noise ratio. Now, within that, some of the stuff you read will still be wrong, but there will just be less of it. Right.

And it’ll be less overwhelming, it’ll be less intimidating. To your point of the people who avoid the news or who because of the last few months, just turned it off completely.

Now, you mentioned one of the through lines. One of the things that have come up a lot in our conversation today is what I hear is a commitment to openness, a commitment to transparency.

Talk about the way in which this platform makes it easier for people, even if they’re just reading the summaries, to dig deeper into particular issues and what’s the significance of you mentioned making it transparent, the way in which you weight some of the AI filtering. Because I think that’s important given the potential for customization.

Right, because you’re allowing people to really create a very custom interface, a custom view to the world. But you also want them to trust you. Right? To trust that you’re doing so in a manner that is ethical, that is transparent.

So do you want to elaborate that both on a philosophical level but also on a technical level?

Well, I think there is the correct, idealistic, philosophical description of what we’re trying to achieve. But the reality is that the news ecosystem today is very far from any philosophical ideal.

And people have become really cynical about what they’re reading.

And if it’s not the resource that they have always read their entire life, they’re just assuming this is either right wing propaganda or left wing propaganda. Propaganda. Right. People just have this inbuilt assumption and you need to prove to them that you’re not the bad guy they’re imagining.

So in an ideal world, we would be able to trust each other. And unless I give you a reason not to trust me, then I would be trustworthy.

But in the world that we live in, at least in the news ecosystem, you are a villain until proven otherwise. So you have to do a lot of things to prove that you’re not a villain. So OtherWeb is registered as a public benefit corporation. Right.

That is one thing that we did so that we are actually allowed not just to pursue the profit motive, but also to pursue our mission, which is to improve the quality of information people consume and to balance those two out. Right. In an open way so that we occasionally have to submit some paperwork showing what our progress towards our mission is. Right.

We have, it’s incorrect to say open source, but we’ve made our source available for audit. Right.

Because again, maybe somebody wants to audit it, maybe most people don’t have the time to, but at least they need to know that it was available and nobody has caught anything bad.

In fact, at some point I announced a bounty where if you can find any instance where we have a left wing or right wing bias anywhere in our reporting, then you get $1,000.

Right?

Hasn’t been claimed yet, but it’s an open bounty. I never announced it ended, so I guess technically, yes.

Okay. Yes. Red Tory listeners, viewers here, here. First, if you’re good at detecting bias, go to otherweb.com right now. Sorry, please go ahead.

It’s been three years. Maybe I should have, but, but, but yeah. So you have to do these things. Right?

You have to build trust and some sort of a non standard way these days because there’s just been so many bad actors in this marketplace for so long.

Now you mentioned earlier we talked a lot from the user perspective, from the consumer perspective, but I want to hear your thoughts on the producers, on the news creators and how they’re making use of the platform. You mentioned you have corporate clients, but even just individuals.

Because I like what you said about making a newswire more accessible, so smaller outfits, you know, bootstrapped outlets have access to quality content. So give us a sense of, you know, culturally, you don’t have to get into specifics, but who’s using the platform in this way?

And maybe more importantly, who would you like to be using the platform kind of moving forward?

Let me try to reframe the question a little bit because I think we need to give the listeners some background on where things are going, please. We talk about what is happening in this kind of intermediate step.

And since we’re talking history, let’s be clear. Those people didn’t lose their jobs, they lost their land, which is probably even worse. But please continue. It was a Great Depression.

That’s what happened.

Sort of, right? No, I mean, not quite the same people. In most cases, the ones who owned land were a minority.

The majority were basically the people who live in the inner cities up north right now. And this is why they still speak with a southern accent.

Yes. Okay.

I acknowledge the racial dynamics there, but certainly here in Canada, where land was in abundance, if you look at the land registry records, it was very much losing their land because they all got land, they leveraged it to the bank, and then the revolution in farm productivity happened and those who were large enough to capitalize on it succeeded and those who didn’t lost their land for pennies. But your point is valid. Please continue.

I think you just avoided the post reconstruction run ta system that the US had and therefore you didn’t have that particular dynamic.

Agreed, but that’s where I alluded to the racial Dynamics here in Canada. Not to digress, we actually invented apartheid because of the way in which we treated our first nations, our indigenous populations.

South Africa then copied it and applied it to. So we’ve got our own guilt and our own need for truth and reconciliation here.

But this is two Jews kibitzing on how people have been oppressing people all around the world because we know that from our own culture. Please, we digress back to AI and content creation.

So hold this frame in mind, right? I think that the number of journalists that will have a job in journalism over the next 10 years is probably going to diminish by 90%.

At the same time, I think the ones who remain in the profession will become so much more effective that they will not produce 10x, they will produce 100x. Right? This is what we’re about to see. Now the question is, what will they produce?

They could produce the nastiest clickbait you can think of because they’re all in competition for eyeballs, because that’s how you show more ads and ads pay per click or per view. There’s no pay per truth or pay per quality or anything like that. Right? So it could go in that direction.

Or we could steer this ecosystem so that they are creating something that is truthful and high quality, because it’s actually easier to create that than to create clickbait, right? Until now, clickbait has obviously been easier, right?

This is why the average BuzzFeed journalist has been creating 10 times more stuff than the average New York Times journalist. Right?

It’s just.

It’s cheaper to create clickbait. But we want to get to a world where creating any content is relatively cheap.

And if that is the case, then you might as well create something good, right? Because you’ll just sleep better at night.

So that’s kind of the point of creating these intermediate tools, you would say, that don’t quite replace journalists. We don’t have the goal of doing that, but they assist journalists in just creating more high quality stuff faster.

And that means looking at what journalists are doing right now and trying to automate particular parts of it while realizing that some parts will take longer to automate. Right? I cannot automate a journalist badgering a politician over and over until the politician accidentally tells the truth.

That, for now, is a human thing. We can’t have AI do that. I cannot automate talking to Deep Throat in a garage in D.C. somewhere.

But we can automate checking something against other sources.

We can automate finding good citations and references for something that you Just found out from one source only so that you have more than one source citing. We can automate researching a topic. We can automate. Even writing first drafts is probably automatable.

Writing the final text, it’s okay in a blog post, probably not okay in a news article. Right? Which is why Autoblogger doesn’t have humans editing, but Press Hub does.

Even though both write, they do use different models because they write in different styles, but still one of those, you want to be an intermediate step. And Audioblogger, it’s still better than what corporations have been writing until now, so it’s good enough.

Now, you’ve mentioned a few times the kind of presence of government and of course within journalism and news, often it’s to hold government accountable.

And you said something that I unfortunately wholeheartedly agree with, which is that governments often don’t have the knowledge to really regulate this technology or facilitate this technology in a way that makes sense, that’s coherent. But journalism as an industry is really fundamentally tied to government, at least oppositionally, if not supportively.

Do you see a role for government in enabling, you know, this kind of future that you’re describing? Which again, I kind of agree with your analysis. I think we are going to see a radical reduction in journalists.

To your point about BuzzFeed, there is no BuzzFeed anymore.

So there is BuzzFeed, there is no BuzzFeed News.

Agreed. Fair enough.

We only killed one division.

Yeah. Yeah, well, I suspect the other one may not last longer either. But my, my point being, you know, to what extent could this be corrupted, right?

To go to the Pravda example, if there is only going to be a few journalists, is it possible that those journalists could be state supported or state sanctioned? And I don’t even mean that in a malicious or an evil way.

Like look at the British Broadcasting Corporation, right, the BBC, which even though it’s independent, it is still technically kind of a state based media. Is there a danger of that happening?

Or do you think that they’re so clueless about this technology that as long as we’ve got young people, as long as we’ve got curious people, there’ll be room for non state actors to be using this technology to provide those independent voices?

I think the danger always exists and I don’t think it’s so much government pressuring journalists directly because even if you go down from 100,000 to 10,000 journalists, it’s still pretty hard to organize that level of conspiracy, right? It’s much easier to go to everybody that has a base model and pressure them to embed a bias into the base model.

And then it would be so much harder for somebody like me fine tuning the base model to get the bias out of there because it’s already built into Llama, GPT and Plaud and all of these base models that are out there. So that would probably be the softer target for a government to try to influence.

But I think the bigger deal is that there is a big push in government to try to hold off this technology. Almost a laudate push. And it’s the wrong impulse, Right?

I understand the impulse, but I think in many ways it exists because unlike the people who lost their jobs in the 20s, the people who are about to a lot of lobbying power, right?

And I think we generally agree that the current state isn’t great, right? And the US trust in journalism is around 20%. So I think that we need to let it progress.

But there are better ways to regulate than to try to freeze things.

For example, one thing that I’ve been throwing out there quite a bit is if government were to mandate interoperability between all the big platforms, that would definitely help because that means that if anybody is better than others, it’s easier for them to convince people to move. And therefore everybody has an incentive to become better.

Whereas if everybody is locked into some sort of network effect now, suddenly it doesn’t matter who is better. Everybody can just show bad stuff and extract more value out of their users, right? So that’s an example of if government focused on that.

It would be great if government focused on actually addressing the problems of abusive competition and antitrust, right?

Instead of just blocking all mergers like we saw in the previous four years, what if they actually looked at, okay, somebody has monopolistic power, what bad stuff do they do with it? Let’s see if we can ban the bad stuff, right? That would actually help the market quite a bit.

Because what I see today in the technology markets, not just in journalism but everywhere, I see a lot of dumping and a lot of bundling. Two things that we’ve known for 100 years are bad. Right.

If you give away a browser for free just to lock everybody into your search engine, you are dumping this product to push your competitors out of the marketplace. Right. That’s generally an abusive, monopolistic practice. But instead government has been blocking mergers randomly.

That’s kind of the wrong target for antitrust.

And to your earlier point, they’re a little slow to the game, right. That they recognize these things are happening kind of after we’re all locked in.

And the lack of interoperability means it’s difficult for us to move. So one last question. What do you see, and you’ve alluded to it so far, but what do you see as the future of other webs?

Like what, what is your kind of ambition in terms of where, what role does it play within this larger media ecosystem?

Well, so I think that information is created, distributed, consumed and monetized in an ideal future. I want to nudge each of these four steps towards higher quality. Right. So we started from the consumption step. Right.

We’re now entering the creation step. Right. But I want to be on the monetization side too.

I want to give a data stream to every single advertiser that it should probably pay less for an ad that appears on bad content because that would create an incentive for the person creating the content to actually improve what they have. So that’s kind of the future I see. To grow as big as we can and influence as many people as possible.

Which is by the way, why it’s been three and a half years and other web is still free and ad free, it’s not normal. Right. But the moment we monetize it, we shrink our audience or at least our growth. Right.

And I want to reach as many people as possible and actually help them consume higher quality content. So hopefully we continue to grow. We’re at 18 million on the consumer side. Hopefully we’ll get to 50, 100.

At some point it will start affecting the ecosystem. So just like whole foods is 3 1/2% of the food market, but they sell organic stuff, therefore Walmart now sells organic stuff.

So we don’t need to be 100% of the market to have an effect across the board.

Right on. And where can our listeners learn more about you and your projects?

You can find me personally on LinkedIn or you can find my work on otherweb.com, getswarmer.com there’s a whole bunch of interesting other places you can find me or you can just email me@alexotherweb.com and tell me what you think. If you like what I’m doing, tell me. If you don’t like what I’m doing, then definitely tell me. I need to know.

Right on. I appreciate that attitude. Well, thank you very much. I really appreciate your time and your perspective.

I think that you’re doing really excellent work.

And the last point about trying to both incentivize the creation of quality content, but from a competitive perspective, influence everyone else so that they’re also motivated and incentivized to create good content.

I like to think that those of us in podcasting trying to create quality podcasting are part of that, and having you as a guest certainly helps make that possible because I think this has been a very quality conversation. So thank you very much, Alex.

And I’m certainly going to spend some more time looking@otherweb.com as a creator and think about the ways in which it can help us spread the word, especially local news, because I think that’s really important moving ahead. So thanks again, Alex.

Thank you so much.