Using Cloudflare? You might unwittingly be blocking your website from search

Episode transcript

Amelia: Hi, I'm Amelia.

Paul: I'm Paul.

Pat: Pat.

Russ: Russ.

Amelia: And this is Fin the Week. Welcome back to a brand new episode. How's everybody doing?

Paul: Really good. Another week. This one feels like it's gone quickly.

Russ: Yeah, good. I can't believe it's Thursday, really. It's flown.

Amelia: It rolls around, doesn't it? So, brand new episode. This week we're going to be taking a look at the thorny issue of AI training and the access that bots and crawlers have to websites. We're going to go straight into it. Paul, where do we start with this?

Paul: So this has all been prompted by an article that was published back in July by Cloudflare, and we'll get into what Cloudflare is in a minute.

On the face of it, AI crawlers and bots sounds quite dry and quite technical, but the more I've looked into this, the more interesting it is. I think it affects everybody who works in marketing and communications, because you're responsible for making sure information about your brand is accessible online. So this is definitely worth knowing about. It all starts with that article published by Cloudflare, called "New AI traffic options for all customers". Just a bit of context: according to W3Techs, Cloudflare is used by over 21.3% of all websites across the internet, as of January 2026. So it's a big deal. It's used by a lot of sites.

Essentially the article was about how Cloudflare categorises crawlers. When we talk about crawlers, we're talking about bots that access websites in order to do certain things. The most common one is probably Googlebot. Google sends this crawler out around the internet, it discovers websites and pieces of content, and then it indexes that content and so on. Cloudflare has simplified how it categorises the different types of crawlers it sees coming to websites.

So they've got search crawlers. That's any crawler that would come to your website, collect information and index it with the intention of eventually sending a real person to it in the future. Then you've got agents. They're automated crawlers, usually working on behalf of a real person in real time. For instance, you're using Claude and you've asked it to go and do a bit of research, or to look at your website and give its appraisal on something. That agent would be viewing your site in real time, so that's a bot accessing the website. And then you've got training crawlers: a crawler that visits a site and hoovers up information in order to either improve an AI's understanding or form part of its training model.

So those are the three categories Cloudflare has set. The article was saying that as of 15 September this year, they'll be setting new defaults for each of these three classifications. For any new site onboarding after that date, the training and agent categories will be blocked by default on pages that display ads. The reason it does that is that it says an ad is a signal that a website owner meant for a person to land there and see it. I suppose they're trying to protect publishers. If you're a publisher with articles and you're showing ads, they want to make sure training bots aren't coming and hoovering up your information so that it's displayed on another platform instead of people landing on your site.

So they're essentially locking things down. But crucially, the other thing they say is that multi-purpose crawlers — those that combine search crawling with training — will be allowed or blocked based on the most invasive part of their behaviour. That's important, because Googlebot, Applebot and Bingbot, the major search crawlers, are all, as far as we're aware, multi-purpose crawlers. Googlebot, for instance, is used for search indexing, but it's also used for gathering information for Google's AI tools, like AI Overviews in Google Search. So if you have set your site to block training crawlers, according to Cloudflare's new defaults you'll be blocking things like Googlebot.

That leaves you with a decision to make. Do I allow Googlebot to come to my site to index my content for search, and accept that it's going to be used for AI Overviews and other AI tools within Google Search? Or has the decision been made higher up the company to block training, in which case you're going to miss out on that search element? Ultimately, Cloudflare's play here is to encourage tech platforms to separate out their crawlers, so it's clear that this crawler is for search and this one is for training, and there's no mixed use.

There's a lot of technical information there. It's probably worth, at this stage, figuring out what Cloudflare actually is, because some of our listeners may have heard of it but might not know what it does.

Amelia: Yeah, I was going to ask that. Give us an overview of what Cloudflare is and what it's used for.

Paul: Pat, you're probably the best person to speak on that.

Pat: Yeah. Cloudflare is a proxy. Think of it in very simple terms as a collection of servers that you send all of your website traffic through. If someone types your website address into their browser and hits enter, the request hits Cloudflare first. Cloudflare then sends it on to your web server, your web server sends it back to Cloudflare, and Cloudflare sends it back to the visitor. That's a proxy: it intercepts traffic. What this allows Cloudflare to do is inspect that traffic and apply actions to it depending on what it looks like.

Cloudflare was originally built for cybersecurity, to prevent DDoS attacks. DDoS attacks aren't so severe these days, mainly because so many sites go through Cloudflare. A DDoS attack is where a hacker uses a network of compromised desktop machines, or other machines or websites, to inundate your website with loads and loads of requests, overwhelming the site and bringing it down. Once the site is down, the hacker reaches out to the website owner and demands a payment to turn off their army of bots.

Before Cloudflare, building architecture that could prevent a DDoS attack was incredibly expensive. It was only really available to huge enterprise-level businesses. So the team at Cloudflare came up with this platform: you route all your traffic through it, and they effectively provide enterprise-level DDoS protection to anyone who wants it. And for the most part it's completely free. You can sign up for Cloudflare and get DDoS protection and all sorts of other useful stuff free of charge.

There are obviously paid plans. There's a $20-a-month plan, which gives you loads more functionality, and there are enterprise plans, which get a lot more expensive. But most small websites are fine on the free plan. It's one of the few platforms out there that hasn't become enshittified over time. They're providing a really good free service and they haven't made it worse over the years. Hopefully it'll stay that way.

They're definitely one of the massive players in the landscape of the internet, alongside Amazon and Google, but you tend not to be aware of them unless you're in the tech world. They've basically got this gateway between the internet and your website that you can configure and control, which allows them to introduce features such as bot control.

Bots are becoming a real problem on the internet. You've obviously got the AI training agents that are scouring the internet, taking a lot of copyrighted content, using it to train models and surfacing that content. If you're a news site or a publisher, for example, that could be problematic. But on the flip side, if you're running a website and you're looking to market your products, you might want that AI agent scraping your content.

Then you've got this huge increase in bots across the internet just searching the web for vulnerabilities, looking for hacks, trying to log into sites. We've seen that across most of our websites — this massive increase in automated bot traffic, and it's becoming more and more sophisticated as well. You've also got a large increase in AI crawlers that aren't operated by the main providers. These are people just doing research. AI-driven business development is huge at the moment, and you've got all these people running Claude, Claude Cowork or Copilot, doing huge crawls of the internet trying to find leads. These bots are hitting all these websites and crawling the whole site to figure out what it's about.

So we're seeing a huge increase in traffic across all of our websites from bots, which increases the server capacity required but isn't generating any leads or any positive outcome at all. Bots have always been a problem on the internet, but they're much more of a problem now, and it's only going to get worse. Cloudflare is probably the only platform out there that can make a real difference.

Paul: This is probably the important part of the conversation, though. There's a real struggle here that people need to figure out, because you can quite quickly veer into the world of "bots are bad" — and, given the examples you've just used, they are. But there are also lots of crawlers and bots that serve a purpose.

One of the categories in Cloudflare's announcement was agent. That would pick up all those cases where people are using agents to do research on their behalf. Some of those will be small scale — an individual person doing a bit of research — and some will be massive scale, people really hoovering up information. So you've got to figure out how you allow a certain level of that but block the enterprise-scale agent use.

Training is a similar decision. One of the questions I was going to pose was: what do we think most people want? If you asked most businesses, "Do you want AI training agents to come to your website and use your information?", I don't have any data for it, but I'd guess most people would say no — that's our information. But if you actually think about it, in some cases you probably do want them to learn about you. So again, how do you figure that out?

I also listed out a set of common crawlers. I spoke about Googlebot, which exists to index information for search, and that includes Google Discover as well as Google Images, Google Video, Google News and various other platforms. You've then got individual crawlers, like Googlebot-Image and Googlebot-Video. These are all standard crawlers that go around the internet finding information.

Then you've got one called Google-Extended, which you can discourage from accessing your site, because that's used for training Gemini's models. A lot of people think, "OK, if I block Google-Extended, I'm not training their models, but Googlebot can still come and index my site." What people miss is that the standard Googlebot is used for the Google Search AI features. It's all a bit convoluted, but essentially you almost have to think of Gemini as a separate thing, whereas the AI Overviews you see in Google search results are still AI, still built on a model behind them, and that's Googlebot doing that.

Bingbot does a similar thing. And then, just a couple of examples from OpenAI and ChatGPT: you've got OAI-SearchBot, which is used a bit like Googlebot to gather information for search, so that if somebody uses ChatGPT to search for information, it comes back with information that search bot has found. But you've also got GPTBot, which is the training bot.

There are loads of these things, and even with the best will in the world, it's going to be very hard to block the ones you don't want and allow the ones you do. Even if you get that right, there are still going to be bots that do multiple things for multiple purposes.

I think the struggle businesses are going to have is that decisions about what should be allowed probably land with IT before they land with marketing. You might find that in a lot of organisations, IT have configured Cloudflare in such a way that, with this recent change, you're unwittingly blocking things like Googlebot, because IT — or even the board — decided "we don't want training bots coming to our site". Because of the way this has now changed, you could well be blocking stuff you do want. And how does that filter down to the marketing team, so they have a say in what gets indexed? It's a real minefield.

Amelia: I was going to say, it sounds complicated to me, but this is your bread and butter, and even for you it's not straightforward.

Paul: No.

Pat: Yeah, it's definitely a difficult one. The challenge is that there's so much value in content on the internet that these big firms will go to great lengths to access fresh content so they can use it to train their models. They all agree that AI is the future of their businesses, and if they're not able to train their models, their AIs are going to fall behind.

It's quite a similar scenario to the classic cookie pop-ups and cookie restrictions that Google have been battling with over the years. Google's primary revenue source is still advertising, and a lot of it is paid search. To track the impact of your paid search, you need to store a cookie — most of the time a third-party cookie. For years, various organisations have been trying to get rid of cookies or control them better, and all browsers apart from Chrome now block third-party cookies by default. However, Google have never implemented that, because cookies are so fundamental to their business model. So while they make all the right sounds and nod to better cookie regulation, they're still not making the decisions that would undermine their revenue.

Similarly, because accessing content for AI training is so valuable, I don't think Google will ever make it easy for someone to switch off AI training while still retaining all the value that appearing in Google provides. They easily could. They could give you an option: "We're not going to use your content to train Gemini models, but we will crawl your content and surface your website in our Gemini answers." But from what you said, Paul, they don't.

Paul: Well, it's a bit of a grey area. You've got that separate Google-Extended crawler that you can effectively opt out of if you put the technical settings in place. Then you've got standard Googlebot, which will crawl and also be used for AI.

But because of a UK regulatory requirement, Google Search Console has had to give website owners the option to manage their site's inclusion in search-related AI features, which is basically AI Overviews, AI Mode and generative AI features in Google Discover. So when you go to Google and see the AI Mode option where you can have a conversation with it, it's that stuff. In Google Search Console, the tool website owners use to manage how their site appears in search, there's a setting that lets you opt out of it.

It's at the global level, though. If you've only got Search Console set up for your overall website, you can only opt your whole site in or your whole site out. You can't choose a certain part of your site. There is a workaround: you can set up a Search Console property for part of your website. So you could set one up for /insights, the insights section of your website, and if you didn't want that included in the AI features, you could opt that out. But it's certainly not an obvious thing to do. Most people will have a property just for their overall domain, and you're opted in by default. Like I say, it was a regulatory requirement in the UK, but it's since been rolled out worldwide from the end of August.

It's a tricky one, because do you want to opt out of that stuff? Google is making it more and more the norm for people who use Google Search to interact with those features. If you search for anything now, chances are the first thing you'll see is the AI Overview section that encourages you to continue the conversation with it. So if you've opted out of that, that's probably a bad commercial decision.

So it's definitely a difficult thing to figure out. The reason I thought it was worthy of an episode was that I imagine there are lots of firms out there with competing departments with different views on this, who maybe haven't even thought about it that deeply. You have the people responsible for information security and IT, who are probably quite risk-averse and will lock things down as much as possible, so given the choice, they'll probably opt out of training. Then you've got compliance teams, who are also risk-averse, and again, if someone says to them, "We've got to make this security decision — do we opt in to AI training or opt out?", they'll probably opt out. I'm just guessing. And then you've got the marketing teams, who are responsible for getting information out there and would probably take a more nuanced view and say, "We need to still be findable online, so we need to opt in to this stuff."

I wouldn't be surprised if there are lots of organisations out there that just aren't managing this. There's not one person overseeing it all who has looked at the whole picture and said, "OK, what does our brand need to do?" So I wouldn't be surprised if there are companies that are inadvertently invisible online because they've got strict security policies.

Pat: Yeah. Think about the example of a news website. A successful news website, like The Guardian or The Telegraph, will have an IT team, a marketing team, and then the journalists themselves. The IT team will probably, like you say, be quite risk-averse. They'll want to block loads of bots, because that site will be absolutely hammered by bots all the time trying to scrape their stories. Then you have the marketing team, who are saying, "We need to rank number one for searches on breaking news about government policy changes," or something like that. And then you have the journalists, who are saying, "I don't want AI bots stealing my words and surfacing my thoughts and hard work in their AI answers without crediting me or the business I work for." So you've got competing agendas.

Perhaps the solution is using Cloudflare to filter which parts of your site agents are allowed to crawl, because you can do that. Rather than trusting Google to crawl certain pages and not others, or completely blocking Google from certain sections of the site using robots.txt or something like that, you make it a bit more nuanced.

Amelia: Is this a new problem? You were talking about these competing agendas when it comes to online content. Is this the first time we've seen this, or is it something that's always been a problem?

Pat: It's reared its head in the past. I remember Google used to surface playable videos in their search results from all sorts of different video platforms, and they were forced to stop that because people were just watching the videos in Google rather than going to the actual platform. It happened a few years ago with news as well, when Google were surfacing news stories directly on the Google page rather than sending people to the news websites. I think they had to stop doing that too — don't quote me on this — at least in certain countries.

But I think it's become a bigger problem ever since Google started introducing AI Overviews. Now they're doing it at a more wholesale level, by providing an AI Overview on whatever you search for. And we've seen that in a huge drop in the number of visits. You still get the search impressions, but most of our websites have seen a huge drop in actual traffic, because people are just getting the answer from the Google page. What did you call it, Paul? You wrote an article on it. The Great Divide or something?

Paul: It was the Great Decoupling. A few people came together on that as a description. If you've ever logged into Google Search Console, you'll see impressions and clicks on a graph, and they always followed each other: if one went up, the other went up. Then, when AI Overviews and other features became more prominent, they started to diverge. Impressions would climb, but clicks would drop off, and it was the first time that had ever happened. It was because AI Overviews were probably increasing the opportunities for a piece of content to get an impression, but that content was much less likely to be clicked, because it was just a reference. If a bit of content was clipped from a page and it just said it came from that page, people wouldn't really click through on it.

Pat: It's so prolific now that it's definitely not going to go away. So I feel like you need to be featuring in those overviews now. Otherwise you just won't get seen.

Paul: Yeah, and I think that's why this Cloudflare change is worth knowing about. They've got those three categories — search, agent, training — and you might think, "I'll just opt out of training," but you won't realise that opts you out of Googlebot, which is a multi-purpose crawler. According to their documentation, Cloudflare would automatically classify that as a training bot, because that's the most invasive part of what it does.

Pat: It's confusing, isn't it? Like you say, Paul, there are definitely going to be a lot of organisations out there with the wrong Cloudflare settings, because someone in IT has been tasked with configuring Cloudflare and they've just gone, "Well, I think we'll do that and that," and inadvertently blocked all the search crawlers. And the marketing team is scratching their heads as to why the traffic's dropped off.

Paul: Yeah. And the other thing that's quite telling: I logged into Cloudflare yesterday just to have a poke around and get my head around where it all lived, and it took me ages to find this setting. It actually lives under Security. There's a security section, you go in there, and a couple of clicks later you find those settings. That's interesting, because it is a security consideration, but it's also a huge visibility and communications setting.

That tells you how it could happen. It's a security setting, so it gets handed off to your IT security team, and they'll just lock it down — why wouldn't they? But there could be a whole other team that has no idea about this and is thinking, "Hang on a minute, why are we suddenly not getting visited by Google? Why are our pages not showing up in Google?" People will look at the common stuff, like robots.txt, where you can opt your site out of things, and say, "Well, that's all fine, our site's accessible, what's going on?" I wouldn't be surprised if there are lots of people in companies scratching their heads, and the root cause might be in there.

Amelia: Based on that, what should businesses be doing? Thinking about everything we've spoken about today, what should the takeaway be for people listening who are thinking, "What do we do next?"

Paul: The whole reason I wanted to raise it was purely to bring it to the forefront and make sure people understand that Cloudflare classifies bots the way it does and gives them a category. If a bot is multi-purpose, the strictest category applies. And there's a strong chance that if nobody in the marketing team has looked at it, the settings might not be optimal for keeping you visible online. So being aware of it is the first thing, and then dig a little and find out what your settings currently are.

Russ, you're on the call and you're probably slightly outside of this. My background is in online search and visibility, so my incentive is always to find a way to make a site visible. Pat, you've got a bit more background on the security side, so you might be stricter about locking down bots to make sure a site isn't hammered. But as someone who sits outside all of that, Russ, what's your take? Philosophically, what do you think a brand should default to? Should it lock things down or make the site available? Is it fair game that training bots come to our site and learn from us?

Russ: I think if the objective of the business is to rank better in search and bring in more traffic, because it brings leads into the business or increases the visibility of the business — and if we're talking about thought leadership, for example, which is an important part of most companies' marketing these days — then it would go against that if IT were saying the site needs to be locked down, because these bots then can't pull the information needed to rank the site higher in search.

So it depends on the objectives of each team in the business. Are we saying the website is less secure because it's being visited by bots? I don't think so. I think it's just that the traffic perhaps isn't a true representation of human traffic, but ultimately it is adding to the overall exposure and visibility of the business. So I'd probably say that locking it down — unless there was a good reason, like it making the site insecure — isn't in the best interests of a business that's trying to market itself, rank higher and get more visibility in search and in those AI summaries.

What's quite concerning is if an IT team didn't consider that, went in and changed these settings and locked the website down, and then rankings drop, and the marketing team is presenting a report to a client, benchmarking against competitors, and they're dropping off and not sure why. That's what's quite worrying in terms of the ecosystem required to keep these websites high up in view.

So I think it's probably down to someone sitting above both the marketing team and the IT team to be aware of it. I'm not sure where that sits — maybe at operations or C-suite level — someone who can orchestrate it all and make sure both teams know what this does. There's an education piece, isn't there? Otherwise it could have a real impact on the business.

And, slightly outside of those thoughts, we actually have our own scanning tool, which we've used on previous podcasts as well, and it got me thinking about how that appears. It scans a website, looks at whether it's accessible and conforms to the four outcomes of Consumer Duty, and where there might be gaps. But how does that appear in analytics? We're crawling these websites, and it actually goes to the website and accepts the cookie pop-up like a human would, and analyses the front end. So how many of these crawlers are out there that mimic the human visiting experience and maybe can't be distinguished from human traffic, as opposed to bots that don't go through that process of dismissing the cookie banner? So, two separate streams of thought on my side while I've been listening.

Pat: My view is that if you're putting something on the internet and it's publicly accessible, it's fair game. If you don't want something to be crawled, don't put it on the internet. If you don't want your staff emails to be crawled, don't put them on the internet. So my view is you should stop worrying about AI agents and bots crawling your site for information. Surrender to it. Just let them in.

And off the back of that — it's strange that these AI crawler settings are under Security within Cloudflare, because, like you say, they're more about visibility. Giving AI lots of access to your site doesn't mean you're opening up security holes. The security of your website is a separate consideration. You should definitely make sure your website's secure — that goes without saying. Make sure it's patched and up to date, and that you've got good coding guidelines, code review, a CI/CD deployment process, all that kind of stuff. Let the agents in, let the bots in, but filter what you put online. If you've got premium content you don't want scraped, put it behind a paywall or block it in robots.txt.

What you will see if you let all the bots in is a huge increase in traffic. A lot of that traffic will be people researching you as a potential client, and yes, there are millions of sales emails in my inbox every day, which is irritating. But similarly, you could have clients researching in depth for an agency that's a good fit for them, and we could appear in those results because of our website and because of the model's training. So that research swings both ways.

I think the actual challenge here isn't blocking bots. It's how you report as a marketer. How do you know the numbers coming through your analytics tools are an accurate indication of how many humans are visiting your site? The reporting tools have become really ropey recently. Google Analytics is kind of useless for actual numbers, because it's largely blocked these days unless someone clicks approve on a cookie pop-up, which only about 20% of people do.

We've started using Plausible quite a lot, and the paid Plausible plan has a database of known data centre IPs. It will just exclude IP addresses that come from a data centre, which are likely to be bots, and that seems reasonably effective at filtering out high-volume bot traffic. But if you're in Claude Code or Cowork and you're running a piece of research from your own machine, all the web requests come from your own machine as well, so that won't hit Plausible as a bot — it'll hit Plausible as a human.

So I think it's a losing battle to fight against bots. You can surrender and let them in, unless they're launching a proper nefarious, malicious attack on you, in which case Cloudflare has tools like Bot Fight Mode and Under Attack Mode that you can turn on to mitigate it. But those definitely shouldn't be the default. I think reporting is the challenge, and Paul, you're on the pointy end of this. I'm sure you can relate.

Paul: Yeah, it's probably an episode in itself. I've written before about GA4 being effectively useless. My general view on reporting is that you should think of it as trend-spotting rather than trying to find absolute numbers. People who are still seeking an exact number for how many visits they had last month — that's where you're fighting a losing battle. You just won't get that. Every tool will have its flaws. But what you can do is combine data from multiple sources and spot trends, and to me that's much more valuable. If you can see the direction of travel and you're aware of the flaws of these systems, that's what matters.

As for bot access, I agree. Hiding these new categorisations in the security settings seems a strange move to me, because it just isn't a security consideration. If you're talking about the security side of bot access, you're really talking about a site being hammered by an attack driven by bots. That's nothing to do with training versus agent versus search bots. That should be entirely separate.

Pat: Completely different.

Paul: Yeah. And I think this could catch a lot of people out.

Russ: It almost feels like traffic to a website is so hard to quantify now that it's more about the visibility of your brand when someone searches for you — whether that's in Claude, ChatGPT or Google Search. It's how often your brand appears when someone searches, rather than them going to your website to find you, if that makes sense.

Paul: Yeah. And there's another really interesting thing on that. Currently in Search Console, you can see which search queries people have typed in that generate an impression for your website. There are certain ways you can filter that data that show you what people are almost certain are things like ChatGPT carrying out searches in Google, and you can see what it's searched for and where you've shown up. I can't remember the exact string of text you need to use to filter this, but you can basically see where AI tools are carrying out searches on behalf of real people as part of generating an answer.

Pat: That's massive, isn't it? You want to appear in that. You still want to rank in the answers to those questions, because people are leaning on these tools so much for everything.

Paul: Yeah. So, wrapping all of this up, my main takeaway for anyone would be: just check Cloudflare. Chances are you're using it. Speak to whoever controls it and figure out what your settings are at the moment, because you might unwittingly be blocking something you actually want.

Amelia: There we go. I think that's a good place to leave it. Before we go, shall we play some Jargon Busters?

Pat: Yeah.

Russ: Yep, let's do it.

Amelia: You've had a pretty good run recently. If you've not caught this before, I have a list of industry terms, and every week I put the guys to the test to see if they know what they mean. This week's term is quantitative easing. What do we reckon?

Pat: I know what that is.

Russ: Quantitative... I can try and go first. I know in UX research you get qualitative, which is a subjective kind of evaluation, and quantitative, which is metrics. So quantitative easing would be seeing a diminishing chart on metrics or numbers — easing, not going up so quickly, perhaps. That's my thought. I'm not sure what segment we'd be talking about, though.

Paul: It's a term I've definitely heard a lot, and it was used a lot during the financial crash a number of years ago. Is it the process of effectively putting new money into the economy in order to control the direction of inflation? Something along those lines?

Pat: Yeah, it's basically governments creating money. They don't just magic money out of nowhere, though. What they do is issue bonds. During the financial crash, and during COVID, the government issued billions of pounds' worth of bonds. When a government sells a bond, someone buys it, and the government guarantees a fixed payment over a number of years in return. So if you buy a bond for £100, the government will pay you £5 a year for 25 years or something, and you get a bit more back at the end. What that does is give the government a sudden influx of instant cash that they can then distribute, and the effect of that is diluting the government's currency.

So it's controversial, because one could argue that by diluting the currency you're effectively taxing anyone who owns it — you're reducing the value of the currency they hold and claiming some of it back so you can do stuff with it. But it saved our bacon in the 2008 crisis, because without quantitative easing the whole world economy would have imploded. It saved our bacon in COVID, because without all that extra money to pay everyone's salaries, even though they weren't working, the economy would have imploded.

On the flip side, it's also used extensively to fund wars that countries otherwise might not have been able to carry out. Is it a good thing that a country that can't afford a war can effectively devalue its currency and use that money to go and fight a war anyway? And off the back of that, it's also one of the main reasons Bitcoin came about. One of the fundamental ideas behind Bitcoin is a currency that isn't controlled by a government and is completely immune to any form of quantitative easing. That was one of the benefits to society it was invented for. But I'm going to go off on a tangent in a minute, so I'll stop talking now.

Amelia: Well, that was an extensive answer!

Russ: Amelia, I really hope that's wrong.

Amelia: Well, actually, that was much more extensive than what I have here, which is: when a central bank buys government bonds to inject money into the economy. But it was good to get a bit more context there from Pat, so well done — absolutely right. We'll have more Jargon Busters next week, and we'll be back next week for a brand new episode, so we'll see you then. Have a great weekend.

Pat: Cheers, guys.

Russ: Thanks, everyone. See ya.

Paul: Cheers!