Ray Gillespie

Blog

  • Should I Be Worried About AI Taking Over?

    The recent explosion of interest in AI was triggered almost entirely by the public release of ChatGPT. While industry insiders had been excited about this for a while, not much detail was known about it until its public release in November 2022. That all changed as soon as regular users got hold of it, and the hype has pretty consistently been building since then, but is it something to be worried about?

    To answer that question, we need to break down a few concepts. Firstly, what is AI? Unfortunately, it’s a term that has many meanings to many people, and changes with context. Broadly speaking though, I like to think of AI as any code that is designed to creatively solve problems. Is a calculator AI? I’d say not, and that’s because it’s missing any kind of creative element. Is a search engine AI? Probably, at least in some situations – if the search engine just provides a list of links based on a basic index, then maybe not. But if a search engine looks through user behaviour and produces an optimised list that has taken account of this behaviour then that would be AI to me. Finally, what about ChatGPT? Absolutely – ChatGPT is pretty much a textbook definition of AI.

    So we can think about what is and isn’t AI, but that’s not the whole picture. When people think of AI from science fiction books and movies, they typically think of a (usually malevolent) sentient, artificial being – sometimes in a synthetic body (e.g. a Terminator), sometimes just running within a computer system (e.g. HAL from 2001 A Space Odyssey). Is that the same as ChatGPT? Thankfully not. ChatGPT is a form of AI known as ‘generative AI’, whereas HAL/Skynet et al are ‘artificial general intelligence’ (AGI). What’s the difference? Well, pretty much everything.

    Generative AI is designed to do one thing really well – generate content. Whether it’s text-based, image-based or video-based, generative AI takes existing content (loads and loads of it), and looks for patterns in that content, and attempts to reproduce them. For example, if you tell a generative AI model (known as ‘training’) to look at 100,000,000 photos of a fox, it will start to ‘learn’ that a fox usually has a bushy tail, two eyes and red/orange fur. Eventually, it will have learned so much about images of foxes that it can attempt to create its own – the more learning its done, the better the end result will be (on average).

    This is different to AGI. An AGI model attempts to replicate the general knowledge and abilities of a human mind, as opposed to just one thing, and this is the real trick that we’ve yet to master. We can create AI that is able to beat the best chess players in the world, and we can create realistic-looking fake videos of politicians doing silly things, but if you asked ChatGPT to operate custom-built hardware in a car factory it wouldn’t even know where to begin. AGI would, in theory, be capable of learning from any situation, and would have a human-like ability to apply knowledge to areas it has no direct experience in.

    When people talk about the dangers of AI, what they are typically referring to is the existential threat of AGI – if we succeed in creating AGI on a level that is above our own intelligence, what happens if that AGI decides to be antagonistic towards humans? If AGI becomes widespread enough and self-sustaining, would we even be able to stop it?

    These are the big questions, but fortunately for now they appear to be a long way off. ChatGPT and other generative similar AI models have opened people’s eyes to the possibilities of generative AI, but thankfully we are likely a long way off true AGI (though some people in the industry claim its only a couple of years away). However, we have seen such rapid progress in AI recently that we do need to start having some difficult conversations, as ultimately (as long as we don’t wipe ourselves out in the meantime) it’s inevitable that we will eventually create AI models that surpass the capabilities of the human mind in every way.

    The biggest question for me, is what happens when we create an AI model designed solely to convince us that it’s a real conscious being – would we ever be able to tell that apart from a genuinely conscious being, and if not, how do we know it’s not real?

  • Who Owns My Website Domain?

    It’s a question that doesn’t get asked often enough – just who does actually ‘own’ your domain name? The reason it’s seldom asked is that most of the time it doesn’t cause any problems – either you built your own website, in which case you probably registered the domain yourself, or you paid someone to do it, and they registered it for you – but what happens if something goes wrong, and how can you ensure you retain access to your domain?

    The first thing to note is that domains aren’t ever really ‘owned’, but they are ‘registered’, and that’s the key piece of information. Ultimately, whoever the domain is registered to has full rights to that domain, and no-one can take it from the registrant without their permission (with a few exceptions relating to crime/bankruptcy). What makes things difficult is that a domain can, in theory, be registered to ‘person A’, paid for by ‘Person B’, held in the account of ‘person C’, be controlled by the nameservers of ‘person D’, and pointed to the website of ‘person E’. That’s five different entities involved in the process, and only one of them ultimately has the rights to the domain. Let’s break down these stages so you can understand the process better.

    Registration

    As pointed out above, this is the really important stage. Whenever a domain is purchased, the contact details of the registrant are entered as part of the registration process. The details include the name, address and, arguably most importantly, the email address. I say ‘arguably’, as the name is probably the most important theoretically, as this is used to identify the owner in the event of a dispute, but in practical terms, the email address can be used automatically verify ownership and bypass the any disputes with the registering bodies (ICANN/Nominet). In any case, having the details of the registrant correct and up to date is absolutely vital, so if you ever use a third-party to buy or manage a domain (i.e. an agency or freelancer), make sure you ask them to confirm who the registrant of your domain is.

    Purchase

    This one can be confusing legally, as in theory you could buy a domain and pay to renew it for 20 years, but if you registered it in someone else’s name, they technically have the rights to it. Now, there are situations in which whoever paid for it may ultimately be granted access to it in the event of a dispute – if you ask an agency to register a domain for you but you never actually pay them for it, they would be within their rights to hold it back from you until payment has been received. However, even under those circumstances, if they have registered the domain to you there are ways you can take it away from them without them even knowing – if you buy a .co.uk/.uk domain, you can log in to the Nominet website, enter your email address and any domains registered to you will be available for your to manage – including the ability to change the domain’s TAG and transfer it to a different provider.

    Domain Hosting

    It can be odd to think of a domain as needing hosting, as typically it’s the website itself that needs a host server, and the domain just sits in a website portal somewhere and is pointed to the host server. However, the domain itself is still held within a hosting environment of sorts, and that’s what the ‘domain host’ or ‘domain provider’ is. This doesn’t normally cost anything like as much money as a proper web server hosting package, and typically the costs are around £10 / £15 per year for standard domains. Examples of a domain hosting provider would be GoDaddy, 20i, Heart Internet, Namesco etc. Each of these providers has their own domain ‘TAG’, and when a domain is registered, the TAG is set to whichever provider will be hosting the domain for you. For example, 20i’s TAG is EXTEND. If I want to host my domain with 20i, I just set the TAG to EXTEND and ask 20i to migrate it into my account. However, just because it’s in my account doesn’t mean I am the registrant, so be careful to check this for any domains you think you own.

    Nameservers

    A step down from the TAG are the ‘nameservers’ of a domain. In addition to having their own TAG, domain providers usually have their own nameservers, and these are used to host the ‘DNS records’ of a domain. Without going into too much detail, the nameservers are effectively the broad level controls of where a domain is pointed to, whereas the DNS records are the detailed control on a level below this. Most of the time, if you want full control of where your domain is pointed (i.e. which web/email host servers) to, then you will want to keep the nameservers set to whichever company you have an account with. This enables you to control every aspect of that domain’s usage, such a which email provider it works with, and which host/server the website itself is held on.

    DNS Records

    As mentioned above, the DNS records are used to ‘point’ a domain to a particular server. Each service (e.g. email, website) has its own DNS record or set of DNS records, and changing these allows you to control where that service is set. They normally take the form of an ‘A’ record, ‘CNAME’ record or ‘TXT’ record, though there are quite a few others too. Each record has an address (either a web address or an IP address), and your email/web hosting provider will provide a list of DNS records for you to use in order for their service to become active. It’s a job best left for experts, as if you make a mistake your entire website/email service can stop working, but once set you rarely need to change any of these until you come to change your web/email hosts further down the line.

    Website Hosting

    Even after all that, there’s still sometimes confusion between hosting a domain and hosting a website on that domain. You may well end up in a situation where someone else hosts your website for you, but that doesn’t necessarily mean they have any control over the domain or where it’s pointing to. This can often be a problem when you have been with a provider for several years, and can’t remember the detail of how the website was set live in the first place – there are normally three situations:

    1. You control the domain via the nameservers, and you pointed the relevant DNS records to the web server yourself.
    2. You host the domain in your own account but you set the nameservers to those of a third-party provider and they set the DNS records to point to the web server.
    3. Someone else registered and hosts the domain for you (perhaps a previous provider) and they pointed either the nameservers or the DNS records at the new web server.

    In the event of any confusion, you may need to go back through some old emails to find out what you did at the time. Fortunately, there are various lookup services that you can use to find out when a domain was registered/updated, which provider’s TAG/nameservers it’s using, and even what server the various services are hosted on – it’s best to ask an expert to do this for you, as they can typically dig into more detailed lookups, but you can also get some top level information just by entering the domain in a website such as https://who.is/.

    Ultimately, what seems like a simple question of ‘who owns my domain’ can become a real journey down a rabbit hole, but if you plan things correctly and ask the right questions there’s nothing to worry about – just make sure you keep notes on whatever you or anyone does on your behalf for future reference.

  • Microsoft Will Pay You to Use Their Edge Browser

    Microsoft has faced significant challenges with its browsers over the past few years, as Firefox and Chrome have rapidly gained users from the Microsoft base. Despite the improvements in recent versions of Internet Explorer, which have largely overcome the issues of earlier versions like IE6-9, their latest browser, Edge, is struggling on Windows 10.

    Edge was introduced as an alternative to IE11, marketed as faster and more responsive. However, many users fail to notice significant improvements. The only visible difference is a slightly altered user interface, which requires some adjustment. This issue is compounded by Windows 10’s inconsistent mix of interface styles, including fullscreen and desktop apps, making it harder for users to embrace Edge. As a result, many continue to use the classic Internet Explorer or turn to third-party browsers.

    In an effort to boost Edge’s usage, Microsoft has begun offering reward points to users who switch to Edge. These points can be redeemed for vouchers or credits, but there’s a catch—users must also use Bing as their search engine, which has been unpopular, especially in Europe. Currently, the rewards scheme is only available in the US, and there doesn’t seem to be much enthusiasm for it outside the region.

    For more details click here.

  • #Friendmageddon Coming Soon to a Facebook News Feed Near You

    Facebook’s original news feed was designed to show users interesting stories and updates from their friends. However, over time, as businesses and organizations have increasingly turned to Facebook for its marketing potential, the feed has become flooded with promotional content. According to Facebook, this has led to users seeing more promotional posts and fewer updates from their friends and family. As a result, Facebook has announced a significant update to address this.

    In the near future (though the exact timeline hasn’t been released), posts from Pages will be shown less frequently in the news feed. This change means that running marketing campaigns through Facebook Pages will become more challenging, as the organic reach of posts will continue to decline. In the past, similar changes have led to a 40% drop in audience reach, and this trend is expected to worsen. This could also be seen as a strategy to push advertisers toward Facebook’s paid advertising platform, which may become more appealing as the organic reach decreases.

    For personal users, the impact may be less noticeable—while promotional posts can be annoying, users can simply unfollow Pages that post too much unwanted content. Paid and boosted posts, however, remain unavoidable and often the most disruptive.

    For Page owners, this is more bad news. Their organic reach will be further diminished, which means they’ll likely need to invest more in paid advertising to maintain the same audience reach. While there could be some hidden benefits, such as potentially higher engagement due to less promotional content in the feed, the update seems to make life harder for social media marketers.

  • The UK’s Most Expensive AdWords

    If you’re using Google Ads (formerly Google AdWords), you’re likely spending anywhere from 20p to £5 per click, with most of our clients’ accounts falling within that cost range. While these typical cost-per-click (CPC) prices can add up, they’re still relatively modest compared to the sky-high costs for keywords in industries like online gambling, technology, and financial services.

    Take a look at some of these eye-watering CPCs:

    • “What is invoice factoring” – average CPC £93.23
    • “Cloud services UK” – average CPC £95.80
    • “FTSE day trader” – average CPC £112.69
    • “Play live blackjack” – average CPC £148.51

    Just one click on an ad triggered by the phrase “play live blackjack” can cost the advertiser nearly £150. Given that only a portion of these clicks will actually convert into paying customers, it’s clear that online casinos must make substantial profits per customer to justify these high costs.

    For those interested in more keyword data, you can check out the full list of the top 500 keywords compiled in a recent research report here.

  • Twitter Moves Into Broadcasting

    The world used to be a simpler place. Major television networks controlled what we watched, and major newspaper and media outlets controlled what we read. Then, the internet happened, and everything changed. Power and influence shifted from the few to the many, and suddenly anyone with a keyboard could become a journalist, and anyone with a camera could be a broadcaster.

    A key player in this transformation was Twitter, which gave anyone and everyone a voice, albeit in just 140 characters. Over the years, many social media platforms came and went, or evolved in significant ways. Yet, Twitter remained largely unchanged—aside from ditching its original, much-derided logo.

    Now, Twitter might have just made its most significant move since its inception ten years ago. Despite being primarily a micro-blogging platform, Twitter announced that it has secured the rights to stream Thursday night NFL games. What’s interesting is that Twitter has landed this deal for a fraction of its market value—paying $10 million for a ten-game deal, whereas Yahoo! paid $17 million for a single game.

    The big question now is how this broadcasting will play out—will people watch the games directly on Twitter’s website, or will the company create a new platform for the content? We’ll find out in the coming weeks.

  • Apple, Google, Microsoft Logos – See How They Started

    Every now and then, a big brand updates its logo. Initially, the new design looks strange, but after a few days, it starts to feel normal. Eventually, you completely forget what the old logo even looked like. That is, unless you happen to stumble upon the shocking early logos of companies like Apple and Google, which they likely hope you’ll forget!

    Luckily, a clever website called Evolve has preserved some of the earliest logos from a variety of modern companies. It allows you to hover over the logos to see the transition from the old design to the current one. It’s a fun way to track how some iconic brands have evolved over the years. Check it out here!

  • Adobe Throws in the Towel for Flash

    Poor old Flash has had a rough time lately, with criticism coming from all directions. In a long overdue move, Adobe has finally admitted that it’s probably time to move away from Flash. In a typically verbose blog post, Adobe shared its new strategy for web animation tools, announcing that they would be dropping the “Flash” name from their Creative Cloud suite. The animation software will now be called “Animate CC” instead of “Flash Professional CC.” Even more significantly, Adobe is now “encouraging content creators to build with new web standards,” which is as close as they’ve come to acknowledging that using Flash is not ideal for modern developers.

    While Flash hasn’t been all bad, it’s clear that its time has passed. Sure, Flash was notorious for draining battery life (a major issue in the age of smartphones), introducing security vulnerabilities, and unnecessarily increasing load times. But there were some fun moments, too. Flash gave us animation gems like Magical Trevor and Line Rider, which became iconic distractions for office workers around the world. Flash also helped usher in a new era of web development, moving away from static, table-based HTML websites to more dynamic, engaging web experiences.

    However, Flash is now living on borrowed time. The first major blow came with Apple’s decision not to support Flash on its iOS devices, a bold move that turned out to be a pivotal factor in Flash’s demise. As major websites and platforms started ditching Flash, the writing was on the wall. Now, with Adobe’s official announcement, even the last defenders of Flash have disappeared.

    So, to mark the end of an era, I thought I’d share a classic Flash animation, now available in glorious HTML5 on YouTube. RIP Flash, and thanks for the good times:

    Watch here

  • Water on Mars

    There’s a very cool and interesting article on The Verge discussing the implications of discovering liquid, flowing water on Mars. While many people—myself included—are excited about this discovery, the article highlights the massive obstacles that still need to be overcome before we can do anything with this knowledge. However, the exciting part is that these are relatively short-term, solvable problems, and they are likely the only things standing between us and what many consider to be one of the most important discoveries in history.

    What makes the discovery of liquid water on Mars so significant, compared to finding it 20 kilometers below the icecaps of Europa, is that there is a genuine possibility of accessing this water within the lifetime of many people alive today. Once we can test the water, assuming contamination can be avoided, we’ll face a major crossroads in our exploration of the Red Planet.

    Given how resilient life on Earth has proven to be—imagine surviving at the bottom of a boiling sulfuric volcano in pitch-black conditions—it’s highly plausible that some form of life could exist in Mars’ water. The results of this research could go one of two ways: we could either discover extraterrestrial life, or we might not. (There’s also the terrifying possibility of inconclusive results due to contamination, but let’s not think about that yet.)

    If no life is found, we continue our search elsewhere, but if we do discover life from another world, it would be an event that could upend history, religion, and human belief systems. Imagine how many of our questions about the universe would be altered in an instant. The discovery of alien life would dwarf questions like, “Did God create the world?” or “Did life start in a primordial soup?” These questions, while significant, would suddenly seem small in comparison.

    Ultimately, the discovery of liquid water on Mars has put humanity at the gateway of a discovery so monumental that it’s hard to put into words how significant it would be. It’s like the first time a curious primate picked up a branch and thought of using it as a club—he had no idea where this newfound ability would take him, but he knew something big was coming. That’s where humanity stands now. We may be on the cusp of a discovery beyond our imagination, and that uncertainty only fuels the drive to explore. Now we just need to wait—and hope it doesn’t take too long.

  • Music Company YouTube Takedown Denied

    Copyright content in YouTube videos has become a contentious issue over recent years, with a delicate truce between rights holders, YouTube, and content creators that seems to work reasonably well. Currently, when you upload a video containing copyrighted material, YouTube has two main options. It will either automatically monetize the video with ads, directing the proceeds to the copyright holder, or it will remove the video entirely.

    The first option allows the uploader to showcase their video without censorship while the copyright holder makes some revenue, which appears to be a win-win situation. However, the second option means the uploader loses their video, and the rights holder doesn’t earn anything further. While this might seem counterproductive, it still happens occasionally, especially with certain artists and labels taking a very firm stance on their content appearing on YouTube.

    An example of this is a video featuring a 13-month-old toddler dancing to Prince’s “Let’s Go Crazy,” where the record label issued an immediate take-down request because the video included 29 seconds of copyrighted music in the background. However, a Ninth U.S. Circuit Court of Appeals in San Francisco upheld the federal judge’s decision to dismiss the record company’s claim, ruling that it constitutes fair use of the music.

    According to one of the legal team members involved, the ruling “sends a strong message that copyright law does not authorize thoughtless censorship of lawful speech.”

    I have to say, I agree with them.