You can find the podcast and other links here.

This podcast is recorded in French, but using something like a Whisper model, you can get a transcript and then translate it into English or another language reasonably quickly and easily.

Show notes and topics discussed:

French

• La dépendance au cloud pose des questions de résilience.
• Un mix entre solutions locales et cloud est nécessaire.
• L’intelligence artificielle est mal nommée et mal comprise.
• Intel doit s’adapter aux nouvelles réalités du marché.
• La spécialisation est essentielle dans l’industrie technologique.
• La gouvernance d’Internet est en évolution constante.
• Les libertés d’expression numériques varient d’un pays à l’autre.
• Les technologies doivent être évaluées pour leur impact sociétal.
• Les modèles d’affaires doivent évoluer avec la complexité technologique.

English

• Cloud dependence raises questions of resilience.
• A mix between local and cloud solutions is necessary.
• Artificial intelligence is misnamed and misunderstood.
• Intel must adapt to new market realities.
• Specialisation is essential in the technology industry.
• Internet governance is constantly evolving.
• Digital freedoms of expression vary from one country to another.
• Technologies must be evaluated for their societal impact.
• Business models must evolve with technological complexity.

My thanks, as always, to my friend Jean-François Nantel and to Éric L’Heureux for hosting this podcast.

It has been a challenging week and not one conducive to reflection or giving me time to think about what I want to write about here. But as this is an exercise in getting me to write regularly, as much as it is a space for me to think about tech and how it relates to our lives, particularly here in the Caribbean.

If you haven’t been following or don’t get to see news from the French West Indies, we’ve been having a week of unrest and violent protests on the Island after a number of peaceful protests seem to stir up a general malaise in the population, which seems to have been hijacked by a small group of well-organised individuals that set about looting, pillaging and burning shops, cars and just about anything that could be burnt. Roadblocks were set up to disrupt circulation, as is always the mode operator here when there is a protest, and several deaths were recorded.

As its relation to tech? I’ve been surprised. When things started to get out of hand overnight, I was very early in calling a curfew; note: I said “calling”, not “calling for”. I knew the authorities (the Prefecture) would instigate one, and as I write this, an overnight curfew is still in place (Edit: The curfew has been extended until the 21st). At the same time, I also said that the government would block TikTok, as it had shamefully done so during the unrest in New Caledonia. I was wrong. To my knowledge, TikTok was left to function normally and was instrumental in communicating the main revendications of the now popularised protests. I’m not sure why. 🤷‍♂️ The French government didn’t hesitate in New Caledonia, so why did they refrain in Martinique?

I’d like to believe it was because they got a lot of criticism from organisations like Access Now and other human rights organisations for denying access to these platforms. But I’m not convinced that is why, and I’d be fascinated to know the real reason. To be fair, I think it was absolutely on the cards at least at one point, but as things started to simmer down, it was perhaps deemed unnecessary. I still believe the French government will use this tool as it sees fit if things don’t improve. When you’ve done it once, the second time is always easier.

Negotiations for the main requests are due to restart tomorrow (Tuesday), so we have a little respite today, but I have a feeling that things will get a little tense again tonight. We’ll see.

It has also shown that governments are poor at understanding and interpreting platforms like TikTok and others, leaving them entirely flat-footed and with no meaningful response to today’s complex situations. Technology amplifies anything that it touches, and as technology gets further and further into our lives and society, it will amplify anything it touches, good or bad. This is not to exonerate technology as benevolently neutral and therefore devoid of blame. No. Humans make technology, and humans bring all their biases and shortcomings to it. Those amplified biases can be disastrous on a much bigger scale.

An example I gave last week in my rant, and probably many times before, is Facebook’s direct involvement in a genocide that took place in Myanmar. It is something that I wanted to write about a long time ago but I recently discovered this feature-length series about Facebook’s involvement. I urge you to read it.

I have another project dear to my heart that I just can’t seem to kick off yet. It’s about the Amiga computer. One day. One day.

France’s decision to intentionally block a platform shows us that no democratic government is above using draconian and rights-limiting powers to enforce its own policy, good or bad. Some are more willing than others, and each time it happens, it just emboldens others to try.

This begs the question about how it can be avoided and what we can do about it. Given that it is a highly complicated subject, as we can see from the EUs attempts at breaking open encryption in messaging (bad!), arguing one side or another only leads to whataboutisms and evermore polarised positioning.

I can only speak from a personal perspective, and my feeling is that basic privacy between individuals and small groups should be preserved and enshrined in law. Mass communications should be treated differently, in how monopolies are treated differently than smaller-sized businesses. This is because the incentives change with scale.

Distributed social media, like Mastodon may be part of the solution, as its moderation model (where instance holders can decide on what its users can and cannot do or say locally, all the while being subject to federation or not by the outside world) are built-in to the system. I say ‘like’ as I’m not entirely convinced that this is the perfect model either, but it is what we currently have at our disposal. Essentially, these are small local islands of “like-minded” users that can communicate locally and across the entire platform if they are connected and accepted (federated). If you, as an instance holder, can decide to federate or not, and crucially, you can choose to defederate if needed. It doesn’t require a little of the “who polices the police” in moderation policies and implementation. But as I’ve discussed previously, this is work already being done and is publicly available. Many of these operate a sort of small-scale multistakeholder bottom-up governance model, too.

So go and fire up your instance of Mastodon or another Activity Pub system and use the guides I’ve linked to as a starting point.

I recently asked on a region-wide forum if there were any known instances of Mastodon in the Caribbean. I haven’t heard of one yet, but they might be out there, so if you do know of one, let me know.

Reading

Speaking of tech speeding things up and making things easier, it should come as no surprise that the dark side of the web is also benefiting from the AI “revolution” (snarky quotes, btw). LLMs are supercharging Pig Butchering schemes.

As this Wired article outlines:

In addition to buying written scripts to use with potential victims or relying on templates for malicious websites, attackers have increasingly been leaning on generative AI platforms to create communication content in multiple languages and deepfake generators that can create photos or even video of nonexistent people to show victims and enhance verisimilitude.

Google’s 2.5 billion Gmail users are at severe risk of being hacked:

Sam Mitrovic, a Microsoft solutions consultant, has issued a warning after almost falling victim to what is described as a “super realistic AI scam call” capable of tricking even the most experienced of users.

On a subject that is dear to my heart —Internet governance— an unexpected decision by the U.K. government to cede authority of a small group of islands called the Chagos Archipelago to Mauritius has caused consternation in the world of DNS. If you remember, Domain Name System (DNS), is one of those essential backbone services of the Internet that without, things would be much more frustrating, especially as we move into the IPv6 era.

The islands in question were given the ccLTD of .io (ccTLD = Country Code Top Level Domain, think .uk, .fr, .bb, and that money-spinner .ai). It turned out to be a money-spinner like the .ai domain, only many years earlier, and .io was the darling of the tech startup world. If you didn’t have a .io, you were nobody.

Just as an aside, .io evokes a computer processing idea of Input/Output, I/O for short.

As nerdy as it was, it was the domain to own for a while, and .io was given to the islands because they were the Indian Ocean territories of the U.K. A 2021 UN decision ruled that the U.K. has no sovereignty over the islands, and hence, the process to relinquish started and is now complete.

The question is, what will happen to that domain now that the Indian Ocean territories as an ISO country code will no longer exist? Not so fast! The ISO-3166 code may disappear, but no timetable has been set. This would also be independent of any decision by IANA (the ultimate decision-makers on the ccTLD) to retire the corresponding domain. They could easily decide to make an exception, as already done in the case of .su.

As for now, all .io domains will remain active, and renewals should happen without interruption of service.

Absolutely no LLMs were harmed during the writing of this piece. Have a great week.

I don’t know. I must have gotten up in a bad mood or something. This is to announce the colour of this issue 😉. These are a few thoughts and are certainly not fully thought out, but I thought I’d put them out there for discussion anyway. 🤷‍♂️

There’s something seriously wrong with Big Tech at the moment. From the proclamations of Mark Zuckerberg (the laughably pathetic Julius Caesar cosplayer) that he no longer should care what he says and does to something like half of the world’s population through his privacy-ignoring “social” media project to Eric Schmidt’s declaration that climate change is bad, but a parlour-game guess-the-next-word text generator will solve it, to Sam Altman’s insistence that his particular brand of LLM is going to “solve all physics”, not that he knows anything about physics of course, and who couldn’t ignore Elon Musk’s Ketamine-fueled descent into fascism and continual abject stupidity.

I could go on more, but it’s just too depressing to hear these people talk this way, with absolutely no responsibility and absolutely no pushback from the media or the business world. They’re held up as “geniuses” or “business gurus”, but they have collectively enabled vast wealth-extraction machines that they are the sole beneficiaries of, marketing poor-quality products and services in most cases.

I’ve been trying to understand why this is happening now and why it didn’t happen before in earlier periods of tech. I have no evidence, but I have experience and a bit of spidey-sense, for what it is worth.

I think it just boils down to the fact that previously tech was confined to silos of some areas and departments of business and a bit of exposure to the general public, where some services were made available, and a few things were made easier. That is no longer the case. Tech has butted up against the real hard stuff of the social makeup, empathy, perception, knowledge and the general humanity of the world. It is now confronted with that hard wall of governments and societies, with their inherent incomprehensible complexity. Throwing logic gates at that scale of the problem (if this, then that) just doesn’t work anymore. The fact is that exceptions are the rule, and the basics are the exception. Virtually everything we do in life is a massively complex set of operations that sometimes follows logic but mostly doesn’t. And Intel or ARM processors with a stack of OS and software piled on top of them are just not enough to work this stuff out. It can’t, and it most likely never will.

When the tech was being democratised and available to the public through sites like the Amazon bookseller of the time, it was simple. Books were listed on a webpage, and you could purchase one with a simple click, paying with your bank/credit card details. A few days later, the book (a physical thing) was shipped and in your hands. That was it. The only amazing thing about it at the time was that the catalogue was immense and bigger than any physical bookstore. Due to how databases work, you could easily and quickly add a new book for sale. This was used to continuously evolve the catalogue from the distributor’s and publisher’s lists of titles and the requests users made when searching for a book by title, author or ISBN. If the book wasn’t in the catalogue and appeared in the search history of the site frequently, a couple of clicks later, it would become available.

It ended up mostly killing bookshops. Niche shops exist, but the industry of buying and selling books has changed forever. Digital distribution didn’t help, of course, but it was the online sales methodology that killed bookstores.

But I digress.

Today, technology is immersed in virtually every aspect of our lives, but it is no longer a benevolent force for the most part when looking at the most popular platforms in use around the world. And I think this results from tech becoming more dehumanising, reducing us all to units of MAUs, ASPs, CSATs and other ultimately useless indicators (KPIs, if you will).

I also think that this will not last. And I think there will be some kind of reset, but I don’t know when or what that will look like. But if history is anything to go by, I don’t think it will be done through negotiations around a table of stakeholders calmly negotiating on what we all want for the best. There are too many stakeholders and too many diverging opinions that something will have to give. I don’t look forward to the future of tech as it intrudes on my/our personal space and encroaches on my/our private life, following me everywhere like a vile little stalker. Did you get a robot vacuum cleaner? Great, now that maker has floor plans of your home, as well as video and sound recordings (possibly) of the most private conversations. It has sold them to advertising companies with little to no scruples about what these other businesses will do with that data. They have also sold that data to the police so that when they come knocking on your door for that benign transgression of taking off a micro-second before the green light at the late-night deserted crossroads, they’ll know where to shoot first because you’ll likely be in the bedroom or the living that has been pre-identified by the crappy dust mover, replete with information on all the furniture in the room. (Think this isn’t feasible? Let’s talk.)

This will not end well.

I mentioned about the Social Web. A term I’m happy with using, but there has been some development in that area. A foundation has been set up, but unfortunately, it is cohabited by undesirable companies such as Meta. If you want to know what is happening, I’ll let you search and look for yourselves. Still, I am of the opinion that it is incredibly naive to believe that Meta has good intentions in their participation. I believe that they will do whatever they can to destroy Mastodon as it stands as a platform today (no advertising and no useless engagement algorithms) and the decentralised social media sphere for their own gain at the expense of the world at the slightest hint of danger to their surveillance operation. Their participation will be nothing more than an embrace, extend, and extinguish policy popularised by Microsoft a couple of decades ago.

Contrary to what you might think my position is, I do feel Meta should have a seat at the table, but it is clear to me that their inclusion should be restrained precisely because they are the single biggest player on the Internet. They wield too much power and power beyond anything that a small piss-pot association or foundation can have. As such, they absolutely will use it if they feel the slightest bit threatened. Anything they do or say should be scrutinised, and they should have absolutely no power to push through any decision unilaterally. One seat, one vote, and not an iota more.

Reading

Here is a quick list of the things that caught my attention this week:

Think your small Caribbean island is not a target for cybersecurity attacks? Think Again. The Barbados Revenue Authority had 230GB of data exfiltrated, and it was put up for sale on the Internet. Theories are being bandied around, but I don’t think it is helpful to discuss that here.

The Walrus has a good article expressing and discussing the way facial recognition technology is not fit for purpose. It opens with this line:

THE FIRST TIME Karl Ricanek was stopped by police for “driving while Black” was in the summer of 1995.

Another article from the same publication does a good job of expressing some of what I’m thinking about social networks in an article entitled, The Collapse of Self-Worth in the Digital Age.

I hope you had a great week, and I wish you an excellent week ahead.

Good morning. I hope you’re doing well. I’m writing this at the end of a busy week that hasn’t left me much time to think about a specific subject, so this will probably be both short and a little unstructured.

After a lull in activity, it seems the Atlantic is firing up for a busy end-of-season. To remind you, Hurricane season is from June to November, with a peak in early September and a spike in October. I’m looking forward to the end of this season.

Source: NOAA - 30th September 2024

One of the things that I have prioritised over the last few years is the importance of reading articles over short posts on Social Media. I’m not particularly active on Social Media; in fact, I only have three accounts, and only one of them is on a major platform. I have a “professional” account on LinkedIn, where I cross-post this article from time to time (I’ve automated it now, so it should be more frequent), and I have an account on Mastodon and one on Bluesky.

I tend to use these latter accounts for a little news and a little research. By curating a list of things I’m interested in, I avoid the usual traps of gamed social media that optimises for ‘engagement’ (i.e., advertising) over real connection. My timeline tends to contain research papers and news articles from major and minor publishers. The downside is that I have a never-ending mass of documents and articles to read, which is impossible to do in the time I have.

To give you an example, here’s a list of the articles and documents I set aside to read in the coming days and weeks:

CYBER-ACTIVISM WITHIN THE GLOBAL DIGITAL DIVIDE: A CASE STUDY OF PERU - This Master thesis is from 2014. I’m interested in how we used to think about these topics a decade ago, and I’m wondering if things have changed and, if so, how.

Towards a Digital Politics of Multiplicity: Social Media Networks and Global Justice Politics - A paper from 2021 to help me think about Social Media’s positive aspects. I’m a little net negative on social media. Having another perspective helps calibrate my thoughts.

THE CYBERSPACE MYTH AND POLITICAL COMMUNICATION, WITHIN THE LIMITS OF NETOCRACY - Another old-ish paper (2017) in Internet terms, but another one delving into the history of the Internet and the implications of its self-built myths.

The Politics of Cyberconflict - This is a long one, and I’m not sure I will get to read it all. But the title interested me, so I added it to the list. It’s another older paper from 2004, but again, it is helpful to understand how we thought of the Internet 10 to 20 years ago.

This is in addition to all the news articles and blog posts I’ve bookmarked for further reading:

‘We will coup whoever we want!’: the unbearable hubris of Musk and the billionaire tech bros - The Guardian

An interesting analysis of fair use and generative models – Baldur Bjarnason

Hacker plants false memories in ChatGPT to steal user data in perpetuity - Ars Technica

Signal’s Meredith Whittaker: ‘I see AI as born out of surveillance’ - Financial Times

In other news concerning the Internet, the Web Foundation has announced that it has shut down. The foundation was the brainchild of none other than Tim Berners-Lee, the inventor of hypertext and the web as we know it now. The foundation championed an inter-connected and decentralised web, only to have it steamrolled by the big social media companies with destructive business models. It is not the end for Tim. He’s going to concentrate on the Solid Protocol. I haven’t looked into it yet, so I don’t have an opinion. It’s on the list.

It’s probably a good point to remind you that, no, the free market didn’t innovate the Internet. It is entirely a project brought about by public funding, from the networks to the applications running on top.

Of note

I’ve started messing around with the site that hosts this text to try to customise it to my liking. I’ve made a couple of tweaks, and I hope I can do more in the future.

In order, I have modified the section divider in the text of these articles to show three stars (only on the website for the moment). I’ve always liked that look, but I couldn’t figure out how to modify the code to do that easily, so I gave up and prioritised other things. This weekend, I decided to make an effort and figured it out with some help from people on the forums. Thank you to them.

I renamed the Reading section “Working Library.” I like this term better because it better reflects what I’m doing with this list of books. The list is a bit of entertainment and a lot of learning, so it is more than a simple reading list.

The next big task is to improve the email newsletter template. It is rather rudimentary and utilitarian at the moment. That suits me fine, and I don’t expect to make it into one of those ghastly marketing “newsletters.” I just think it could do with a little spruce up here and there.

One thing that I should reiterate, if you didn’t know, is that both the website and the newsletter collect zero data on visitors. I have no idea if I’m writing into the void or not. I simply don’t know if the emails are being read or even if the site is being visited. Zero. The only way I can tell that the email is, in fact, being sent is the odd out-of-office message I get back from a subscriber.

I prefer to respect your privacy.

Hitting reply on the email will land in my inbox. Feel free to reach out.

Entering the last quarter of the year, thinking, “Where the hell did the year go?”. Have a great week.

If you live in the Caribbean or have interests in the region, you’ll know that we’re now past the peak intensity of hurricane season on average. The statistical peak is the 10th of September, and we’re now two weeks past this (on the day this is sent as a newsletter), and the remarkable thing is that, so far, we haven’t experienced much heavy activity with the lead-up to the peak being particularly quiet in the Atlantic. And aside from the major hurricane at the start of the season that devastated several of the Windward islands to the south, it felt like it was an announcement of a very active season, as predicted by the NOAA and other prognostications, the reality has been very different for most of us. So I’m writing this on a Sunday evening as the sun sets to an orange sky with some angry-looking clouds, thinking about how I’m pleased that things have been calm for us so far. But… there’s often a sting in the tail, so I’m not celebrating just yet.

ARIN 54 is fast approaching, and if you’re interested in IP, DNS, and the Internet, it is an excellent free online conference to attend. I’ll be there both days because, as I mentioned, I’m a second-time fellow and on the electoral list for a place on the ARIN Advisory Council. I’d appreciate it if you could add a couple of words of endorsement for my candidature to the site. It’s simple and only takes a few moments, and it goes a long way to helping. Just ensure to take a look at the rules to ensure you comply (they’re nothing onerous, trust me).

I’ve been busy with two main topics this week: project management and document management.

I teach at a couple of local further education establishments on the island, and one of the topics I teach is Project Management for IT projects/Software Development. Last week, I spent far too much time trying to wrestle with a well-known project management SaaS application to abide by basic PM rules. And despite it being sold as a “real” PM product, it is woefully inadequate and allows you to do things like planning a task that needs to start only if another task has been completed BEFORE that task. WTF?

This led to an interesting interaction with ChatGPT by a colleague. The interaction was surprising and got me thinking about human-computer interactions and how we haven’t learned much since the original ELIZA experiment, where a simple computer program simulated human-like conversation through pattern-matching, substitution and crude reflection of language (i.e., repeating the words of the user) to generate an illusion that the user was speaking to a device that ‘understood’ or ‘felt’ the human’s feelings during interaction. Of course, it was nothing of the sort and was only a decent parlour trick. This new generation of LLMs is, similarly, a more elaborate version of the very same parlour trick, and more and more people are falling for it.

So now, I’m starting to see more people uncritically trusting the words statistically regurgitated from a website over those from an experienced expert. This is not good, and no good can come from this. And I’m sure I’m not the first to encounter this, but I fear that this is going to become more prevalent, making discussions with people more complicated, given they’re based on misinformation (incorrect information unintentionally repeated or shared), thus continuing to polarise society. I’m pretty concerned.

On the document management front, I’ve been helping a client consolidate and structure their business documents (the usual stuff, Word, Excel, PDFs, etc) to a SharePoint-based model. Things have been going well despite the self-sabotage that Microsoft seems to do when you’re using its products. The end result will be more flexible, scalable and user-friendly, but getting there is not. When I say self-sabotage, I mean from a technical point of view. To be brief, the documents in question were all stored on OneDrive, and a migration to several SharePoint libraries was necessary to structure the data better, and several libraries were required for data separation between company entities. So OneDrive provides a simple way of moving (or copying) data to SharePoint Libraries (and Teams), but it doesn’t expose the paths to all your libraries, only a couple of them. That is, you cannot directly select where you want the data to go unless it happens to be one of the libraries listed that (for the life of me) I don’t understand which ones get listed on which do not. The workaround is dumb, but it works. You have to open the library in SharePoint or in Teams, then create a shortcut to that library in your OneDrive. Only then can you copy data through the shortcut from OneDrive. Way to go Microsoft, let’s make it simple. 🤦‍♂️

Reading

One popular topic that seems to have been displaced by everything AI is Quantum Computing. While not ready for general use, much work is being done behind the scenes in universities like Sherbrooke in Québec, Canada. Le Monde profiled the lab in an interesting article recently. A nice place to study, apparently (a close friend of mine studied there).

I’ll leave it up to you to judge this, but a major Hollywood company, Lionsgate, has sold its catalog to Runway, enabling it (it is hoping) to remove all creative staff from its payroll. There’s an argument for this being determined as immoral. If you’re unfamiliar, Runway is an AI startup with the byline “Tools for human imagination”. I’m guessing imagination means, “Imagine all the money you can keep by getting rid of those pesky employees and their need to eat, live and breathe.”

Let me ask you a simple question. How many domain names do you think it is legitimate for an organisation to possess? 1? 10? Maybe even 30 or 40, that wouldn’t seem odd to me and wouldn’t spark any concerns. Meta (née Facebook) has TWO THOUSAND THREE HUNDRED AND FIFTY NINE known domains. KNOWN DOMAINS. Yeah, that’s not shady at all! For reference, it has four main products: Facebook, Instagram, Threads and Quest. Why so many domains? Sneaky skirting of detection for fraudulent advertising is why. Just look at the list yourself to see it clearly.

Last week, I wrote about a weird billionaire who got his pet social media company banned in Brazil. Refusing to do so, then climbing down in the face of the rule of law (despite bleating to the world that he’d “never” do such things, all the while doing exactly that in other territories), implementing a workaround to bring it back online illegally, irking Brazilian regulators even more, then climbing down once more, and implementing what had been initially asked for, i.e., legal representation in the country and the removal of a couple of provable disinformation tweets. I keep asking myself the same question when it comes to tech billionaires, what the fuck is wrong with these people? Rest of World has a good background article.

I’m not quite sure what to make of this yet, but I thought it would be right to share: The disturbing impact of exposure to 8 minutes of TikTok videos revealed in new study. To ensure my sharing of this is not taken the wrong way, I’m not on board with the moral panic about social media and apps like TikTok, and the holier-than-thou attitudes of some are nothing more than contemptible. Please read the article with an open mind and take the time to read the study.

Lastly, if this wasn’t depressing enough as a newsletter issue 😀, this short article is absolutely worth your time: “The Department of Energy Wants You to Know Your Conservation Efforts Are Making a Difference” from McSweeney’s Internet Tendency. Thank you for all your efforts.

If you’re interested, I’ll be speaking at the second Caribbean IIC meeting this coming Thursday. It’s free to attend, and you don’t have to be a member to register. The roundtable is entitled “AI Ethics & Governance, Cybersecurity, and Data” and I’ll be on a panel discussing AI Ethics and Governance. Mostly.

A human produces this newsletter —admittedly, poorly— but completely free from a parlour-trick algorithm. Have a great week.