🗞️ January 2024 - An Open Internet and Thoughts on Generative AI

This is my first newsletter of 2024, and it’s a long one. I look forward to writing more during the year. I won’t promise they will be sent on a strict schedule, but I’m setting an overall goal to get back into the rhythm of writing these long-form posts here and in newsletter form for this year’s subscribers list.

Enjoy, and let me know your thoughts by email or on Mastodon.

I’ve set up a new site to consolidate all the public writing I’ve been doing. I mentioned it before in the previous emails, but I’ll take the opportunity to plug my site again. I’ve added a page with what is essentially my CV to the site; the idea is to give people an easy one-stop shop to see what projects I have worked on over the last few years. The list isn’t exhaustive; it’s more representative.

An Update on the Newsletter Migration

In the last newsletter email, I talked about the distasteful issues and goings-on at Substack. In that update, I said I would be moving to a different platform and that I had my sights on either WordPress or micro.blog. After a lot of research and discussion with the support at WordPress, I took the plunge and decided to go with a WordPress site. That didn’t turn out to be a good decision for several reasons.

Sadly, WordPress’s idea of a newsletter is not really aligned with mine. Secondly, it was a challenging task to get the site up and running and looking the way I wanted it to. I needed to take a few training sessions to get started before I could get the site edited to look like something I’d be happy with. Domain purchased, WordPress plan purchased, I did the transfer and moved across the entire library of articles I’d written on Substack. The migration was easy enough until I reached a limit of subscriber numbers, which, to be fair, was easily resolved but annoying to run into and be taken completely by surprise.

However, I wasn’t happy with the way things work over at the site. Number one, a subscriber has to create a WordPress account to use the newsletter properly, and I don’t think that should be necessary. Secondly, there is a non-optional amount of tracking performed by sites like WordPress (Substack did this, too), which I didn’t want to keep de facto endorsing. I don’t need to see ‘stats’, and I don’t need them feeding the anxiety bucket. I want to write informed and interesting articles, put them out there and see what happens over time. I don’t need tracking stats to know where you’re from, what you read and when, what you had for breakfast or anything for that matter.

All this to say that I have performed a second migration in the space of one week from WordPress to micro.blog. It is a small and independent company I knew about a few years ago, as I was one of the early backers on Kickstarter to get the platform up and running. For some reason, I didn’t find a use for it back then, but recent events made me reevaluate that, and I’m here now.

You should continue to receive the newsletter as previously, but the look and feel will be a little different, as I noted in the last email from Substack (So long and thanks for all the fish).¹ The new platform doesn’t have all the bells and whistles of Substack, but I’m okay with that, as I think it is the content that is the most important, not the flashiness. I think it speaks more to who I am and what I do.

The Internet’s Past, Present, and the Movement for a More Open Future

I haven’t been as enthusiastic about the Internet since I first started using it back in 1989 when the Internet was a series of clunky command-line tools like Gopher, WAIS, and a few others. It blew my mind back then that I could communicate in almost real-time with a student in San Fransisco from my university DEC VAX VMS terminal in London, UK. But logging on to baymoo.sfsu.edu became a ritual and a pastime that shaped how I used the internet and thought about the future. Shortly after that, ISPs (Internet Service Providers) started popping up in the UK, and I was one of the early clients of a dial-up service based in North London. I even applied for a job with them and went through an interview (and failed), but I remember seeing the hundreds of dial-up modems they had in the office for the connections from their customers like me. I got myself a ‘real’ email address and sent an email to myself from my university account to that personal account, racing home to check I got it as intended. My car didn’t go as fast as electrons, so I lost that race too.

What set the Internet apart at that time was its truly open nature. Open, as in having not walled off, private, or for-profit-only tools. Tools like GOPHER², WAIS³, and TELNET⁴. This presented an almost limitless opportunity in its time for people to develop new ideas and new applications. The most notable of those is the very system that you might be reading this on now, the web, the World Wide Web, or WWW.

In 1991, Tim Berners-Lee and CERN released the HyperText Transfer Protocol (HTTP)⁵ protocol and a rudimentary browser called Nexus.⁶ This transformed the Internet entirely, and its technologies developed into what we have today: visual, virtual spaces on the Internet. It democratised the Internet for anyone able to get online through an ISP by allowing people to create easy-to-navigate, easy-to-use and interactive websites. If you’re interested in the specification of such things, the RFC for HTTP can be found here.

What developed after this amounted to what I would call a Cambrian explosion of websites and innovation on the Internet, which, eventually, made it big enough for the financiers to step in. Slowly but surely, bits of the Internet got walled off. Here and there at first. Little by little, then, all at once. The Internet was no longer an open system. Sure, there are still some open systems, but they are dwarfed by the platforms such as Facebook, Microsoft and Google. All are responsible for intimating and pretending to be open whilst closing down the real openness of the Internet so they could sit in between all Internet things and extract money from anything that happened in either direction —Site to user, user to site.

And that’s where we are today. An Internet with a rich tapestry of site designs, features, and opportunities (primarily for grifters). But it is a sad Internet, one with plenty of bad things despite an enormous amount of innovation and ideas for an open Internet. Most of them are stillborn or are stifled or bought out by giants as soon as they make enough of an impression on the masses and possibly threaten an incumbent. Instagram is the canonical example. It was a lovely app for amateur and professional photographers alike to share ‘olde filtered’ square photos taken using smartphones that had only just gained decent camera parts. It is now a disinformation machine entirely driven by advertising, most of it absolute garbage or downright dangerous. There is only one winner, Mark Zuckerberg. It is now a platform that could be subject to health warnings or regulated to change if some of the proposals to control the platform get implemented. It, and others like Twitter, are being targeted by the EU for abusive privacy practices and flat-out violations of the GDPR. But even that doesn’t stop them trying to squeeze the last drop of cash from people. It’s just a cost of doing business. Take a recent example, Facebook. They recently announced a convenient feature called Facebook Link History. Convenient for who? Facebook, of course. It is essentially a key-logging Javascript injected into every site you visit and monitors everything you type or tap on, including your passwords! It should be illegal. Facebook has ignored GDPR since the law has been in force, believing it is too powerful to be taken down.

You should understand that advertising incentives are not aligned with you, the customer, or the seller. If you want to know more about online advertising and understand how the machine actually works and, importantly, why this type of advertising isn’t as efficient as we are led to believe, I’d suggest looking at this EU Commission document. For the record, I don’t subscribe to the notion that you are the product if you’re not paying for it. This is too reductive of an explanation and doesn’t adequately describe what really happens. Advertising giants are squeezing both ends of the value chain, you and the would-be advertiser, by telling you both lies about reach, accuracy, and the other largely made-up metrics.

I believe we should try to get to a modern version of the open Internet of before. I don’t mean dialling back the clock as it is impossible. I don’t believe in the “things were better before” doctrine either. I’m advocating getting back to a point where anyone could have and, this is the crucial part, control their own plot of cyberspace. A more distributed Internet, one that values quality, not quantity. One that values truth, not who can shout the loudest. In trying to explain what I mean in clear terms, I’m thinking about the British Broadcasting Company, the BBC —one of the world’s oldest and most respected media companies.⁷ The Internet link it promotes on its News programs is www.facebook.com/bbcnews. The site and brand is Facebook. Not the BBC! It should only ever be www.bbc.com/news. It should only ever be a space that they control, not a Facebook walled-garden portal.

I’ve been reading a lot about the distributed Internet, and I believe it is a good start. Note: Don’t confuse the web3, crypto, etc model of “distributed” with what I’m thinking about. That is an entirely different “distributed” and a discussion that has somehow damaged the image of distributed in its meaningful form. I want to write more on that in the future as I think it is at the heart of the reason why, in the Caribbean, we don’t have value in using the ccTLDs, with businesses not benefiting from that visibility and attractiveness as in other regions. Anguilla would disagree with me here, but they are the exception currently riding a wave of popularity. The .ai ccTLD is a hot property currently earning the tiny British dependency millions of pounds in revenue.

Harnessing AI Responsibly: Insights from Training Business Leaders

I wanted to mention a little about the new hotness, AI. I’ve been teaching a reasonable number of business leaders about these tools over the last six months.

It is clear to me that I have been surprised by the interest from such a broad range of managers and business leaders for a product that is so technical and so linked to ICT. The OpenAI hype machine has galvanised the public into believing that these tools can make them one hundred or more times as efficient for 100 times less money than they are spending at the moment (on personnel). This, of course, is not true at all, and I find I have to temper expectations and canalise those runaway thoughts they often have about generative AI and how it will make every person redundant.

For the record, I remain enthusiastic about the technology from a basic productivity point of view. I do think it brings something to the table that can be helpful when used responsibly. I liken it to the automated systems on some cars that ensure the correct security distance between you and the vehicle in front without human input. It’s not self-driving. It is just an assistive technology that needs guardrails and human verification. If an accident occurs where you run into the back of the cat in front (despite the technology being activated), who is responsible? The assistant software in the car or you That’s exactly what we’re dealing with when we use these systems. You, the user, remain responsible, and you, the user, should ensure you use it responsibly.

I don’t think discussing accuracy, efficiency or other measures of “intelligence” is helpful at this stage, as these systems are changing rapidly. To give you an example, I have had to modify the training materials no less than ten times in the last six months. I would suggest a wait-and-see approach before integrating them into fundamental or central processes in your businesses that would provoke significant consequences in the case of error or failure. I would also suggest you integrate human-based verification and validation to the output generated to ensure you don’t fall foul of mis and dis-information, obviously wrong answers, and poor analysis that these LLMs can produce. That doesn’t mean that I don’t support the use of them. Please do. However, please don’t rely on them too much, as you may be sorely disappointed and dissatisfied with the results. For once, I’m bullish on Microsoft’s approach, but I would still exercise caution handing over the car keys to Copilot, ChatGPT, Bard and other LLMs.

Thanks for reading, and I hope to email you again soon.

/comittedtodisk