Something happened, which is now the only reason this arrives.
In the last issue I retired the schedule. No more Sunday deadline, no fortnightly obligation and no publishing simply because the calendar said I should. This newsletter now turns up when something happens that is worth roughly eight hundred words of your time. I half expected that to mean the next issue would be a month away. It lasted three days.
I have been writing a book for seven weeks and only noticed this week.
It was not deliberate. From the beginning of this project, I have written down every mistake as it happened, together with the numbers, the dates, the things that broke and the theories that quietly died once reality arrived. This was meant to be record-keeping so that I could look back and understand what worked. Read from beginning to end, it turns out to be a manuscript. Twenty-one chapters of one.
Somewhere in the middle of putting it together, I finally wrote the sentence that I think explains the entire project. It is something I have been trying to say since the beginning without ever quite getting hold of it: I wasn’t trying to escape Australia, retire early or become an internet entrepreneur. I was trying to stop buying my life as a bundle.
Politics hands you two enormous shopping trolleys, tells you to choose one and then expects you to defend every item in it, including the things you would never have picked off the shelf yourself. But politics is not the only place this happens. Career says climb or quit. Retirement says work until the official age or somehow become independently wealthy. Housing says buy the largest one the bank will allow. Geography says choose a country and arrange your entire life around being permitted to remain there.
Every one of those packages was assembled by somebody else, and they are all offered on roughly the same terms: take the whole thing or explain yourself. I am becoming less interested in the trolley. I want the basket. I want to choose the parts that work, leave the parts that do not and stop pretending that rejecting one item means I must reject the entire shop.
That is closer to what The Long Game is actually about. It is not disappearing, escaping responsibility or spending the rest of my life drinking something with an umbrella in it. It is about having enough control over my income, time and geography to make individual choices again, instead of accepting somebody else’s bundle because it arrived labelled “success.”
A warning first, because it is probably the most useful thing I learned this week.
By the third video I thought I was getting good. I introduced my AI agent using a hard, full-frame flash transition that I was genuinely pleased with. In the editing software, the effect is called The Glitch, which should perhaps have been more of a warning than I allowed it to be.
I then discovered, entirely by accident, that there is a published accessibility standard covering flashing content. The guidance says content should not flash more than three times in any one second unless it falls below defined thresholds. The reason is not that flashing effects are annoying or make a middle-aged man look as though he has discovered nightclub editing. Flashing content can trigger migraines, nausea and seizures, particularly for people with photosensitive epilepsy.
I did not take the video down. This is a warts-and-all project, and that particular wart is still sitting on YouTube. If you go looking for it, take the warning I failed to provide: there is a hard, full-frame flash in the video. I still do not know whether it crossed the threshold because I never counted the frames before publishing it. That is the honest version. I published something to strangers that I could not vouch for.
The video has nine views. It is possible YouTube watched it as well and quietly reached the same conclusion.
If you make video, the standard is WCAG 2.3.1. It is free to read and takes about four minutes. Of everything I have learned during the last seven weeks, this is the only mistake where the person who could pay for it is not me. That makes it worth knowing before you discover it by accident.
Now to the thing I actually sat down to write.
There is an enormous amount of noise around AI at the moment. Depending on which post you read, it is about to replace half the workforce, remove entire departments and allow one person with a laptop to run a global company before lunch. The rest of us should apparently either prepare for life after employment or become prompt engineers by Friday.
I have spent the last month allowing AI to run a fair chunk of a small media company. It has written content, scheduled it and published it. It built the website, connected the newsletter, created the sign-up forms and reported that everything was working. Every task was completed on time and followed by a reassuring green tick. Nothing broke, which turned out to be the problem.
I eventually tried to work out how many of my 2,131 LinkedIn connections had joined the newsletter. The answer was that I had absolutely no idea. The system had carefully recorded which of my own pages people used to enter their email addresses, while recording nothing about what had sent them there in the first place. Nobody had asked it to collect that information, so it had not.
The subscribers who arrived through $100 of Instagram advertising were tagged, dated and perfectly identifiable. Apparently Instagram becomes extremely organised when it is spending your money. LinkedIn may also have sent subscribers, but I had built no way of proving it.
I then went back and read the last twenty things I had published on LinkedIn as if I were seeing them for the first time. They were fine. Some were probably better than fine. Not one contained a call to action. There was no mention of the newsletter, no link to the website and nowhere for an interested person to go next.
AI had written the posts, scheduled them and published them exactly as instructed. I had produced twenty reasonably good pieces of content, delivered them on time and sold absolutely bugger all. The technology had worked perfectly. The person writing the instructions had forgotten to connect any of it to a business outcome.
That is the part I think is missing from much of the current AI conversation. AI can write things, schedule things and post things. It can work at a speed that makes a month of human procrastination look particularly embarrassing. What it cannot do is automatically know whether the work matters, whether the result is correct or whether you have asked it to solve the wrong problem.
In 1999, if you built a system, you tested it. You opened it up, walked through it like a customer and checked what came out the other end. We were still doing that in 2023, which is roughly when a lot of us seem to have stopped. AI has not removed the need to check the work. It has simply made it possible to skip that part and produce a great deal more of it before anybody notices.
A green tick means the task was completed. It does not mean the business outcome happened. AI did not fail me. It did exactly what I asked it to do for an entire month, without complaint and without once pointing out that some of what I had asked made no bloody sense. If you do not check the work in the same way you would have before AI arrived, shit happens—only now it happens faster and gets automatically posted to six platforms.
For thirty years I was paid very well to sell software that worked. This month I took my own word for it.
The numbers, and then I will leave you alone.
The newsletter now has 25 subscribers. Thirteen of the 24 people who received the last issue opened it, giving it an open rate of 54.2%, the second highest this newsletter has managed. Nobody clicked a link, making that three issues in a row with zero clicks.
The website has had 25 unique visitors in total. According to the information I have, exactly none of them came from LinkedIn, TikTok, Instagram or X, although the entire point of the previous section is that I am no longer willing to swear that the information I have is correct. There have been no unsubscribes, which is encouraging, and revenue remains firmly committed to $0.00.
So, an ask—which, given everything above, I should probably start doing.
If you know somebody in their fifties who is quietly wondering whether the life they built is the one they would choose again, please forward this to them. That is the whole ask. There is no form to complete and no reply required.
The book will be written either way. When there are chapters worth reading, they will land here first because this is the only list I actually own, and twenty-five people is still a room I can picture. I’ll be back when something happens.
— Mark
The Escape Report documents one person’s attempt to build a way out of the cage in his fifties — the real numbers, the mistakes, and what actually happened. No schedule. It arrives when there is something to say.
This newsletter runs on beehiiv. That’s an affiliate link — if you start a newsletter through it, I earn a commission. I’d have recommended it anyway; it’s what I chose before there was any commission in it. I’ll always tell you when a link pays me.