He began by saying his name — Sam — and by naming the shape of the thing he was about to do. First in a series. Walking you through my work. He was already comfortable on camera; he had a YouTube channel, had for a couple of years. But this format asked something different of him, and he said so plainly before proceeding: talking about work in a linear fashion is not how his brain works, not how he thinks, not how he talks. He allowed that it might be clumsy. He allowed that it might improve. Then he stated where the videos would go — the freelance platform where he had once been successful and which now seemed largely dormant, the other one he was trying to put himself back on, his homepage, the transmission archive, the YouTube audience where all his videos end up. A distribution map laid out before the first artifact was even shown. Then he turned toward the thing itself and said it was maybe a little bit sentimental. This one always meant a lot to me.
The site came up on screen with what he called that old tech vibe. He named his own position in it immediately and without hedging: front end is not his strongest skill set, he is a back-end programmer, exceptionally skilled at that. The stack was PHP and MySQL. The layout was probably Bootstrap, because there were not really CSS frameworks then. Underneath all of it ran a content management system called Enet Wizard Matrix Server, which he had started building in his teen years and worked on for many years in his free time, open-sourced for a little while, then set down as CMSs blew up and WordPress and the others arrived. He had already moved on. He was probably working on other things — things like this. He was a freelancer on the project, though he described it as more like a partnership, the first long-term one he had: ten years with the company, roughly half of that on this particular site.
He had built a slideshow to accompany the walkthrough, and it required a little bit of effort to keep working — literally running on a local server on his own back end. He adjusted it live, mid-recording, and narrated the adjustment as it happened. Okay, it worked. Cool.
Then the archaeology. A lot of the older projects are simply not online anymore, so he had pulled this one up on the Wayback Machine, and there he arrived at the fold in the record. Fifteen years ago he had written custom back-end processes to prevent scraping. Every visitor examined, every IP address checked; when something claimed to be a bot, his scripts verified whether it was actually the bot it claimed to be. He described it as very sophisticated, and it had worked — it had worked on the Wayback Machine too. Roughly 2,700 captures sat in the corner of the screen, and across the span the site was live, most of them held nothing but the message he had written to output to bots. There's nothing else to show you except this, basically. The defenses had outlived the site and were still doing their job on the only surviving record of it. So he narrated what the images could not.
What came back came back in layers, not in order. The writer's dashboard he could no longer show — writers recruited from around the world, writing independently, managed by him, revenue split with them and tracked, the numbers visible to them in their own dashboard. They could see what was trending and where the traffic was coming from, so they could prioritize: celebrity biographies, movie reviews, episode recaps. There was a calendar for the shows they followed. Then the ad revenue split engine. Then the gamified point system for the users — and here he stopped himself mid-sentence, having just retrieved it in the act of speaking: anybody could submit reviews, and if they met the standards, they were posted and earned points. I actually forgot about this till just this moment. Points from writing, from rating pages, from writing to a favorite celebrity. Points spent in auctions he built, bidding on merchandise donated by companies wanting to reach the audience, auctions running for a while and then closing, and once a month he would mail out the prizes, sometimes hundreds of little ones. They did that for a number of years.
The letters had their own architecture. Celebrity love awards, tied to users writing to the people they cared about, a whole section where the letters could be read, and whoever received the most by year's end was the winner. For the top ten, he burned custom CDs — a static website on a disc that let a recipient browse every letter written to them without ever visiting the site, and reply to them, and he would send the replies out. That was a fun thing that we did. Behind all of it, the origins of the database: scraping across the web, APIs, link exchanges, partnerships with other entertainment sites of the era, TV Guide among them. Trending detection. A very longtail structure, deeply SEO-friendly, which is why the traffic arrived from so many random directions, especially on the celebrity side.
Near the end he came to a set of notes on screen and named their source without ceremony — they came from artificial intelligence, from a conversation he'd had about the project and what this video might look like. It had said the site was designed not just for engagement but for participation. He agreed with it, and then extended it past the site: that's probably how I approach just about everything. He described what the platform had become — a community, before social media really existed, a place where people could gather around something they cared about. That to me felt meaningful. Reciprocity throughout the system, for users and for writers both, always about meaningful contribution. One of the early long-term projects. He said he was very proud of it, and that he wished there was more of it he could show. Then, with the archive still open on screen and holding only what it had been permitted to keep: yeah, there it is.