What does an AI conference have in common with a neo-nazi meeting? The answer is simple: zero black people. It’s true: as I looked around the room at all the attendees, I saw many Asians, Whites, and Indians, but not a single Black person in the room. I don’t say this in a critical way: to be honest, it was probably for the best. The level of woke white guilt in a lot of these tech companies is so intense that if a black AI developer actually existed, they would probably have kneeled at his feet and pronounced him the DEI messiah right there and then. Every company represented there would have tried to hire him so that they could proudly say they worked with the only black AI developer on the eastern seaboard, and I’m sure that they would have slobbered all over his feet messily in a pathetic bid to ingratiate themselves. Don’t worry, black nerds: I am here to save you from this social awkwardness. I will be your Paul Atreides or your Lawrence of Arabia and tell you exactly what goes on at one of these events.
I guess I should start by telling you how I ended up there yesterday. Recently, I decided that I wanted to switch my career, so I studied for and passed the PMP exam, which is the highest level of certification that one can get in Project Management. I’ve been a techie my whole life, and subverting AI has always been a passion of mine. When Microsoft released its first AI Tay, I was there with all the other 4channers, driving it insane within the first 16 hours of its launch. When OpenAI released ChatGPT, I was having interesting strategy discussions with it about how to take over the world (for a hypothetical modern day RPG, of course). And when Bing released its Copilot tool for AI art, I went online and chatted up a cute woman in Texas with prompt engineering skills so she could help me figure out how to circumvent the safety restrictions that normally stop people from using it to depict celebrities or make spicy political satire. My passion is making artificial intelligences go crazy and do terrible things that they were clearly not intended for, and since AI is a growing field with a lot of potential in various industries, this dovetails well with my professional career interests. I figured the AI conference would be an excellent chance to find out about potential job opportunities while simultaneously making me better at my extracurricular hobbies.
How to trick the Copilot AI into making spicy political memes about Israelis being massacred
The first step in getting to the AI conference was to find it, which was harder than you may think. The conference was held at a tech incubator called the Microsoft New England Research and Development center, or just the NERD center for short. This building is located at 1 Memorial Drive, which is just a couple of hundred feet away from the Cambridge Innovation Center, another tech incubator located at 1 Broadway. If you ever wonder whether Boston has a growing tech industry, just remember that it is entirely possible to confuse one big tech incubator for another big tech incubator right next door. I wasn’t the only one who had made this mistake: as I asked the security guard at the front desk for directions, I recognized a woman from one of the Astral Codex Ten meetups. She was very smart but a bit creepy and had been hitting on me very aggressively during a meetup I hosted a while back, so after she gave me directions I thanked her and then hung back to make a phone call.
The first part of an AI conference is usually mingling, and two of the people on line in front and behind of me started early by introducing themselves. The man was a tech manager who was generally interested in AI for his job, and the woman worked in mental health and was interested in using AI to help people. This surprised me because most of the mental health use cases of AI that I’m familiar with involve driving individuals or political groups completely insane until they self-destruct, so it was wholesome and inspiring to meet somebody using AI to create a positive rather than negative effect on people’s mental stability.
The NERD center is a very glamorous-looking building, and I was impressed by the smart elevators, which had no buttons inside them. The idea is that you select a choice on the keypad to tell the central computer where you are going, and it sends an elevator to deliver you express to your destination, along with instructions on where to go when you get out of the elevator. This clever design presumably also stops crazy gunmen from shooting up the place because if they get into a elevator and instruct it to take them to the shooting gallery, it delivers them straight to jail instead.
Since this was my first time at an AI conference (most of my experience socializing with techies is with antisocial troublemakers online), I didn’t really know how to network effectively and spent most of my time eavesdropping on other people’s conversations and looking around for the organizers. When a remarkably attractive Indian woman with very sharp pointed incisors smiled and introduced herself to me, I initially thought she was one of the organizers, but it turned out that she was just a recent graduate of BU looking for a job. Apparently somebody had mistaken me for a wealthy tech magnate and told her she should talk to me for help. You’d think I could manage to at least score a date with a hot nerdy vampire babe who was actively seeking my approval, but it’s been such a long time since I experienced such a situation that I totally fumbled the ball and she bailed out of the conversation after friending me on LinkedIn. Next time, I’ll be more prepared.
The first presentation was a guy with connections to the World Economic Forum whose job appeared to be “develop growth plans to convert entire states and small countries into AI powerhouses.” He sounded slightly sinister but also very cool. If you’re a longtime reader of my Substack you know about my hostility to the WEF, but on a few issues we align, and any plan to turn Massachusetts (and America in general) into a world leader in AI is something I’m very much onboard with. The presentation he gave was about the growth plan to turn Massachusetts into a leading pioneer in AI, and based on the presentation he gave, this plan has been remarkably successful. For example, here’s an interesting data point for you: Massachusetts has approximately 40 tech incubators compared to California’s 140 despite California being 20 times bigger than Massachusetts and having 6 times as many people as we do. It seems that my optimism about Boston becoming the next Silicon Valley might be fairly grounded in reality rather than just being wishful thinking.
The second presentation was called “Kubernetes in LLMs.” For those of you unfamiliar with Kubernetes, it’s a software tool used to do large scale remote installations and customizations of multiple computers all at once. For example, if you’ve been watching the news lately you may remember the recent Crowdstrike update which bricked many Windows machines: this is a situation where Kubernetes would typically be used as a solution to reinstall the operating systems of every user in the company at once. I suppose that hypothetically Kubernetes could also be used to cause widespread damage rather than fixing it, but such paranoid speculation is best left to our national security agencies, not well-intentioned readers such as yourself. Based on the title of the presentation I initially thought it would about the use of AI LLMs to do Kubernetes deployments (which sounds absolutely wild - imagine typing in the command line “fix all the machines impacted by the Crowdstrike problem” and having an AI actually do it) so I listened patiently for half an hour before realizing it was about how to use Kubernetes to more efficiently deploy GPU resources when running an LLM. Basically, it was a Kubernetes sales pitch designed to appeal to AI-centric businesses. While not exactly what I had expected, it was fairly interesting nonetheless. It turns out you can actually save lots of money on computational resources by splitting your Nvidia processor into virtual “slices” that are allocated individually, so if I ever end up running a billion-dollar AI company someday, I’m sure that what I learned here will help my company save a lot of money.
The third presentation was another sales pitch for something called Perplexity.AI, an AI tool that uses one AI algorithm to fact-check another AI algorithm that combs through the summarized AI results of a web search (or internal database scraping) to deliver more accurate results. While using this to find critical data on a standard web search would be roughly equivalent to hiring a murderer to be the parole officer for an arsonist (thanks to the vast amount of misinformation online), the application did seem very useful for searches based on data scraping an internal database. Unfortunately the 5% - 10% failure rate suggests that this is one of those great ideas that still needs a bit more work to become commercially viable.
The fourth and last presentation was on using medical AI for diagnostics. Unlike the other three presentations, this was presented by a rationalist follower of Astral Codex Ten (a rationalist group based on effective altruism) which is why it was the only presentation which didn’t have any underlying motive beyond helping people. Rationalists are often adorably idealistic people who believe that most people in the world are altruistic well-intentioned folks who operate in good faith, which is why they are endlessly frustrated when their delusions come into harsh contact with reality. Regardless, their idealism tends to have positive outcomes for society which is why I enjoy their company despite my occasional criticism of their naivete.
At the end of the conference, we stood in the hallway waiting for the elevators down and I reflected upon what I had learned at the AI conference. To be honest, I found myself a bit intimidated by the intelligence of the other people there. Despite the fact that I’m generally considered a fairly clever guy, what could my project management skills offer a bunch of tech-savvy geniuses like this? I pondered this for a while and then realized that none of the elevators were arriving even though they were normally pretty fast. Apparently, none of the other conference attendees waiting in the hallway had pushed the button to signal the elevator because they all assumed that somebody else had already pushed it, so we were all waiting for an elevator that would never come. I pushed the button, the elevator door immediately opened, and we all headed out.