Like many nerds, I am a Douglas Adams fan. And the answer provided by one of his iconic inventions, Deep Thought, (that name has made its way through the culture, hasn't it?) made an even deep impression on me. I am thinking about 42 being the answer to the mysteries of Life, the Universe and Everything Else. That was Adams’ way of ridiculing legions of deep thinkers and snake oil salesmen, but long have I nursed this ambition of doing something with that number.

The advent of consumer AI makes Deep Thought a near future accessory. What if we actually had a superintelligence that could answer the mysteries of life, the universe and everything else? How might we ascend the ladder of enlightenment with this superintelligence as our guide?

This series of 42 posts is an attempt to explore that question.

A level one problem, is simple; you should be able to answer it right away - no thinking required. Two is a little bit harder. It might take you a few minutes. These days it might take a ChatGPT query. Three and four are further up the hill. It's like a term paper or a particularly hard problem you might have solved for extra credit. ChatGPT might well collapse the distinction between 1, 2, 3 & 4: prompt and ye shall be answered.

You're getting to some serious stuff at 6 or 7 —undergraduate thesis level, or a master's thesis for 7. This is a logarithmic scale and the difficulty rises fast. By the time you get to 14, we are well past a PhD, it’s more like the life’s work of a brilliant mind. 21 is very, very hard— think of Fermat’s Last Theorem - a finding that combines the sweat and toil of very smart people working over centuries. 42, of course, gets you to life, the universe, and everything else. A civilization's worth of insight.

That’s my classification, the 42 scale.

In the coming weeks and months, I'll be sharing 42 musings; some will be pure questions, some backed by arguments and some challenges for the future. I have been told that in the age of AI, the victor’s spoils will go to the ones who ask amazing questions, since our mechanical friends will answer any question faster than you and I can.

Let’s get questioning.

Note: I started with the factors of 42, but I might cheat by turning it into a continuous scale, with every number between 1 and 42 fair game. Factorization remains a good mental model, because difficult problems can be broken down into easier problems. Even doable problems.

Let me give you an example of something I have been toying with as a lazy side project. One day it occured that AI might be the technology that helps us create an internet of dreams. A web of consciousness is a 14 level challenge at the very least. But I started hacking away at it at level 4. For the last few days, I have been speaking my dreams every morning into a voice notes app, that transcribes the dreams and turns them into text. I feed that text - after some minor tweaking - into Midjourney and after a few iterations, I produce a seven second video of the previous night’s dreams.

TBH, the video is nothing like my dreams; it’s at best a creative representation. But who cares?

But this stage 4 prep already suggests the next few stages. For example, can we create an app where anyone can record their dreams? Of course we can. It took me ten minutes to prototype that app on one of those AI services that the cognoscenti have heard about. Makes sense to vibe code a dream platform.

The hard problems at this stage aren’t technical, they are in making the right design choices. For example, should we be able to record our dreams at any time? I think not. You should record a dream when it's fresh in your head. So the app should be allowed to record your voice the first time you pick up your phone in the morning, with a window of a minute or two or five, but no more than that.

The next stage: how about if we connect each other's dreams? But what's the logic of connection? I am thinking some light semantic markup should do the trick. Let's say my dream and your dream have a common theme, or perhaps even the same person or location.

That semantic markup affords dream surfing: if our dreams intersect, I should be able to hop from my dream to yours. But that, of course, requires us to have our dreams available publicly - isn't that a problem? Aren't dreams some of our most private experiences? Before we start dumping our innermost thoughts into the commons, we need to be convinced about their security. Then there’s the issue of why would anyone want to make their subconscious public? Feels like a level 9 moral problem, one our matrixed culture is primed to address.

Starting at 10: a wearable device like a portable EEG monitor that streams your dreams live as you're dreaming them. You're no longer the agent in charge of narrating your life. That's the job of the machine. It will etch them into a video that could potentially be mined for posterity. Very dystopian. There are level 10 technical problems to be solved. We will need accurate translation from subjective experience to a public record. And that, in turn, will need excellent data? Lots and lots of people would have to record their dreams with great precision to be able to train these devices that translate brain waves into dream narratives.

I think it can be done with today’s tech, if Mark Zuckerberg or someone less creepy puts a few millions into the task. The level 12 questions are all about individual agency and the dilemmas associated with letting our collective subconscious loose on society. I am not saying it's a good thing, but if we unleash that Pandora's box, it will be the foundation of a whole new culture that doesn’t exist today.

(Level 12) Do we want that culture?