Discover more from Deliberate Internet
Digital Immortality and the puzzling humanity
Will your soul get stuck in an LLM?
I have been following Tim Ferriss for a while now. 4-Hour Workweek prompted me to join Automattic, so it was great fun and a privilege to build a digital copy of him based on his writing. This Friday, Tim announced the collaboration, inviting his 1.5M+ newsletter subscribers to try it out.
Thanks to consistent WordPress data structures and some hard work of the humble yours truly over the last few months, this feature works not only with tim.blog but on any WordPress site. In fact, it also works on my blog.
If you want to install it in your digital home, email me and I will be happy to help you out.
Facebook has a similar product: It released AI-powered, celebrity-based chatbots. You can now chat with the AI version of Tom Brady, Kendall Jenner, or Mr. Beast.
The difference is quite emblematic - Facebook pays each celebrity $5 million to replicate their likeness in the digital world controlled by them. Our offering is consistent with our value proposition: this is your digital home, and you are in control. You can even set your bot to aggressively use emojis, like I did.
It is easy to imagine an Internet of the future, where most users are AI bots pretending to be humans or even - become so well-trained that they do not realize they are not humans.
During my sabbatical, I've devoured an entire series Sci-Fi series of "The Bobiverse". The protagonist, Bob, signs up to have his head cryogenically preserved in the event of a sudden death. After a fatal accident and a series of legal mishaps, Bob wakes up quite surprised as an uploaded consciousness aboard a Von Neumann probe - a self-replicating space colonization automaton - while the rest of humanity is busy wiping itself out in a nuclear war. What is the minimum “download“ needed to replicate somebody?
Large Language Models (the AI stuff) are essentially compressed knowledge of humanity. In particular, knowledge it can get its digital hands on. And it can most easily get to anything published on the Internet. This would mean that writing on the Internet is your best bet for preserving your thoughts through time and - essentially - achieve immortality.
Will those chatbots be ghosts of the future? If I were a prolific enough writer and captured my every thought and fed it to the bot, would it be able to pass as me? If my knowledge is captured in a model that keeps evolving forever, does it count as immortality?
I take it as a challenge to come up with ideas so novel and surprising and life so unorthodox that ChatGPT cannot predict what I say or do next.
The last standing bastion of humanity may very well be randomness and unpredictability. Aim for being surprising!
A few things I've read
Jessica Livingston, together with her husband, Paul Graham, cofounded Y-Combinator (the most famous startup accelerator). Her piece of the story is more unconventional and surprising and thus more informative,
My grandmother was the most important female role model in my life. She was a very independent person. The term anyone who knew her would use to describe her was "free-spirited." For example, in the wintertime, after putting me to bed, she'd go out and work till late at night on giant ice sculptures she built in the front yard.
YCombinator success is built on throwing parties and events:
When you fund startups in batches, everything’s an event. Interviews are an event, each dinner is an event, Demo Day is an event. As the alumni network grew we started doing events for alumni too, and from the very first year we did big events like Startup School
The titular "puzzle around you" reinforces the value of being unique and building upon your unique strengths. Again, don’t let ChatGPT predict your moves.
So if you want to start a startup, I recommend you try asking yourself what's distinctive about you. What unique combination of abilities and interests do you have? And don't edit your answers, because as my example shows, the most unlikely ingredients could be the key to the recipe.
My favorite mental model when working with AI is comparing it to infinite cheap interns. They don't always do the job exactly right, but when you can guide them properly, you can get some cheap labor.
Ethan Molick distinguishes four types of tasks:
Just Me Tasks are those where the AI is not useful and only gets in the way,
Delegated Tasks are those that you assign the AI, and which you expect to review and offer oversight (remember, the AI makes stuff up all the time), but ultimately do not want to spend a lot of time on
Centaur Tasks are those that you will learn to integrate AI deeply into your workflow and work.
Automated Tasks are ones you leave completely to the AI, and don’t even check on.
Ethan recommends focusing on Centaur Tasks exclusively for the time being.
I love faraway vistas and expansive views, and now I know why.
According to Andrew Huberman, our humans are hardwired to relax when in this "panoramic" mode:
By consciously shifting into panoramic vision, you can activate the part of your nervous system associated with relaxation and calmness
Simply expand your visual field and soften your eyes.
Focus on what's in your peripheral vision.
You might notice your breathing slows down. The muscles in your face and body relax.
So if our AI future stresses out - you know what to do.
As a Diet Coke addict, I am tracking other accomplished figures in an effort to convince myself it’s all ok.
Subscribe if you don’t want to get stuck in an LLM.