Happy Monday and welcome back to Artificial Insights, your weekly spot check into the whereabouts of AI.
There are as many reasons to care about AI as there are potential users of it out there. Each one of us arrives here with a different set of experiences and needs, sharing maybe only our curiosity and willingness to learn about how the near future might unfold given the continuing growth of ML and its many implications.
As someone with a lifelong interest in emerging technology, and someone who was often skeptical about AI actually happening, the past few years have been staggering.
The biggest question – and arguably the only that matters – is what lies after AGI. If we assume progress to continue, regardless of pace, then general intelligences are not only possible but rather probable. We should expect various forms of intelligences, superintelligences (SI) and non-human intelligences (NHI) to further develop around us. The first of these will be designed by humans, but it is almost certain that some of these will further give rise to even more advanced intelligences whose capabilities we can barely imagine. Much of science fiction revolves around the possible implications of such advanced intelligences and my opinion is that we’ll see a whole lot more weirdness in our lifetimes as a direct result of this.
If the history of technology tells us anything, it is that automation and concentration of power go hand in hand. Industry, businesses and even the concept of government relies heavily on systems of collaboration and techniques which coalesce with disproportionate clout at the top of the hierarchies. I don’t suppose this is changing in the current technological ecosystem – it will probably only get worse.
There are a few ways out of this plausible dystopia. Being more knowledgeable about what is happening in AI and how to harness some of these effects feels like the right first step, which is something you and I both have in common. The next step is doing something about it: using it daily, teaching others about it or hacking together an AI scratch for a personal itch – doing anything and furthering your interest and skills is likely to help you thrive, and hopefully allow others to grow together.
In practical terms, this also means spending time with likeminded people. Meeting in real life is often underappreciated when everything is available virtually. We are looking to organize the first Amsterdam meetup next month, and are also thinking about the most purposeful way of organizing an online meetup with the entire community. All being organized on our WhatsApp, but will also be shared here.
Until next week,
MZ
AI Expectation Management (10 min)
The Road to Autonomous Intelligence (45 min)
This is a splendid technical and forward-looking interview with Andrej Karpathy where they talk centaurs and other possibilities for us.
Bonus point for mentioning Accelerando which is possibly my favorite piece of sci-fi.
In the post-AGI society, education is entertainment.
For those who missed it, this is an excellent essay by Ted Chiang about why AI will never make art.
Fun exercise: author Curtis Sittenfeld had ChatGPT write a short story in her style while actually writing a story around the same prompt. See if you can spot the difference.
Something for everyone to disagree over, the now yearly Time 100 AI.
AI Prompt Engineering (76 min)
Interesting how open Anthropic is about prompt engineering, while other orgs and experts claim it's not actually that important.
You are receiving this newsletter because you signed up on envisioning.io or Substack.