AI Prediction Reality Checker
I am offering a new service, available for a fairly priced fee, for anyone writing about artificial intelligence, from popular science writers, journalists to academics who have a background outside of neuroscience and computer science. A lot of writings on AI are vastly, spectacularly overblown in what they predict/fear about the future of artificial intelligence, mainly due to a lack of understanding of the technical details.
Hence I am offering services as a
AI Prediction Reality Checker
Alternatively, I like to call myself a
AI Bullshit Meter Technician
Send me what you wrote, and I will give you an honest assessment of how realistic, given the physics of computation, the neurobiology of human brains and the present-day computing technology your thoughts and musings are.
Why is this Service Necessary
The issue of overblowing the prospects of AI also spills over into related fields, like brain-computer interfaces. An interesting article I recently read is this book review. It seems that both the author of the book that is reviewed, and the reviewer have a much too optimistic view of what is possible in terms of such interfaces.
“. Technological advances such as Elon Musk’s Neuralink device are meant to be implanted in people’s skulls to interface their brain with powerful computers via the internet. Neuralink’s device will be designed to bring information to subjects’ minds and, possibly, to help subjects acquire skills.”
Now, the folks at Neuralink are doing very interesting work, and, in my impression, are very competent neuroscientists. I always like to include such a disclaimer; these are cutting-edge technologies. I am impressed by the actual work they are doing, it’s pushing applied neuroscience. But it will certainly not implant skills and memories anytime soon. With “anytime soon” I mean not realistic at all, with the present-day technologies or anything which builds on it in a way I consider remotely realistic, within a few decades. I did my doctoral work in synaptic plasticity (one of the main biological bases of memory), I believe that a lot of fantastic progress has been made in the last decades in finding out how memory is laid down in brains, but we are far from implanting the skill to play championship level chess, or concert piano into someone’s brain, as some people seem to believe we will soon be able to (and then write philosophy books about the consequences of such skill implants). It’s a combination of aggressive marketing of potential future applications of such technologies with a general academic public which doesn’t understand the details of the underlying biology and physics well which causes such misunderstandings.
To use an analogy, the people at Neuralink have managed to take a fine-grained picture of a city with a satellite so that you can tell the color of houses’ roofs (impresssive!), but they are far from reading what’s in the city library. This kind of order-of-magnitude difference in high-flying predictions, versus what has been achieved and what can reasonable be achieved in the next few decades is what a lot of enthusiastic, well-intentioned folks without a neuroscience background don’t realize. That’s why an AI reality checker is necessary.
My Portfolio
What is my portfolio which makes me a good AI reality checker?
This paper which I still like a lot argues that human science is still very far from simulating a human, or even any vertebrate brain:
And this new manuscript shows that the evil world-dominating super AI which takes over from humans is almost certainly impossible, because it would use way more energy than is available:
Stiefel, K., & S Coggan, J. (2022). A Hard Energy Use Limit of Artificial Superintelligence.
And this blog post contains some general thoughts on the topic:
My Professional Fees
So, drop me a line of you need to have your writings on the impeding takeover of the planet/demise of democracy/end of the human race as a dominant species due to a bunch of advanced statistical techniques (“artificial intelligence”) checked. I offer fair introductory rates, and if you are a grad student I might work for the extra donuts which you stole from the philosophy department’s weekly seminar.
