If you were having fun discussion If you're looking to share the holidays with friends and family, you're in luck! Gift a subscription and save 15% through December 24th. discussion Please use the link below.
And for those looking for something more specific, we also have merchandise such as hats and “Libbing Out!” Stickers are now on sale online. Order as soon as possible to receive it before the holiday season ends.
There were some pretty unusual graphs recently. financial times Article about the difficulty of predicting AI. It is based on the following analysis. dallas federal reserve bank It then displayed the following three lines.
One is “Technological Singularity: The End of Scarcity,” in which AI will create immense wealth and productivity, with real GDP per capita about 10 times the current level by 2034.
The second is “Technological Singularity: Human Extinction” (GDP per capita will be zero, but AI will probably still be doing AI work).
Third, with the help of AI, GDP will continue to maintain its existing growth trend over time.
Here's another version of that graph with a line showing the current trend in GDP per capita.
Of course, this kind of prediction is comically useless (and the author clearly knows that). But this analysis is also more honest than many other AI analyzes that tend to fall into the trap of “I don't like AI, so AI won't affect us economically” (I like technology and progress, so it will enrich us nicely and smoothly without being too destructive, like the internet).
These are bad ways to reason about AI risks. As we approach the beginning of 2026, I've been reviewing people's predictions for 2025 regarding AI and seeing how they held up. And I was shocked at how badly these atmospheric predictions turned out.
I don't blame people who hate AI, but I will say this. With AI, we get really bad at predicting AI. Disliking AI (and rightfully so) leads people to believe that AI must be incompetent, useless, or both. However, these are separate issues and are worth considering separately.
“The AI is simply a word guessing program, as it simply calculates the most likely next word.” This was primarily a complaint from bitter machine learning academics. probabilistic parrot (A poetic but ultimately misleading statement that AI, by its very nature, cannot do anything other than repeat input.) And it was mostly exasperated artists who complained about the plagiarism machine.
Lately, I've been hearing this term mostly from internet denizens who are complaining about how its vital heartbeat, actual human-created content, is being replaced with something vulgar in its characteristic speech patterns.
