- Trainwreck Labs Newsletter
- There’s more than one way to make a brain
There’s more than one way to make a brain
Answers for Globle, Chronogram, and Metazooa from Dec 4 - Dec 10
Coming to your inbox every Monday with educational fun-facts and all the answers to Trainwreck Labs games from the past week.
This week, we have…
A fun fact inspired by a recent Metazooa answer
Answers to last week's games
TIL: There’s more than one way to make a brain
What would Artificial Intelligence be like if it was created by octopuses instead of people? Image generated by DALL-E.
One of the hottest topics of 2023 was Artificial Intelligence. This is in no small part because of large language models like Chat GPT (the AI engine behind the guests you talk to in Chronogram and Fictogram) and Google’s recently announced and highly anticipated Gemini model.
When we say AI, we often mean some version of Artificial Neural Networks (ANN) which “learn” by training on large datasets to predict the most likely outputs for similar future inputs. But are they really learning? Do ANN brains work like the brains that grow inside our human bodies? They’re certainly not identical organs, but a brain doesn’t have to resemble ours to be a great problem solver.
Just look at our distant relatives, octopuses (I’m told octopi is wrong, don’t @ me). Whereas we have the vast majority of our neurons snugly contained in our skulls and spinal cords, our squishy cephalopod friends have most of their neurons distributed throughout their arms. As a result, their neuron-dense tentacles can do more than slither: scientists theorise that octopuses can move, smell, and create memories independently of the brains in their heads!
Metazooa players will know that octopuses (including Animal #129, Enteroctopus dofleini) are among our most distant relatives on the phylogenetic tree. Our brains evolved completely differently from each other! That said, no one can deny that they are remarkable problem solvers, just like us… and just like Chat GPT.
So are AI brains just like ours? Not really, but human brains don’t have a monopoly on remarkable thinking and problem solving. We learn a lot from studying and understanding our evolutionary relatives, not the least of which is that there’s more than one way to build a brain.
Godfrey-Smith, P. (2017, January 1). The Mind of an Octopus. Scientific American. Retrieved from https://www.scientificamerican.com/article/the-mind-of-an-octopus/
Answers to last week's games
Monday, December 4 to Sunday, December 10.
"José Costa y Bonells" by Francisco de Goya
Forgery of week, from Dec 5
Play Forgeous for Dec 11.
Goya, F. de. (1800). José Costa y Bonells (died l870), Called Pepito [Painting]. Oil paint on canvas. The Metropolitan Museum of Art. https://www.metmuseum.org/art/collection/search/436546
No major front-end game updates this week because I’ve been upgrading the framework used to build most Trainwreck Labs games and infrastructure. This has been fairly time-consuming but will make future updates faster and less error-prone.
Looking ahead, one of my top priorities for the coming weeks is to translate the Meta games into multiple languages. There are still some technical details to work out, but if you would rather play with a pulpo, bläckfisk, 章鱼, or ऑक्टोपस than a plain old octopus, stay tuned for updates!
Another solid meme from the Trainwreck Labs Discord server this week, albeit a bit niche. Surely the Venn diagram overlap of Donny Darko fans and botanists can’t be that small, can it?
The Trainwreck Labs Newsletter has over 10,000 weekly subscribers. Would you like to become a sponsor? Reply to this email and we’ll get back to you ASAP!
Thanks for reading, and enjoy the rest of your week!
Before you go, please rate this week's newsletter!
Be honest (but don't hurt my feelings)