Sunday Scaries (02/21/21)
A forgotten model of innovation, the Dream Bank, equations that changed the world
Writing to you from Los Angeles, CA.
I spent the weekend working on side projects, including getting my very own Coral TPU from the good folks at Spacepost.
Coral is a small tensor processing unit about the size of a thumb drive. It’s capable of running machine learning models “on the edge” — meaning you can take advantage of applications that use object detection, voice-to-text, and language models without needing to send your data to a central server.
This little device is part of a growing movement to move the power from corporations into the hands of individuals (literally!). Instead of letting companies gather massive amounts of data to serve us recommendations, devices like Coral open up the possibility for “federated learning”. This allows users to update the model weights on their local machines and send companies the results without needing to reveal sensitive data.
Technology is rad!
In this week’s edition
Innovation is unintuitive. Antibiotics were discovered when Alexander Fleming’s bacteria experiment was contaminated with Penicillium fungi. It killed all the bacteria (and ruined the experiment), but led the way to the discovery of penicillin. Einstein was a patent clerk when he had his miracle year and wrote a series of papers that became the foundation for modern physics.
These inventions feel like random chance, yet we pour money into R&D budgets to try and force serendipity. Can we predict when innovation will happen?
Today, the most popular model for innovation is the linear model. This model describes innovation as a series of steps that lead from basic scientific research to commercialization. The linear model of innovation is fundamentally about supply — finding good ideas and applying them at large scales.
However, there is another side to the story. The Demand-Pull model of innovation argues that breakthroughs are not driven by researchers working in a vacuum, but instead as a result of market need. This type of innovation explains the dramatic decrease in energy storage prices over the last decade, as well as innovations in space technology. The Demand-Pull model was popular until a criticism of it was published in academic research. After the criticism, researchers refused to cite the idea so as to not be associated with a flawed model.
Today, researchers rarely refer to the original idea of Demand-Pull, but instead cite the secondary criticism and dismiss the idea without considering it from first principles. This is important because the models that we use to understand the world shape our policy decisions. Without academic backing for the Demand-Pull model we are likely to underfund research that has a nascent market in favor of more basic “supply-side” research.
Lindy score: 2028
This website is fascinating.
The Dream Bank is “a collection of over 20,000 dream reports”, gathered by researchers at the University of California Santa Cruz. It is categorized by the age and sex of the dreamers and includes dreams from individuals throughout their lives.
Some examples of dreamers in the collection:
Phil is a retired humanities professor with no professional or theoretical interest in dreams. He has provided us with dreams from three periods of his life, stretching from age 15 to age 62. We have no other dream series that covers such a long time span starting from such a young age.
Rob Bosnak is a well-known, Jungian-oriented dream analyst and dream-group leader who has written several books on dreams, including one -- Tracks in the Wilderness of Dreaming (Delacorte Press, 1996) -- that includes 53 of his own dreams and his analysis of them. But are his interpretations plausible? Does he understand his own dreams? To answer that question for yourself, you can make your own analysis and then compare it with his.
Tom was a 21-year-old graduating senior when he recorded these 27 dreams in the 1990s. An outgoing, competitive person and very good student, in his dreams he takes great interest in young women, wants to be noticed for his physical prowess, can't find the means to go skateboarding and surfing, experiences hassles at work, and has conflicts with his father and a male housemate.
Samantha is a college-educated woman in her twenties who volunteered her dream journal in September, 1999. It is first of all of interest because it is one of the most contemporary dream journals we have from someone under 30. It is also of interest because the dream narratives are often longer than in many journals.
There are three sets of dreams from this Vietnam War veteran, who had nightmares about his intense combat experience for decades after he returned to the United States. The first set consists of 98 dreams, including many nightmares, that he recorded off and on for more than 30 years after he left Vietnam (1971 to 2008)
Lindy score: 2,070
Lindy score: 4,570
Have a great week,