Invest Like the Best with Patrick O'Shaughnessy
Jeremiah Lowin: Explaining the New AI Paradigm - [Invest Like the Best, EP.307]
My guest this week is Jeremiah Lowin. Jeremiah has b...
more
Dec 13 2022 1h
Chapter 1 2 mins
[First question] What a pre-trained transformer isChapter 2 3 mins
What latent representation means in the context of AI modelsChapter 3 1 min
Models using math to interpret input data and generate images accuratelyChapter 4 2 mins
Whether or not understanding AI complexity in light of the results they arrive at will become a black box scenarioChapter 5 3 mins
A high level history of the companies involved in generative AIChapter 6 3 mins
The precursory technology that makes generative AI art possibleChapter 7 5 mins
What people are doing to improve AI models in between versionsChapter 8 3 mins
Things that are literally happening during AI trainingChapter 9 3 mins
The power dynamics and barriers to entry for building AI modelsChapter 10 2 mins
Whether or not AI models might one day function as a utility like electricityChapter 11 4 mins
Coding using GitHub Copilot and what it’s felt like to use itChapter 12 4 mins
How he’d approach starting an AI company from scratchChapter 13 5 mins
Developing this technology beyond general and into specific use casesChapter 14 3 mins
The secret sauce for defensibility in the AI model spaceChapter 15 3 mins
What he’s watching more closely as the story unfoldsChapter 16 4 mins
Whether or not he thinks that these toolkits will eventually learn how to use other systems like Unreal Engine on our behalf