The core thesis for Unconventional AI is that AI models are inherently stochastic, making them well-suited for analog computation, which is also stochastic and can dramatically reduce power consumptio...

Constantine Bueller is wary of investing in companies that prioritize short-term revenue growth over building sustainable long-term economics.

Constantine Bueller believes the primary limiting factor for technological progress is not a lack of talent, but the 'matching problem' of placing talented individuals in their highest and best use.

Constantine Bueller believes the vast majority of investors did not seriously begin looking into AI and machine learning until 2022.

Constantine Bueller cited Landauer's principle, which defines the theoretical minimum energy required to erase a bit of information.

Constantine Bueller predicts that within 20 years, the vast majority of cognitive work will be performed by machines.

Constantine Bueller asserts that digital computers were widely adopted primarily due to their reliability and the need for precision in calculations.

Constantine Bueller estimates that currently only 1-2% of global cognitive labor is being automated by AI.

Constantine Bueller asserts that OpenAI's early research, which led to modern large language models, was a contrarian effort that the vast majority of the academic community believed would not work.