Topic: Scientific Computing

OpenAI is reportedly developing "Project Strawberry" to enable AI to conduct autonomous deep research. Meanwhile, a "boots on the ground" expert critiques NVIDIA's new Blackwell chips, arguing that their performance increase for scientific computing is underwhelming compared to the high cost and existing alternatives.

The hosts challenge the notion of a total scientific consensus on climate change, citing a letter signed by 30,000 dissenting scientists. They revisit a 1989 Reuters report where a UN official predicted that entire nations would be wiped out by the year 2000 if global warming was not reversed. The reliance on computer simulations for long-term environmental forecasting is criticized as unreliable.

The debate over climate change is characterized as a government-led experiment in mind control, utilizing psychological tricks and the "97% consensus" meme to influence public opinion. Critics argue that climate science fails the scientific method because it relies on non-repeatable computer models rather than empirical observation. The discussion references the "Climategate" email scandal as evidence of a coordinated effort to manipulate data behind the scenes.

The concept of "scientific consensus" is challenged, with the hosts arguing that true science requires repeatable and observable empirical facts. They cite recent failures in computer modeling, such as the 2010 volcanic ash flight bans and errors in economic Excel spreadsheets, as evidence that models are often unreliable and politically driven.