20 karma · joined July 14, 2023
Yet we’re obsessed with framing ourselves as chains of matmul.
Is it accurate to say you’re willing to go into ~20,000,000 USD debt to sell discounted computer-as-a-service to researchers/startups, but unwilling to go into debt to sponsor the undergraduate degrees of ~100-500 students at top-tier schools? (40k - 200k USD per degree)
Or, you know, build and fund a small public school/library or two for ~5 years?
Excerpts: - A person who endeavors to solve problems with a computer should be distinguished from someone who does so in order to spend more time at the computer. A real scientist can go home at the end of the day and think about something important, like their family, and not be kept up by thoughts of the computer. A scientist who is ever-tweaking their code for more dopamine is not a scientist but a computer addict with a side job as a scientist.
- A computer’s memory is a perfect medium wherein any memory configuration is possible. As the programmer’s power over this domain is effectively absolute, one must acknowledge that their corruption and will to exploit this domain is also absolute.
- Technology amplifies power. It is a lever. Just because your choice was impactful doesn’t make it right.
Don’t use centralized and closed-source AI.
Reminder that flux-pinned levitation only occurs when superconductors are cooled from above to below their critical temperature while in a local magnetic field.
The researchers probably didn’t heat up their big sample above the critical temperature in air as that could have mechanically destroyed it. It was already chipped almost in two.
But the papers don’t show that. There’s no obvious contradiction unless you make some big assumptions.