Pinned
Science has swung from "take no one's word for it" on one side to "trust the science" on the other.
I discuss how science has changed over time, and how it will continue to change in the future, in this article:
Geoff Anders
2,484 posts
Geoff Anders
@geoffanders
- People are now thinking about whether they could have told that Sam Bankman-Fried was a fraudster. I told several people, many months ago, to steer clear of the FTX Foundation and that SBF was going to be taken down by the feds. How did I know? A brief thread. 🧵
- The Apollo project is a great example of people being able to construct extremely long sequences of essentially error-free intellectual output. Lots of people actually think that this is not possible.Margaret Hamilton, lead software engineer of the Apollo project, stands next to the code she wrote by hand to take humans to the moon! #ExascaleDay #exascale #WomenWhoCode #womenintech
- Replying to @geoffandersThe next piece of evidence I used was Bankman-Fried’s ideology. He’s a self-declared “hardcore utilitarian,” which can means he’s really serious about doing good… or a sociopath who rationalizes bad actions using math. The evidence of character indicated more of the latter.
- I just learned that the Soviet Union landed on Venus several times in the 70s and 80s and sent back pictures of the surface, including this color picture:
- Replying to @geoffandersThe big takeaway for people should be the idea of *moral character*. Character isn’t destiny, but it creates a strong presupposition. If you want to know what people are doing, you should try to learn more about who they are. /end
- Replying to @geoffandersThe next question was about the extent of the fraud. Here I employee a heuristic, which is that bad actors are typically *much* worse than people expect. This is because people typically expect others to behave like normal people, or like others from their experience.
- Replying to @geoffandersHere’s where some more character reads and knowledge of ideology came in. Many EAs are kind-hearted. They want to believe the best about people… to a fault. Other EAs like to prove how smart they are and can be bullied around by math. (And by reports, Sam was a bully.)
- Scientific advance seems automatic to many people. Like it will just happen, a wave pressing inevitably forward. The more I study the history of science, the less automatic it seems.
- OpenAI’s GPT-3 API automatically generates shockingly high quality text. Here are some samples: read-the-samples.netlify.app What are the societal ramifications once automatic text generators are widely available? Big list of ideas in the thread! h/t @nosilverv for prompting this
- Replying to @geoffandersUnusual people, though, often are *very* unusual, and will do things that normal people won’t expect. This is usually harmless, and sometimes really good. In the case of bad actors though, this means that they sometimes are willing to go much further than people will think.
- Really excited for Tyler Cowen and Emergent Ventures to be supporting our work. Let's figure out where the bottlenecks are to scientific and technological advance and see if we can break them!Honored to be part of this cohort. Thank you to @tylercowen and Emergent Ventures for supporting our work.
- Replying to @geoffandersHowever, if there is *repeated* *agentic* bad behavior, that means bad character. Bad character implies a disposition to act wrongly, given the opportunity. So the next question was whether Bankman-Fried had the opportunity.
- What I want: A news service that lets me know where we are in a story and when the story ends. “Three Gorges Dam still holding, but there is still concern.” “Rioting in cities X and Y has ended.”
