Want to know why OpenAI's safety team imploded?

Here's why.

Thank you to the company insiders who bravely spoke to me.

According to my sources, the answer to "What did Ilya see?" is actually very simple...

vox.com/future-perfect… People speculated that Ilya saw AGI...as if OpenAI was hiding some conscious, shackled AI in the basement.

But reporting this out, I thought: This is not a horror story about AI. This is a horror story about humans.

Like in 2001: A Space Odyssey, the issue wasn't HAL lying.

You’ve probably been hearing lots about Israel — but not about Jews with roots in the Arab and Muslim world. They’re over half of Israel’s Jewish population, yet the American media barely covers them.

So let me tell you a story about my family. 🧵 1/18

vox.com/world-politics… My dad’s side is from Iraq, where Jews lived for 2,000 years and were deeply integrated into Arab society. Jews spoke Arabic and made up 1/3 of Baghdad’s population. We were everywhere — in parliament, the judicial system, the music scene. Here’s what my family looked like. 2/18

For anyone who doesn’t understand why Palestinians have every reason to fear they are now facing permanent expulsion:

Here is a thread with quotes from the Israeli defense establishment this month. No comment. Just direct quotes.

Here’s the first one. Here's the second quote.

I hope effective altruism learns some serious lessons from the fall of its billionaire bankroller @SBF_FTX.

The FTX scandal should get EA to question its philosophical foundations, not just do a cosmetic makeover.

Here’s 6 reasons for that 🧵

vox.com/future-perfect… @SBF_FTX 1) EA skews heavily utilitarian. It teaches people to maximize the overall good. That’s a dangerous ethos unless you’re a god who somehow always knows what the good looks like. Per Holden Karnofsky:

Effective altruism’s most controversial idea is called “longtermism.”

It says we should focus on protecting FUTURE people (potentially more than present people).

It’s a deeply political idea, so the main question is: Who gets the POWER to decide?

🧵
vox.com/future-perfect… The first thing to realize is that there isn't 1 longtermism. There's longtermisms. Think of this worldview as a train that can drop you off at different stations.

Effective altruists sometimes talk about this by asking each other: “Where do you get off the train to Crazy Town?”

You may have heard that AI is biased because the data it’s trained on is biased.

That’s really just the tip of the iceberg.

The fairness crisis goes much deeper than that. 🧵 1/8
vox.com/future-perfect… Thing is, there’s no 1 definition of fairness. Fairness can have many different meanings — at least 21 by @random_walker’s count! — and those meanings are sometimes incompatible with each other. 2/8