jmerz

Rice cooker.

The only thing that would make this whole saga better is if OpenAI actually had a baby machine god in its basement right now and meanwhile the entire organization is collapsing to humans being human right above it.

As of yesterday this situation has crossed into omnishambles territory.

Sunday you could still have reversed this and treated it as a blip. Now it's Tuesday—there's not really an option anymore that doesn't cause some kind of lasting damage.

I think he's not as known in the outside world but it's really difficult to understate the amount of social capital sama has in the inner circles of Silicon Valley. It sounds like he did a good job instilling loyalty as a CEO as well, but the SV thing means that the more connected someone at the company is to the SV ecosystem, the more likely they like him/want to be on his good side.

This is kind of like the leadership of the executive branch switching parties. You're not going to say "why would the staff immediately quit?" Especially since this is corporate America, and sama can have another "country" next week.

This is certainly making a case for the fallibility of human leadership.

I'm hacking on some GPT-for-long-form-text stuff right now and it is _eye wateringly_ expensive once you start generating at anything close to "professional human" token outputs. $80 per month sounds already pretty optimized.

Feels like they've been trying and failing to grow past their core product flow (send pics to friends) for a decade now.

The intersection between politically/socially unpopular and low impact.

We're far more likely to pump a bunch of sulfates into the air than we are to stop driving pickup trucks everywhere. Humans aren't rational, and this is the kind of irrationality that democracy in particular cannot solve.

So there's actually a lot of academic debate on the merits of reparations, and exactly what and how much reparations should be.

A very oversimplified pro argument: if it wasn't for slavery, these families would have generational wealth and better social situations. African Americans in the US ARE disproportionately lower wealth/income and this has CLEAR historical origins.

The oversimplified con argument: Okay, but if you come from a wealthy African American family, why should you have a leg up over a poor (or otherwise more disadvantaged) white student? What about an immigrant, who didn't benefit from slavery at all?

Fundamentally there's a huge swath of different injustices across society, and we obviously can't fix all of them at once, so a big challenge in this sort of debate is how you slice the injustices and how you prioritize fixing them.

You’ll absolutely nab me with a wanikani integration. I think the biggest challenge for these learning apps with me is that nothing talks to each other and everything wants me to use it in a silo.

This captures it well. We're at the "startups are throwing GPT at every possible wall to see what sticks" stage of this.

We're going to see both improvements in application, and parallel improvements in the underlying model.

Who cares if it's AGI if someone figures out how to turn it into a competent tax accountant?

As someone who currently spends time on two continents it's always jarring to return to the USA. The first thing that hits you is the wealth. The second is the wealth disparity. The third is the dysfunction.

It's truly a remarkable place but often feels like a rocket ship pulling itself apart at the seams—a lot of exceptional velocity and a ton of social problems that just no one can seem to get a grip on.