I’ve been seeing this “high-trust society” dog whistle a lot lately, and I think it’s one of the funniest of its kind. You truly want me to believe that the United States, a country with a history of slavery and segregation, a country that went through a historical period dominated by people literally called “robber barons,” was a high-trust society before immigrants from less industrialized places came and ruined that?
CarPlay enables a car-friendly interface with big buttons, shortcuts to navigation/music/communication, and minimal distractions. I don’t know of any way to get that on the phone’s own display.
What's up with the AI narration at the beginning? Or is it just someone with an incredibly steady voice and AI cadence? It's uncanny and weird considering that this is a podcast hosted by actual humans!
The search field loses keypresses on mobile. I haven’t looked at the code, but I’m assuming it uses a React-style value binding but has some synchronous processing before it propagates the new field value back into the state variable. That is a really terrible pattern.
That's 10 cents for a single 35-calorie corn puff, to be clear, and it was a 25% price increase. 10 cents for one corn puff is not actually a good price.
They're willing to admit "5 to 6 days a week" in writing [1]! Crazy stuff. Also a notably huge number of job openings, including for a head of HR [2]. Worrying signs, I would say.
Well, I'd never heard anyone make it before, but sure. (I looked into Mercor a bit and know some people who've worked in data generation/labeling, which is what exposed me to that side of the operation.)
It doesn't absolve them of any theft, but it does make the assertion that they should be required to release their models to the public seem, to me, a bit farcical. There are dozens of free and open-weights models that have all trained on exactly the same web crawls and books as GPT-5 and Opus. The proprietary models are better because of proprietary data.
The vast, vast majority of submarine cable projects are paid for with private money. I'm only aware of a few in the US that have received any public grants (all connecting relatively unprofitable areas like Hawaii, kind of analogous to the Essential Air Service program).
The core of the training data is public, but the part that actually makes these models smart came from (pretty highly-paid) experts via platforms like Mercor. Claude didn't magically learn to write good code by reading all of GitHub - humans trained it in that, more or less manually.