Reminds me of the two hard things in distributed systems:
- 2: Exactly-once delivery
- 1: Guaranteed order
- 2: Exactly-once delivery
Reminds me of the two hard things in distributed systems:
Just cache the common questions.
There are only two hard things in Computer Science: cache invalidation and naming things.
FTC for the past 60 years: “You lied to us? Great job, kid!”
FTC under Lina Khan: “You lied to us? We’ll see your ass in court.”
We might not win, but golly it’s refreshing to see someone try.
Right? Dude’s whole premise is “omg he had to SCALE the models to get them to match up! That proves he faked the whole thing!!!1”
Get outta here with this dumb shit.
And “artists should be able to independently corroborate”? The fuck? Ripping assets is not always a straightforward endeavor for someone who is interested in it… why would you expect artists to automatically know how to rip assets for any game at a moment’s notice? (And spend their time and money doing so)
People that work there, yes. But it was next to a conference area which often had folks from out of town or external to the company. Dudes in suits who are just there to shake hands and give a presentation.
So imagine some highly honored guest chuckling with the CEO and then 5 seconds later opening the bathroom door to find a sweaty nerd in flannel flailing around at the urinal.
My work used to have the bathroom on a one-minute timer, so you’d end up in the dark if you were too still at the urinal.
So you have to choose how you want someone to perceive you if they walk in while you’re mid-stream:
I need help finding a source, cuz there are so many fluff articles about medical AI out there…
I recall that one of the medical AIs that the cancer VC gremlins have been hyping turned out to have horribly biased training data. They had scans of cancer vs. not-cancer, but they were from completely different models of scanners. So instead of being calibrated to identify cancer, it became calibrated to identify what model of scanner took the scan.