YouTuber Markiplier broke down in tears after his self-financed horror movie "Iron Lung" opened to $23 at the worldwide box office: “Fuuuck dude this movie cost so much money I'm so fucked”
I know it’s become pretty cliche and cringey to talk about at this point but if you’re under like 25 I cannot stress enough how one time Obama wore a tan suit and people spent a week arguing over whether or not it was demeaning to the Oval Office and they were serious about it.
The best part about Hunter Biden being overly online is how people in his replies go “How much COKE did you SNORT before POSTING THIS you FUCKING PEDO” and Biden responds “I didn’t snort coke I smoked it” and then they respond “Haha true that, sorry man, have a good one”
WTF timeline are we on. Someone called me the MAGA whisperer and I’ll gladly take the title. Left, right, D or R we all want the same things. We’re being divided on purpose by the Epstein Elite Oligarch class because as long as we’re at each other’s throats, they get fat and rich off of our misery. The second we figure out we agree on more than we disagree, they’re done. Love your neighbor. Be yourself. Radical honesty. No fucks given, no fucks taken. Everything else is just noise. (But still fuck Jake “Brick Tamland” Tapper on any time line)
Moneyball wasn’t really about baseball. It was about discovering that everyone else was allocating capital inefficiently. Once Billy Beane proved a better model, the advantage was enormous. But the real lesson came later: when the Boston Red Sox combined the efficiency of Moneyball with vastly greater resources, they became a dynasty.
AI feels similar today. The industry is spending unprecedented amounts on compute. In image and video generation, the scale of GPU investment is staggering. When capital is abundant, efficiency often becomes an afterthought.
Two years ago, Reve started with a simple constraint: we couldn’t outspend the largest labs, so we had to outthink them. That constraint forced invention. While compute and data matter enormously, architecture still matters. How you build the system matters. Efficiency matters.
Visual generative models today use text as an intermediate representation, leveraging LLMs to plan their outputs before rendering any pixels. Natural language is expressive, but it is ambiguous — and ambiguity is the enemy of efficiency and control. Reve made a different bet. We replaced text with a better, code-like semantic representation — a layout. Layout is the reason we can compete with, and beat models trained on 10× our compute. It also opens up a whole new world of precise, non-verbal visual control.
Today, we have proved that our bet paid off. A team of 50 people has built one of the world’s best visual generative models using a fraction of the training compute of our much larger competitors.
The next chapter is even more interesting. What happens when a system designed for efficiency is given access to substantially more compute? It’s not about doing more with less. It’s about building a better machine first, then scaling it.
The broader AI race may be much earlier than most people think.
I always found it strange that trans people were considered more sexual than the average person because a large proportion of erotic trans fiction has the plot of "What if I was a girl and people were nice to me"
@CyMorgComics until you get the person who either did or didn't text you after hours of deliberation now convinced that if they chose the opposite you wouldn't be in the exact position necessary to get hit
Since FIFA said they'd be in touch about making this official, I just gotta say that I really couldn't think of anyone better to represent the World Cup than the guy who's spent the last three years promoting and learning the world's cultures and breaking stereotypes around them
@ShitpostReels Do I look like I'm gay, or do I look like I can't be trusted around your mother, your sister, your auntie, your wife, hell even your grandmother