🕶️ Stripe x PayPal: frenemies fund Xflow

Stripe and PayPal Ventures just backed the same startup, Xflow, in a $16.6 million round led by General Catalyst, giving the company an $85 million valuation—a rare moment where fintech rivals share the same cap table. The bet is simple: cross-border payments for Indian exporters are still slow, expensive, and full of paperwork, and Xflow wants to fix that by plugging payment rails directly into the software businesses already use. If it works, more payment volume could move away from traditional banks and wire transfers.

Banks may respond with better pricing and faster FX, while players like Wise and Payoneer fight to protect their customers. For founders, the opportunity is offering clean compliance and predictable fees; the risk is sudden regulatory changes or relying too heavily on partner banks.

We’re being lied to about 4K.

Most movies labeled “4K” today were actually mastered in 2K — then upscaled.
Some of your favorite films — Lord of the Rings, Harry Potter, Mad Max — were finished on 2K Digital Intermediates. That means the real master file never had true 4K detail to begin with.

So what are you really paying for?

In this episode, we break down how movies used to be cut from physical film, how Digital Intermediates changed everything, and why resolution isn’t what actually determines image quality. We explain color spaces (Rec709 vs P3 vs Rec2020), bitrates, Blu-ray vs streaming compression, and why your “4K” stream might look worse than an old 1080p Blu-ray.

This isn’t just about pixels.
It’s about color, compression, and the trade-offs the industry made when 4K became easier to market than quality.

And yes — there are ways to get closer to the original film experience at home.

Anthropic says three China-based AI labs may have tried to replicate Claude by running a large-scale “distillation” operation, allegedly using around 24,000 fake accounts to systematically harvest its outputs and train competing models, a claim that surfaces as Washington considers tighter AI export controls and quickly turned into meme fodder online.

If true, the fallout could reshape the API economy: providers may tighten KYC, introduce stricter rate limits and per-seat pricing, deploy more forensic detection and watermarking, and pursue legal action. More broadly, distillation weakens the moat around closed models, while chip bans push competition away from raw training compute and toward inference access and synthetic data.

For builders, that means higher operational risk—sudden caps or tougher onboarding—while China-based labs could face expanded sanctions or service cutoffs, setting up what may look less like a GPU arms race and more like a compliance and captcha battle.