B-TIER news
Open vs closed token share — Friedberg on the 12-week flip and Anthropic risk
100k ft — What is this?
Open vs closed token share — Friedberg on the 12-week flip and Anthropic risk
50k ft — Why care?
David Friedberg on All-In says token usage flipped from 80/20 closed-to-open to 80/20 open-to-closed in 12 weeks, calling it Anthropic's biggest risk factor ahead of S-1s. He expects a bifurcated market: open weights for most use, premium models for hard engineering, math and biology work.
10k ft — Key takeaways
• Open weights are spreading into robotics VLAs and image/video
• Open-weight models now dominate token volume per Friedberg's chart
• Premium closed models keep pricing power on hard technical work
5k ft — Tools & techniques
Anthropic
(no techniques tagged)
1k ft — Deep-dive / try it
⚠️ David Friedberg on the biggest risk factor for Anthropic:
“In the last 12 weeks alone token use has flipped from:
80/20 closed vs. open to 80/20 open vs closed
I don’t think we’ve ever seen in 12 weeks... in the history of all technology markets, we’ve never seen anything like this.
— David Friedberg on @theallinpod
#davidfriedberg #allinpodcast #opensource #anthropic #ai
Transcript excerpt
in the last 12 weeks alone, token use has flipped from 80-20 closed versus open, to 80-20 open versus closed. I don't think we've ever seen a tidal wave of technology. In the history of all technology markets, we've never seen anything like this. And that is the biggest risk factor for Anthropic, is they were first out the gate to get the enterprise on AI, to build tools and whatnot. 12 weeks, oh my God. If you look at this chart, this I think is what blows the lid off the door on what needs to go into these S1s. And that's why I walked through all those open weight releases at the start of the show or this segment, is it's not just that they're there, it's that they're proliferating now. There's so many more of them coming out and the pace of improvement is extraordinary. And they're now in VLA for robotics. They're in image generation for image and video. So the open weight, open source models are now being used for all of the broader AI applications. The premium stuff, I will say, is still the best. For narrow use cases. For engineering, for highly technical, highly complicated tasks and workflows and problems, solving Navier-Stokes, solving mathematical problems, solving biology problems. That's where these models absolutely perform and people will pay a premium of, I would say, close to infinity dollars for that premium application set. So it doesn't mean that their market is destroyed. It means that we're seeing a bifurcation in the market where the majority of use…