Decoding Discontinuity
Subscribe
Sign in
Home
What is Discontinuity?
ARAF
Format & Subscription Tiers
Agentic Era Series
AGNT Podcast
Speaking + Interviews
AGNT Manifesto
Anthropic
About
Latest
Top
Discussions
The AI Sun King Has Fallen. What Happens to Nvidia’s Empire After Jensen Huang?
What history says follows the fall of a king: the realm misses its next turn, the favorites change master, and credit moves to whoever has written the…
Nov 17
•
Raphaëlle d'Ornano
14
1
Anthropic IPO: Can Frontier AI Become a Durable Business?
The long-term case for Anthropic (ANT) rests on compute, frontier models and turning them into products. The first public year may be tough. Can the…
Nov 10
•
Raphaëlle d'Ornano
15
2
1
The Hyperscaler Settlement: Free Cash Flow Does Not Turn Positive Before 2030
Alphabet, Amazon, Meta and Microsoft have found ways to keep building. The larger wager is whether AI agents can create enough paid work to cover the…
Nov 3
•
Raphaëlle d'Ornano
10
2
October 2026
It’s All About Compute: Anthropic’s S-1 Leaves Its Biggest Bill Blank
The filing omits aggregate obligations. A gross sum of disclosed commitments and Anthropic’s capacity estimates approaches $1 trillion, before overlap…
Oct 27
•
Raphaëlle d'Ornano
25
3
2
Two Tenants Part II: Anthropic's Run Rate Doubles in 2027. The Work Behind It Quadruples.
My model puts Anthropic’s December 2027 run rate at $169 billion. Token volume nearly quadruples, while falling prices and agent competition shape who…
Oct 22
•
Raphaëlle d'Ornano
13
2
Is the AI Build-Out a Bubble? The 2000 Lesson Is About Debt, Not Chips
The AI build-out is being called a bubble. That’s true for one layer of it but wrong about the rest. The difference is a matter of which price you…
Oct 20
•
Raphaëlle d'Ornano
14
1
Two Tenants Part 1: Anthropic Does Not Have ARR. Neither Does OpenAI.
OpenAI calls its figure ARR; Anthropic calls its figure a run rate. Both annualize sales from products with different contracts and renewal risks…
Oct 13
•
Raphaëlle d'Ornano
16
1
The Other End of the Compute Trap: What Nscale's S-1 Shows About Anthropic's $45 Billion
Anthropic refinanced its single point of failure through its suppliers. Nscale's filing reveals what one of those suppliers is carrying, and what public…
Oct 6
•
Raphaëlle d'Ornano
17
1
September 2026
TypeSafe’s Jev: The Model That Does Not Talk — and Charges $0 for Output
Jev’s explosive growth points toward a split between low-cost decisions and frontier deliberation. That has consequences for AI pricing and who controls…
Sep 29
•
Raphaëlle d'Ornano
10
3
3
Weights Are the New Code
As open bases get good enough, the fight moves from “who has the smartest model” to “who can turn real work, with permission, into the next version of…
Sep 22
•
Raphaëlle d'Ornano
10
1
The AI Pause Is the Wrong Panic
Dario Amodei wants to pace production of frontier intelligence. That may be good for safety and very good for Anthropic. And it may, paradoxically…
Sep 15
•
Raphaëlle d'Ornano
8
1
GPT-6 Astra: OpenAI Pushes Frontier Models That Act But Explain Less
The frontier's new flagship acts instead of answers, competes with employment instead of software, and needs so little written reasoning that the next…
Sep 8
•
Raphaëlle d'Ornano
13
2
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts