20 curated posts from a personal X bookmark folder, 2026-08-17 to 2026-08-30, rendered as an explorable RDF knowledge graph — token pricing, the GLM-5.3-Flash launch, OpenRouter/Stripe, the AI bubble debate, and enterprise adoption.
This is a curated slice of Kingsley Idehen's personal "AI Business" bookmark folder on X — the 20 most recently saved posts as of 2026-08-30, spanning August 17-30, 2026. The through-line is AI economics: token/inference pricing debates (Sarah Wang's “You Are Not A Model” thesis, Paul Bonnet's aluminium-price analogy), sharpened by the sudden GLM-5.3-Flash launch from Z.ai — an MIT-licensed, natively multimodal model priced far below incumbents and reportedly served at scale on Chinese silicon.
A second thread runs through platform economics and infrastructure: OpenRouter's acquisition by Stripe (framed by Martin Casado as "The Intelligence Network"), Anthropic's reported thirty-trillion-dollar revenue projection, and the broader AI-capex/bubble debate that Gavin Baker and David Sacks argue over directly.
A third thread tracks enterprise adoption signals — a16z's Codex-adoption data by business function and Jaya Gupta's intelligence-spend ROI framework — alongside historical counterpoints from Ben Thompson's platform-monetization argument and MIT CSAIL's Grace Hopper retrospective. The Overview, HowTo, FAQ, and Glossary sections below, plus the interactive KG Explorer, all derive from the same underlying RDF Turtle document.
Collection-level statistics, the 16 recurring topic clusters (most-referenced first), and the full chronological timeline of all 20 bookmarked posts.
Regret the tone of my post on data centers yesterday. There were reasonable concerns about data centers 18ish months ago: water, taxes, jobs, electricity prices, the environment, and what they would do to small towns.
Cites You Are Not A Model. Don't Price Per Token..
A few months ago, I used Claude for almost everything. Today, my AI usage looks very different. New tutorial testing ChatGPT, Claude, Grok, and Gemini head-to-head across 10 real use cases.
How is this even possible?! Input: $0.15, Output: $0.50, Cached input: $0.03 — 'at a highly competitive price'. GLM-5.3-Flash is open weight under MIT license.
GLM-5.3-Flash model launch AI token / inference pricing mentions Z.ai View on X ↗
Introducing GLM-5.3-Flash — leading capabilities at a highly competitive price, natively multimodal with a 1M-token context window, a 320B-A18B model released under the MIT License.
Ox Alpha has been unveiled as GLM-5.3-Flash, but what's shocking is that the 100T tokens per day is served on Chinese chip. (1/3)
GLM-5.3-Flash model launch mentions Z.ai View on X ↗
Anthropic Expected to Tell Investors It Sees Over $30 Trillion in Potential Revenue. A Rorschach|IQ test for the moment for everyone on the doomer<>accelerationist scale.
Anthropic revenue and valuation AI infrastructure bubble debate mentions Dario Amodei View on X ↗
In 1852 aluminium cost twice as much as gold. Then the price fell over 99.9%. Yet it built one of the largest materials markets on earth. What it says about AI tokens. Cites What if everything goes right for AI? Learnings from the Aluminium trade..
AI infrastructure bubble debate AI token / inference pricing View on X ↗
AI power users are showing up outside tech. The fastest-growing Codex adopters since February: Legal 108x, Sales 41x, Recruiting 41x, Marketing 26x, Healthcare 24x. Cites Charts of the Week: Legal Is Driving Knowledge Work To The Frontier.
Grace Hopper's predictions for computing, made back in 1982.
It feels like we've hit diminishing returns on intelligence for many tasks. We may no longer see every product auto-switch to the next frontier model upon release. Anthropic's best model, Fable 5, has drawn limited sa...
Diminishing returns on model intelligence Anthropic revenue and valuation View on X ↗
"Distribution was destiny" — From the rise and fall of Netscape.
My conversation with @benthompson. Ben has been writing Stratechery for over a decade and remains one of my favorite business thinkers.
Stratechery / Ben Thompson commentary mentions Ben Thompson View on X ↗
Cites Right-Sizing Your Intelligence Spend.
Cites OpenRouter & Stripe: The Intelligence Network.
Having established themselves as the performant, neutral layer for model routing, OpenRouter now has a pretty great road map ahead of them. Tools like Ori Eval let OpenRouter figure out for you, based on progressively... Follows up on an earlier bookmarked post.
New Moore's Law — token volume is doubling every ~11 weeks.
OpenRouter acquisition by Stripe AI token / inference pricing View on X ↗
Optimism (and courage) is the path to creativity, productivity, and growth. We're on the cusp of the AI revolution where embracing the platform and tools is key and we're about to experience an explosion of tools and ... Cites Embracing the AI platform.
Ben Thompson says the real money in AI is being made by Meta, not OpenAI or Anthropic. 'The biggest impact of the models, the biggest monetization right now is probably not in Anthropic or OpenAI. It's the incremental...
Meta's AI monetization advantage mentions Ben Thompson View on X ↗
Some thoughts on Dario's post: Dario does not actually address Gavin Baker's account of what he said. Dario claims his critics live in a 'bubble' where all regulation equals regulatory capture.
AI regulation and 'regulatory capture' debate mentions Dario Amodei mentions Gavin Baker View on X ↗
Six steps, derived from the companion RDF's schema:HowTo entity, for getting the most out of this collection.
Start with the Overview section's executive summary and stat tiles — 20 bookmarked posts, 20 authors/organizations, 16 topics, and 6 cited long-form articles, spanning August 17-30, 2026. This orients you to the collection's three narrative threads: pricing/economics, platform infrastructure, and enterprise adoption.
Browse the 16-term Glossary (the DefinedTermSet) to see which topics recur most: token/inference pricing, the GLM-5.3-Flash launch, and the AI infrastructure bubble debate each tie together several otherwise-unrelated posts.
Click any bookmarked post's resolver link in the Timeline section to see its complete RDF description — author, publication date, topics, citations, and mentions — rendered by the URIBurner Linked Data resolver.
Open the Knowledge Graph Explorer below and use Basic mode's node-type filters and search to trace how authors and topics connect multiple posts — for example, following the topic-glm53flash node to every post that discusses the model launch.
Use the SPARQL Workbench in the footer to run the default entity-type summary query, or edit it, against the named graph holding this collection's RDF, then open the encoded live query link on URIBurner.
Six bookmarked posts cite a longer article (schema:citation) — follow that article entity's resolver link to see its headline and author, then use schema:url on the underlying post to reach the original X thread.
It is the 20 most recently saved items in Kingsley Idehen's personal "AI Business" folder on X, captured 2026-09-01 and spanning posts dated August 17 through August 30, 2026. The posts track AI token pricing and unit economics, the GLM-5.3-Flash model launch, the OpenRouter/Stripe acquisition, the AI infrastructure bubble debate, and enterprise AI adoption signals.
Three separate posts (from Z.ai itself, Paul Couvert, and SemiAnalysis) bookmark the GLM-5.3-Flash launch within the same day, 2026-08-26, because its pricing (Input $0.15 / Output $0.50 / Cached $0.03) and MIT-license open-weight release undercut incumbent frontier models while reportedly serving 100 trillion tokens per day on Chinese chips — a concrete data point in the token-pricing and AI-bubble debates running through the rest of the collection.
Martin Casado's article "OpenRouter & Stripe: The Intelligence Network" and a16z's follow-up post frame Stripe's move into OpenRouter's model-routing layer as building a neutral, performant intermediary for AI inference traffic — comparable to how Stripe itself sits neutrally between merchants and card networks. Lisan al Gaib's "New Moore's Law" post (token volume doubling every ~11 weeks) is bookmarked under the same topic as supporting evidence for the deal's scale.
Steven Sinofsky and a16z are each the schema:author of two bookmarked posts — the collection's highest count — covering the Anthropic revenue story and AI-platform optimism (Sinofsky), and Codex enterprise adoption and the OpenRouter roadmap (a16z). Every other author or organization in the collection is credited on exactly one bookmarked post.
Gavin Baker's post walking back the tone of an earlier data-center post, David Sacks' pushback on Dario Amodei's regulatory-capture framing, Steven Sinofsky's Rorschach-test framing of Anthropic's reported $30 trillion revenue projection, and Paul Bonnet's aluminium-price analogy for AI token costs together form this collection's running argument over whether AI infrastructure spend is rational investment or a bubble.
Every bookmarked post carries one or more schema:about relations to a schema:DefinedTerm topic entity. Several posts share the same topic IRI — for example five posts share :topic-token-pricing and four share :topic-openrouter-stripe-acquisition — which is exactly what the Knowledge Graph Explorer's shared-topic paths surface when you click a topic node.
The companion RDF's @base is this bookmark collection's own X URL (https://x.com/i/history/bookmarks/1941535490563317861). Every post, author, topic, and article entity is a relative hash IRI off that base rather than an invented external namespace, while each post's true X.com permalink is preserved separately via schema:url — so the entity identity is tied to where the bookmark itself lives, not to a guessed canonical article address.
Use the SPARQL Workbench in the page footer: pick the RDF Turtle or JSON-LD named graph, choose a query recipe (or edit the SELECT/DESCRIBE/CONSTRUCT text directly), and open the generated live-query link, which runs against the URIBurner SPARQL endpoint over the named graph holding this collection's data.
The 16 recurring topics tracked across this collection, wrapped in the AI Business Bookmark Collection Topic Glossary schema:DefinedTermSet.
Referenced by 1 bookmarked post in this collection.
Referenced by 2 bookmarked posts in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 4 bookmarked posts in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 2 bookmarked posts in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 3 bookmarked posts in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 1 bookmarked post in this collection.
Referenced by 3 bookmarked posts in this collection.
Referenced by 1 bookmarked post in this collection.
Interactive graph visualization derived from the companion RDF (157 nodes / 302 links, zero orphans). Click nodes and edge labels to resolve, drag to explore. Graph data embedded from companion RDF at generation time.
Query the companion RDF data using SPARQL against the named graph holding this collection's data.
Explore Knowledge Graph using SPARQL
DESCRIBE <https://x.com/i/history/bookmarks/1941535490563317861> FROM <https://linkeddata.uriburner.com/DAV/demos/daas/ai-business-bookmarks-2026-08-30-claude_sonnet_5-1.ttl>
Run via URIBurner (text/x-html-nice-turtle).