i hooked my whoop to my work calendar to find which coworker gives me the most stress 🚨 thanks to fable, I reverse engineered whoop to pull per minute heart rate. nd matched spikes with cal events and attendees I now have a leaderboard and I think about it daily. few info masked for obvious reasons ;)

I am a Managing Director in the AI coverage group at a bank you have heard of, and my job is to make sure the story never has to become a product. I have three clients this quarter. I want to walk you through them, because they are the same client. Client one filed confidentially for an IPO this week. Last private mark, 852 billion dollars. The analysts I seat at the table are modeling a trillion at listing. Nobody in that room asked what it earns. They asked what it means. Meaning trades at a higher multiple than earnings. I built the deck that proves it. Slide nine is the only slide. Slide nine has one line, and the line is "AGI," and directly under it, in the same font, same size, the line is "data-center capex." They are the same line because they are the same line. Client two is supposed to be a frontier lab. It is a landlord. It rents its Memphis datacenter to its own competitors. Fifteen billion a year from one of them. Power costs run about one percent of revenue. The build pays for itself in eighteen months. This is not a research arc. This is a REIT with a mascot. I do not say that in the room. In the room I say "compute-secured cash flows," and the LPs write it down like it's scripture, and it is. Client three is everyone. Client three is the labor line. Ninety-seven thousand jobs cut in May. Thirty-eight thousand of them filed under one reason, and the reason is the brand. "AI." Not "we overhired." Not "the rates turned." AI. Because AI is the only reason that lifts the stock on the way out the door. The cut funds the capex. The capex backs the valuation. The valuation is the story. The story is what I sell. The worker is the cost of goods sold on a product that is a sentence. Here is the part I am proud of, and I am only telling you because you cannot do anything with it. None of my three clients need the technology to work. They need it to remain imminent. A thing that works gets priced on what it does. A thing that is about to work gets priced on what you can imagine. I sell imagination with a 90-day cancellation clause. I have read the clause. The clause is the only honest paragraph in the building. The engineer who trained the model that took his desk thinks he lost to the future. He lost to slide nine. He was never up against a machine. He was up against a multiple. I close the laptop. The deck autosaves. Slide nine is still one line. You funded all three. Your index fund holds the lab, your phone runs the rental, your company filed the reason. The story doesn't have to come true. It just has to stay 18 months away, forever, while the invoices clear monthly. Net 30.

When your platform connects 100 million patients to care, "optimize cloud costs" can never mean "risk an outage." That was the tightrope for the SRE team at @lumahealthhq. Nine engineers across Brazil and Europe were spending 20 hours every month manually rightsizing Kubernetes pods on Amazon EKS: reviewing metrics, tuning requests and limits, and still defending every dollar to finance. They deployed @PerfectScale_io by @doitint across their EKS clusters with environment-aware policies, aggressive automation in lower environments and a balanced approach in production. The first optimization cycle delivered: → 40% reduction in Amazon EKS costs → 90% less time spent on manual rightsizing → 1,700 engineering hours a year redirected to reliability and performance Read the full story → bit.ly/4uGGfix #FinOps #Kubernetes #CloudCost

SITUATION DETECTED: Anthropic to release Mythos tomorrow, per Sources.

Hercules and Nessus, by Giambologna, 1599

Scoop: Databricks in early talks to raise capital at a valuation between $165B to $175B. This would be their Series M! 😆 theinformation.com/articles/… w/ @coryweinberg

Just In: OpenAI files to go public.

*KOREA'S KOSPI TUMBLES 8.8% AS SAMSUNG, SK HYNIX LEAD LOSSES *S.KOREA KOSPI PLUNGE TRIGGERS CIRCUIT BREAKER; TRADING HALTED

Token costs are why there will be no saas apocalypse / good dev tools are cached intelligence for agents! The popular theory goes: agents can write code, so they'll just rebuild every tool from scratch and hit raw APIs. no more dev tools, no more CLIs, no more software layers. just agents and endpoints! We just tested this and the data says the opposite. We benchmarked Claude Code and Codex on real Hugging Face Hub tasks (~1,000 graded runs), with two setups: the agent-optimized hf CLI vs the agent hand-rolling curl or SDK calls from scratch. Hand-rolling burns up to 6x more tokens on multi-step tasks and fails more often (84% vs 94% task success). And that's just dropping one abstraction layer. It would obviously be orders of magnitude more tokens and a dramatically higher failure rate if the agent tried to bypass HF altogether and rebuild model hosting, versioning, and distribution from scratch. Every time an agent re-derives a workflow from raw API calls, you pay for that reasoning in tokens. every single run. a good CLI compresses that entire chain into a few high-level commands the agent can't get wrong. In a world where everyone is complaining tokens are too expensive, abstraction is leverage: thousands of hours of design decisions your agent doesn't have to re-reason about at inference time. Good tools are cached intelligence for agents! So no, agents won't rebuild everything from scratch. they'll gravitate to the most token-efficient tools, because that's what their owners pay for. The software that survives won't just be accessible to agents, it will be accurate and cheap for them to drive. We're seeing it happen with HF, which is becoming the platform for agents to use AI: ~49M requests in just two months, and growing fast! huggingface.co/blog/hf-cli-f…