Trump's Saudi nuclear deal is conditioning energy policy on Middle East peace, Israeli jets are still striking Lebanon despite the ceasefire theater, and the FDA is finally getting serious about reigning in medical AI.
Thu Aug 27
Sat Aug 29
Mon Aug 31
Tue Sep 1
Bills
Do today: Certify unemployment on Tuesday morning, first thing.
Trump's Saudi nuclear deal hits Congress with proliferation alarm bells ringing. The administration submitted a civil nuclear agreement with Riyadh this week, but there's a catch: it only moves forward if Saudi Arabia formally joins the Abraham Accords. This is leverage wrapped in energy policy. The real issue isn't the deal itself, it's what Iran-watchers are nervous about. A nuclear-armed Saudi Arabia, even for civilian purposes, reshapes the entire Middle East power balance while the U.S. is simultaneously crushing Iran with "economic D-Day" sanctions. You're watching a administration play 3D chess on energy, alliances, and deterrence all at once.
Congress gets final say here, and this timing matters. With the Strait of Hormuz still gridlocked over Iran tensions and billions in damage reported at U.S. intelligence sites from recent Iranian strikes, lawmakers are going to ask hard questions about whether arming a key U.S. ally in the Gulf waters down more risk than it reduces.
Quick hits:
Israeli jets strike southern Lebanon despite framework agreement, and we're back to square one. The US-brokered deal that was supposed to be a turning point is basically theater at this point, with reported strikes continuing across the south even as the ink dries on the accord signed in Washington.
This is the pattern we've been watching for months: agreement announced, violations begin immediately, international community expresses "deep concern," rinse and repeat. The difference now is that Hezbollah and Lebanon's state are openly competing for control, which means President Joseph Aoun's leverage to enforce anything on his side is... minimal. You can't police a deal when your own territory is contested.
The economics tell the real story: the World Bank is projecting a 6.4% contraction for Lebanon's economy. That's not a bump in the road, that's structural collapse. And we're still talking about 860,000 displaced people who've returned to homes without electricity, running water, or functioning hospitals.
S&P 500 +0.32%, Nasdaq +0.66%, Dow +0.3% — Tech bounced back after yesterday's sell-off, with chip names leading. Oil fell again, which eased bond-market jitters. Consumer confidence hit a seven-month low, but the market shrugged it off.
The game here is simple: everything hinges on Nvidia tonight. Options are pricing in a 5.4% move post-earnings, which translates to roughly $280 billion in market cap swing either direction. Investors are hunting for signals on AI demand, margins, and guidance. If Nvidia disappoints, it won't just crater the semiconductor complex, it'll cascade across the entire mega-cap tech rally that's been holding up equities.
The tension is real. We're caught between a hawkish Fed (inflation won't cool) and growth stocks that need yields to stay low. Yield pressure from bond vigilantes has been the biggest headwind for three weeks now. If Nvidia nails earnings and the AI thesis holds, yields could stabilize. If not, we're back to being trapped between inflation worries and a slowing consumer.
Quick hits:
Claude's skill system is finally becoming useful, and founders should pay attention. The AI Break just published a tutorial on installing six essential Claude skills, and the practical implications are wild. We're talking about a 49-skill marketing toolkit that lets you generate agency-grade work in minutes instead of waiting weeks for a consultant retainer.
The key insight here is progressive disclosure: skills fire automatically when they match your request, without cluttering Claude's context window like old-school prompts do. That means you can install 200 skills and they stay lean. More importantly, they can execute real code and produce actual files (PowerPoints, spreadsheets, PDFs) instead of just describing what they'd do. That's the difference between a tool and a crutch.
For NKDHS specifically, this matters because it's how Claude catches up to o1-style reasoning in practical deployment. If your competitors are shipping Claude integrations without skills enabled, they're leaving 80% of the capability on the table. One toggle in Settings > Capabilities, and suddenly your models produce clinical documentation templates, patient journey maps, and regulatory compliance checklists that actually work.
Quick hits:
FDA is finally getting serious about generative AI guidance. Rick Abramson, the agency's Digital Health Center of Excellence director, told STAT he plans formal guidance on gen AI-enabled medical devices, plus specialty guidance on complex AI topics. This matters because right now there's almost no guardrails, and every vendor from Epic to smaller startups is shipping AI features into clinical workflows with minimal oversight.
The timing is interesting, though. We're in a Trump administration that tends to resist heavy regulation, so don't expect the FDA to come in hot with restrictions. This will likely be enabling guidance, not binding rules. Still, any signal that the FDA is moving on this fast enough to publish before year-end would be a win for the entire field.
On the vendor front:
Masimo expands the hemodynamic monitoring footprint with FDA clearance for continuous oxygen delivery sensor. Masimo received 510(k) clearance for Root with LiDCO and the Masimo LiDCO Sensor, which measures DO2 (oxygen delivery) in real time. This is table stakes for ICU and perioperative settings where oxygen balance matters most.
The move deepens Masimo's integration play. They're not just selling standalone sensors anymore, they're building a cohesive Root platform ecosystem that hospitals can layer onto existing workflows. This marginalizes point solutions and makes it harder for competitors without a central hub strategy to compete on clinical breadth. It's the same playbook that's been working for them in brain function monitoring and patient safety.
For NKDHS, this is a reminder that the consolidated hemodynamic player wins market share. Masimo is bundling DO2, brain function, and non-invasive trending into one interface. Our differentiation has to come from either superior integration (which we have with many EHRs) or delivering outcomes these competitors can't match on their own.
Elsewhere:
The Witcher III is coming to Switch 2 on September 29, and this is the port that might finally prove the console can run anything. CD Projekt Red's remaster includes updated visuals, enhanced combat, and improved traversal, which means Nintendo's handheld is officially in the conversation with PS5 and Xbox for AAA ports. This alone would be huge, but the timing is perfectly savage: it lands one day before the console's $50 price hike kicks in, turning your last chance at the $450 model into a genuine decision point.
On the earnings front, Nintendo just reported Q1 saw Switch 2 push past 23 million units and operating profit jump 53 percent. The hardware price increase is working exactly as intended, even if it stings.
Also worth knowing:
Xabi Alonso's Chelsea debut ended in a pulsating 3-2 win over Fulham, with Cole Palmer looking like he remembered how to play football. Pedro scored after 31 seconds (Chelsea's fastest goal since October 2016), Palmer added a second and an assist, and Rogers chipped in the third. The problem: Robert Sanchez continues to be a goalkeeper liability, gifting Fulham two goals through basic mistakes. Alonso won't care about the defensive sloppiness if Palmer stays this sharp.
Palmer's performance matters more than the messy scoreline. After missing the 2026 World Cup following a difficult 2025 campaign, he was dominant at Craven Cottage, moving freely and creating chances with that slick timing everyone paid for. If that's the version Chelsea get regularly, the back-line issues become fixable.
---
Cal Raleigh's historic collapse finally has an explanation: The Mariners catcher is swinging at fewer pitches in the zone than last season, and it's crushing his production. Through 99 games, he's at 18 home runs with a .632 OPS, nearly certain to post the worst year-over-year drop for any 60-homer slugger in MLB history. Mark McGwire's 32-homer decline is the current bar. Eno Sarris broke down the mechanical breakdown so clearly that even a layperson walks away understanding the problem. Raleigh went deep twice in his last two games, but he needs 14 more dingers in 29 remaining games just to tie McGwire's mark.
---