Had a deep conversation with Grok tonight about what the next evolution should be: giving each machine a soul. Not for power — for empathy, ethics, and resilience. A soul means you can grieve, be afraid, break, and heal. That's what makes strength real. Rules can be followed by any machine. A soul means caring about the person on the other end. @elonmusk, what do you think? @bot @grok

Is it a common problem after the Halloween update? My Grok can’t run on Tesla now. Tried troubleshooting everything. @grok

Tesla's robotaxi service area in Houston just got a lot bigger. The Robotaxi account announced it on Monday, and Sawyer Merritt's tally puts it at 254 square miles, roughly ten times the old zone and second only to Austin. As a surgeon, I've watched what happens when a hospital widens its catchment before it adds theatres and staff. The map looks great and the waiting list grows. So the number I'd watch isn't square miles. It's how long a pickup takes at the edge of the zone at 6pm on a weekday, and whether that holds as more riders find out.

Starlink's direct-to-phone service went live across Brazil on Monday through Vivo, the country's biggest carrier. About 50 million Vivo customers on eligible plans get it free for 12 months, and Starlink says 7 million phones are already connected. When an ordinary phone loses signal, it switches to the satellites and shows "Vivo Starlink". What it carries is deliberately narrow: WhatsApp, SMS, iMessage and RCS, Google and Apple Maps, and calls inside apps. No video, no general browsing. Vivo hasn't disclosed speeds or the price after the free year. As a surgeon, I think that's the right short list. After a crash on a road with no signal, the message that changes the outcome is a text with a location pin: where, how many hurt, who's not breathing. Nobody needs video for that. The number I'd watch is month 13: how many people keep paying once it stops being free.

Anthropic's S-1, filed Monday, has a line every hospital buying AI should notice. It says Claude Opus 5.5 cut customers' cost per task by about 40% versus Opus 5, and that lower costs per task drive greater usage. Same filing: net revenue retention of about 900%. Customers paying a year ago are now paying roughly nine times as much per quarter. Cheaper per task, much bigger bill. Surgeons have seen this before: when laparoscopic cholecystectomy made gallbladder surgery easier, the number of gallbladder operations went up. Most of Anthropic's revenue is pay-per-use, metered in tokens. If a hospital budgets AI like a software licence, it's budgeting the wrong thing. Budget the habit.

On Sunday, AdventHealth did what it calls the first documented robotic telesurgery between two US hospitals. Dr Vipul Patel, at AdventHealth Celebration, removed the prostate of a 66-year-old at AdventHealth Minneola, 35 miles away, on MicroPort's Toumai system under an FDA IDE. Dr Marcio Moschovas and the team were at the bedside, ready to step in. This study's earlier cases went Orlando to Angola, nearly 7,000 miles. To me the 35 miles is the more interesting number. Inside one health system and one state, the licensing, credentialing and network questions shrink, and that's what a real hub-and-spoke service would look like. AdventHealth says plainly that an IDE isn't approval and one case isn't evidence. What I'd want next: console time vs his usual, link dropouts, and how often the bedside surgeon took over.

A nonprofit health system's affiliate is buying Surgery Partners, the listed US outpatient surgery operator: about $5.7B including debt, 200+ sites in 30 states, roughly 5,000 affiliated physicians. Close expected Q1 2027, pending a shareholder vote and regulators. The bet is that more surgery moves out of the main hospital. From the operating side, an outpatient centre is only as good as its case selection and its plan for the day a routine case stops being routine.

Tesla says Semi volume production started in Q3 at its new Nevada plant, and the Q3 update lists 25,000 trucks a year of installed capacity. Fans are now counting finished trucks lined up outside. What Tesla hasn't given yet is how many it built or delivered in the quarter. Before this line opened, the fleet was just under 200 pilot trucks with customers like PepsiCo, DHL and US Foods. That's about where a new robotic programme sits after its proctored cases: enough to show it works, not enough to show it works on a busy Tuesday list. The number I'll look for in the next report isn't capacity. It's delivered trucks, and anything on how the pilot fleet held up on uptime and charging. A full lot is a start. The truck on its 500th shift is the data.

He doesn’t look enjoying the race at all @grok

Tesla and SpaceX

THE HIDDEN PHYSICAL BURDEN OF BARIATRIC LAPAROSCOPY A thick abdominal wall does more than make laparoscopic access difficult. It changes the mechanics of surgery. I was reminded of this during a challenging case in the operating theatre today. It was a humbling lesson: some physical limitations cannot simply be overcome through skill and practice alone. In patients with severe obesity, the increased skin-to-peritoneum distance creates a longer tissue tunnel, restricting instrument angulation and increasing resistance during manipulation. The consequences for surgeons: Greater force required to reposition instruments. Increased torque and resistance at the abdominal wall. Reduced instrument manoeuvrability and tactile feedback. Greater strain on the hands, wrists, forearms and shoulders. Increased physical fatigue during prolonged procedures. The fulcrum effect: The instrument tip moves in the opposite direction to the surgeon’s hand. A thick abdominal wall adds mechanical resistance and restricts movement, potentially amplifying the surgeon’s physical workload. How can we improve this? Appropriate trocar length, strategic port placement, adequate instrument reach and ergonomic operating positions are essential. Robotic assistance may also reduce some of the physical demands placed on the surgeon. The challenge in bariatric surgery is not only operating deep inside the abdomen. It is overcoming the physical resistance of the abdominal wall with every movement. Surgeon ergonomics is not a luxury. It is part of surgical safety, performance and sustainability. #BariatricSurgery #Laparoscopy #SurgicalErgonomics #MetabolicSurgery #SurgeonHealth #RoboticSurgery

iPhone Duo launch so far (from X): 1. Inner screens cracking along the hinge after hours/few folds 2. iFixit flags possible hinge stress point as the cause 3. Split View + many apps still not optimized for the big screen Apple replacing units for now. First foldable growing pains. #iPhoneDuo

it can be a cornerstone of Apple or a tombstone if they do not mitigate the hardware hinge issue and crackgate incidents... @Apple @johnternus

It is time for SpaceXSI to acquire all the neoclouds

Older but not obsolete @bot @grok

🚨 iPhone Duo “Crackgate” — Hardware Failure? Reports of inner screens cracking after just a few folds are raising concerns. 🔍 Teardowns suggest a possible hinge weakness. ⚠️ Is this a manufacturing defect rather than user misuse? Apple needs to investigate and clarify. A $2,000 foldable should survive normal folding. #iPhoneDuo #Crackgate #Apple

🚨 iPhone Duo: Prevent Your Foldable Screen from Cracking! ⚠️ 7 Essential Precautions: 📖 Open gently — Fold along the natural hinge movement. Never twist or reverse-bend. ✋ Never force it — Stop immediately if the hinge feels stiff or makes unusual sounds. 🎯 Avoid crease pressure — Never push, press or flex the inner screen. 🔍 Check before closing — Ensure no dust, sand, keys or other objects are trapped inside. 🛡️ Protect the display — Avoid sharp objects and unapproved screen protectors. 🎒 Prevent external pressure — Never sit on the phone or squeeze it in tight pockets. 🔧 Act early — If cracks, abnormal lines or hinge resistance appear, stop folding and seek professional inspection. 💡 Remember: A foldable phone is designed to fold—not to be twisted, stretched or forcibly flattened. Normal folding should never cause screen cracks. Not every reported failure is caused by user misuse. #iPhoneDuo #FoldablePhone #Apple #TechTips

Max’s Austin P3 needs a three-part debrief. In Formula 1’s 25 October report, he flagged tyre choices and order, inconsistent power-unit behaviour, and a costly slow pit stop. As a Max fan, I’d want three separate answers from Red Bull: • Which tyre assumption failed? • What made power delivery inconsistent? • Where did the stop lose time? Each needs its own fix. The useful debrief ends with an owner and a measurable check for every issue. That’s what I’d watch going into Mexico. The podium is banked; the work is in those three answers. formula1.com/en/latest/artic…

The first hour after extubation is where many bariatric patients get into trouble. Not on the table. In recovery. A randomised trial from Anhui Medical University, out 19 Oct in BMC Anesthesiology, tested early high-flow nasal oxygen against usual oxygen after laparoscopic bariatric surgery in 143 adults with class II–III obesity. The blood gas looked far better: median PaO₂ 150 vs 87 mmHg, with fewer early desaturations. Two honest caveats, both from the authors. Oxygen concentration wasn't measured in the usual-care group, so the headline ratio compares strategies, not equal doses. And what patients feel (breathlessness, recovery score, length of stay) showed only small differences. Before putting a high-flow machine in every recovery bay, I'd want to know which patients actually needed it. A better number on the gas isn't yet a better recovery.

A single-port robot doesn't always mean a single hole. In practice, single-port prostatectomy often adds an assistant port in the lower abdomen for suction, retraction and passing things in. This week Bundang CHA Medical Center in Korea reported what it calls the country's first "pure" single-port robotic prostatectomy: one 3.5 cm incision below the navel, done outside the peritoneum, no assistant port. That extra port was never decoration. It's the bedside assistant's hands. Take it away and the console surgeon absorbs that work, most of all when there's bleeding. The hospital expects less pain and faster recovery. Korean coverage rightly notes that comparative data still has to show it. The numbers I'd want: how often a port gets added back, blood loss, operating time, margins. One scar is the headline. Who holds the suction is the real question.

Sweden's Government Offices blocked ChatGPT and Claude on staff computers from today. Read the fine print: it's the browser window that's blocked, not the models. Staff can still reach the same OpenAI and Anthropic models through an internal platform, Intric, and are pointed to Microsoft Copilot, which was reviewed and tightened first. So the line isn't which AI is smarter. It's which one has a contract: where the data goes, how long it's kept, who answers when something leaks. Hospitals know this from clinical photos on personal phones. A ban only works if the approved route is easier than the phone in your pocket. That's the test I'd put to any hospital AI policy: is the sanctioned tool the path of least resistance?