Google AI takes over hurricane forecasting

0:00Flowing ad-bunch it on metrics that look great, till the CFO sees them, that's Bullspend. And marketers are calling it out in dashboard confessions. I remember telling my boss, it'll be good for the brand when leads were slow. Yeah, it wasn't. Cut the Bullspend. LinkedIn lets you target by company, job title, and more. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit go to LinkedIn.com slash campaign term secretions apply. Call us footballers back. So, Hilton called to me the superstition concierge to make your fan rituals a reality. Need a room to match your lucky number? We got you. Want to make sure our team doesn't wash your lucky jersey? Oh, that smells lucky. Hilton's unmatched hospitality can keep up with any superstition. Even a marching bandwick up call it 555 and 55 seconds. Hit it! When you need a team that will do whatever it takes on game day, it matters where you stay. Hilton, for this day.

1:00I get so many headaches every month. It could be chronic migraine. 15 or more headache days a month. Each lasting four hours or more. Botox on a botchaline mtoxin A prevents headaches and adults with chronic migraine. It's not for those who have 14 or fewer headache days a month. Prescription Botox is injected by your doctor. Effects of Botox may spread hours to weeks after injection causing serious symptoms. A lurcher doctor right away is difficulty swallowing, speaking, breathing, eye problems, or muscle weakness can be signs of a life-threatening condition. Patients with these conditions before injection are at highest risk. Side effects may include allergic reactions, neck and injection site pain, fatigue, and headache. Allergic reactions can include rash, welts, asthma symptoms, and dizziness. Don't receive Botox if there's a skin infection. Tell your doctor your medical history, muscle or nerve conditions, including ALS Lugeric's disease, myosthenia-gravis, or Lambert Eaton syndrome, and medications, including botchaline mtoxins, as these may increase the risk of serious side effects. Why wait? Ask your doctor. Visit BotoxCronicMigraine.com or call 1-800-44 Botox to learn more.

2:00The National Hurricane Center just handed the tracking of a category 3 Atlantic hurricane named Azius entirely over to an artificial intelligence model built by Google DeepMind. Yeah, I mean, you look at a decision like that and you realize we are just completely bypassing traditional meteorological institutions now. Like right now, Alabama and Florida and Georgia are under a state of emergency. Amtrak has shut down routes down the east coast and offshore oil production in the Gulf has ground to an absolute halt. All of these massive physical systems are reacting to a line on a map drawn by a neural network. It just removes the traditional atmospheric physicist from the driver's seat entirely. We have millions of coastal residents packing their cars, draining gas stations, and boarding up their windows based on data generated by the exact same company that runs your search engine. Which brings up a question that I think sort of hangs over everything we are talking about today. When a major storm forces entire coastlines to evacuate,

3:05what happens to the physical world when the emergency response is dictated by a tech company's algorithm? Right. An Andrew Friedman and Michael Fomel have been looking closely at how we even got to this point. They point out that for years, predicting severe weather was essentially a soap opera of competing models. Oh, yeah. We had the European model and the American GFS and the German icon model. If you lived anywhere near a coast, you knew those names. Local news anchors would stand in front of green screens every hurricane season and just pit them against each other. Exactly. The Euro model says the storm hits Miami, but the GFS says it veers out into the Atlantic. People would basically place bets on which government agency had the better math, but that entire debate is just dead now. Completely dead. Google's weather next three is the undisputed leader in accuracy, and the forecasting community is no longer arguing about which traditional model is best. They are just watching Google system run circles around everything else. And the reason the traditional models lost really comes down to how they process information.

4:06A traditional physics based model relies on raw math. Meteorologists take the current temperature, bear-match pressure, wind speed, and they plug those numbers into incredibly dense equations governing fluid dynamics. They are quite literally trying to calculate the friction of air molecules moving against each other in the atmosphere. You need building-sized supercomputers crunching those physics formulas for hours to figure out where the storm is going. Yeah, they have to break the Earth's atmosphere into a grid of 3D boxes, then they calculate the physical state of every single box, and then calculate how that box affects the box right next to it. It is exhausting, mathematically. But Google's AI ignores the math entirely. It uses deep pattern recognition. They fed the system decades of historical weather data and trained it to recognize what happens next. It doesn't solve the underlying physics of the atmosphere at all. If you're not subscribed yet, take a second and hit follow on whatever podcast app you're using. It helps us keep making this. We appreciate you being here.

5:07It actually treats the weather more like a language translation problem. Right, like translating French to English. Exactly. If you translate French to English, you don't need to know the physics of how vocal chords produce sound. You just need to know that one pattern of words usually corresponds to another pattern of words. So Google's model looks at hurricane forming off the coast of Africa and says, well, based on the millions of weather patterns I've studied, this specific arrangement of clouds and pressure usually results in a category three storm hitting Florida in six days. And for the people working in meteorology, this is a profound cultural shift. You spend a decade in higher education learning atmospheric physics to predict the weather. Now your job is basically shifting from predicting the weather to interpreting the output of an AI. You essentially become a translator for a black box. You have to explain to the public why the storm is moving a certain way. Even though the tool they gave you the forecast doesn't actually understand the physical reasons itself, which means centralizing that much critical data requires a hard look at the vulnerabilities.

6:11We are taking something as unpredictable as severe weather and bottlenecking the entire emergency response through one corporate tech leader. You are putting the safety of millions of people on the assumption that this specific pattern recognition system won't encounter an unprecedented atmospheric variable that it is never seen in its training data. And climate change is constantly generating weather anomalies that have no historical precedent. Right. If the flowing ad budget on metrics that look great, till the CFO sees them, that's bullspend. And marketers are calling it out in dashboard confessions. I remember telling my boss, it'll be good for the brand when leads were slow. Yeah, it wasn't. Cut the bullspend. LinkedIn lets you target by company, job title, and more. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit go to LinkedIn.com slash campaign Terms and conditions apply. College football is back. So, Hilton called to me the superstition concierge to make your fan rituals a reality.

7:14Need a room to match your lucky number? We got you. Want to make sure our team doesn't wash your lucky jersey? Oh, that smells lucky. Hilton's unmatched hospitality can keep up with any superstition. Even a marching bandwink up call it $555.55. Hit it! When you need a team that will do whatever it takes on game day, it matters where you stay. Hilton, for this day. I get so many headaches every month. It could be chronic migraine, 15 or more headache days a month, each lasting four hours or more. Botox on a botchaline mtoxin A prevents headaches and adults with chronic migraine. It's not for those who have 14 or fewer headache days a month. Prescription Botox is injected by your doctor. Effects of Botox may spread hours to weeks after injection causing serious symptoms. A lurcher doctor right away is difficulty swallowing, speaking, breathing, eye problems or muscle weakness can be signs of a life-threatening condition. Patients with these conditions before injection are at highest risk. Side effects may include allergic reactions, neck and injection site pain, fatigue and headache. Allergic reactions can include rash,

8:14welts, asthma symptoms and dizziness. Don't receive Botox if there's a skin infection. Tell your doctor your medical history, muscle or nerve conditions, including ALS Lou Gehrig's disease, myasthenia-gravis or Lambert Eaton syndrome, and medications, including botchaline mtoxins, as these may increase the risk of serious side effects. Why wait? Ask your doctor. Visit BotoxCronicMyGreen.com or call 1-800-44 Botox to learn more. AI encounters a completely novel variable and gets confused. It might just hallucinate a forecast. And if everyone relies on Google because it is currently the most accurate, you lose the diversity of modeling. You don't have the European or American models running hot in the background with equal funding to catch the error. You just have a bad prediction that entire state government's act on. But predicting where the wind goes is only step one anyway. Predicting where the water goes is step two. And that brings us directly into the modeling of the flood waters. Yeah, so hell Rod Far has been running experimental modeling with a system called AutoCaf to predict the actual flooding for myases.

9:16And to do this accurately, they have to ingest an enormous amount of varied data. They pull Noah, Wind and Rain data, River Discharge Drates, Global Water Levels, then they process all of those inputs through a hydrodynamic solver called Sphinx. Which you probably define what a hydrodynamic solver actually does. Because if you were just drawing a line on a map based on elevation, you are basically guessing where water might pull. Right. But Sphinx, which stands for Superfast Inundation of Coasts, calculates how water behaves when it is forced over complex terrain. It takes the topography of a city, the height of the seawalls, the slope of the streets. And calculates how the water will physically flow. It maps out where it will accelerate, and where it will trap itself in those spaces. They are looking specifically at compound flooding, which is really the most destructive element of a coastal storm. If you are living on the coast, compound flooding is the absolute worst-case scenario. It is not just a wall of ocean water crashing over a beach.

10:19It is more like a bathtub. Exactly like a bathtub. You have the ocean pushing inland, driven by the hurricanes wind. That acts like someone plunging the drain so nothing can escape. At the same time, you have torrential rain falling directly over the city. Which is like pouring a bucket of water from the top. Right. That rainwater collects in the rivers, and the rivers try to flow out to the ocean. That is the faucet running fully open. But the river water has nowhere to go, because the ocean search is acting like a wall. So the water piles up from every direction simultaneously, flooding areas that normally never see water. And to calculate how that water piles up, Radfar's model uses a 200-meter grid layout of the entire impact zone. They basically map the entire coast in 200 meter squares. They run that simulation on a GPU and get a multi-day compound flooding forecast in roughly 80 seconds. Which is an incredible tech flex. Calculating fluid dynamics across a grid of an entire coastline

11:20used to take days of super computer time. Doing it in 80 seconds means you can constantly update the model as the storm shifts. If the eye of the hurricane wobbles 10 miles to the east, you just hit refresh. Yeah, you run the model again, and you know exactly which neighborhoods are going underwater based on the new trajectory. It gives emergency managers a dynamic map instead of a static guess. But there is a severe friction on the ground regarding this data. There is a disclaimer attached to Radfar's work. Right, it is explicitly labeled as a research forecast, not an official warning. You have highly accurate, lightning-fast experimental AI models generating predictive flood maps. But residents are told to wait for official emergency guidance from government channels. Because the technology is moving faster than the bureaucracy's ability to verify and distribute it. If you are sitting in a coastal county, and an experimental AI model says your neighborhood will be under six feet of water in eight hours. But the local emergency management agency has an issue to mandatory evacuation yet. Because their traditional models are still running. What do you do? An 80 second computer simulation gives the physical economy mere hours to react.

12:25You can run a model in a minute, but you cannot move a fleet of semi-trucks out of a flood zone in a minute. No, and Zack Shire looks at this exact reality from the perspective of the logistics industry. While residents are at the hardware store buying plywood to board up their windows, the freight still has to move. Truck drivers are dealing with road closures, wind restrictions, and completely upended delivery schedules across Mississippi, Alabama, and the Florida Panhandle. The disconnect between the AI models and the physical reality of a truck driver is intense. A logistics dispatcher sitting in an office in Chicago might look at a weather map until a driver to push through or out in Alabama. But out on the highway, that driver is trying to keep an 80,000 pounds 70-foot vehicle on the asphalt while hurricane force crosswinds try to flip the trailer over. Right. So a freight carrier representative during a hurricane isn't just sending emails and updating spreadsheets. They have to know the psychology of individual drivers. You have to know which driver is willing to run a dangerous lane, who has the experience to

13:27handle heavy rain, and who will just panic. The rep has to have the authority and the judgment to look at a route and tell a client we shouldn't run this. The risk is too high. You see a stark contrast here with the automation trends dominating the supply chain. Walmart recently discussed hitting peak complexity with their automation efforts. They are building pristine, hyper-efficient robot-driven warehouses where every variable is controlled. The temperature is constant, the floors are flat, the robot move on exact coordinates. But the moment that freight leads the warehouse and enters a disaster zone, all that pristine automation falls apart. It relies entirely on a single human being in a cab, navigating flooded coastal highways and making split-second survival decisions. And the pressure applied to these supply chains often ignores the reality of the weather entirely. Customer deadlines are hard-coded into contracts. A distribution center expects a delivery to specific hour because their inventory model demands it. If they don't get the delivery, the carrier faces financial penalties.

14:30The spreadsheets don't care that a sius just made landfall. Which is exactly where the friction between rigid corporate systems and human safety peaks. You have a moral imperative when dealing with a category 3 hurricane. You have to get that driver home safely. But that moral imperative conflicts directly with the financial imperative of the freight contract. The dispatchers caught in the middle, staring at an AI weather model on one screen, predicting compound flooding, and a demanding client on the other demanding their freight. And none of this logistics network functions without communication. A truck driver cannot navigate and a dispatcher cannot communicate if the cell towers go down. The entire logistics web relies on connectivity, which triggers a massive corporate infrastructure response long before the rain starts. Yeah, the Verizon frontline team is staging mobile assets around the clock. They are working to ensure that first responders have mission-critical connectivity. They are not waiting for towers to fail and the wind to go fix them. They are moving emergency network equipment to the edge of the impact zone before the storm

15:31even makes landfall. Their deployment is entirely proactive and it is based heavily on the AI forecast we discussed earlier. Now they park mobile cell towers, which are essentially heavy-duty trucks equipped with satellite uplinks, massive antennas, and their own generators right on the boundary of the predicted flood map. They position them so they can drive in the minute the windstie down. Now while Verizon is setting up the digital infrastructure, the government is trying to clear the physical space. Alabama, Governor Kay Ivey issued mandatory evacuation orders for Baldwin and Mobile Counties. Those orders specifically target tourists and Gulf shores, orange beach, and Duffen Island before addressing the permanent residents. Evacuating tourists is a completely different logistical problem than evacuating residents. Tourists don't have local networks. They don't know the alternate routes when the interstate floods are the bridge closes. They lack the supplies to shelter in place because they are staying in short-term rentals or hotels with maybe just a weekend's worth of groceries. Clearing them out first is the primary focus for local emergency management because

16:34they are the most vulnerable population in blackout. But inside that evacuation zone, you have people who aren't leaving. The orange beach fire rescue and Spanish fort fire rescue are setting up to work 24 hour shifts directly in the eye of the storm. When the local power grid inevitably fails and the permanent cell towers lose their backup generators, those fire rescue teams are relying entirely on that pre-stage Verizon network to coordinate water rescues and medical emergencies. The government and the telecon companies set the stage, but the actual recovery depends on a completely different network of people who arrive right behind the wind. Steve Powell categorizes this group as the second responders. And the second responders are a vast, complicated mix of organizations. You have insurance property claims professionals who need to assess the damage immediately so rebuilding can start. You have nonprofits like Operation BBQ Relief setting up mobile kitchens. The American Red Cross bringing in cuts and blood supplies. Kage and Navy relief preparing flat bottom boats and the Salvation Army staging logistics hubs.

17:37They all stage just outside the cone of uncertainty, waiting in massive parking lots for the winds to drop below a specific threshold so they can rush into the disaster zone. It is a highly organized response, but it runs parallel to a very personal reality for the people living there. Ashley Ro shared her experience as a first-time Florida resident going through her first hurricane. She described an intense cultural phenomenon where neighbors check on neighbors. It is a very specific mentality that emerges on the coast when a disaster is imminent. People who barely speak to each other during the year are suddenly sharing generators, clearing storm drains together and pooling bottled water. Are your ad campaigns lighting up the dashboard? But not the pipeline. That's bullspin. And marketers are calling it out in... dashboard confessions. Cut the bullspin. See revenue? Not just reach. LinkedIn delivers the highest return-on ad spend of major ad networks.

18:38Advertise on LinkedIn. Spend $250 on your first campaign and get a $25 credit go to LinkedIn.com slash campaign turn sick edition supply. And rituals are reality. Need a room to match your lucky number? We got you. Want to make sure our team doesn't wash your lucky jersey? Oho, that smells lucky. Hilton's unmatched hospitality can keep up with any superstition. Even a marching bandwink up call it 555 and 55 seconds. Hit it! When you need a team that will do whatever it takes on game day, it matters where you stay. Hilton, for this day. I get so many headaches every month. It could be chronic migraine, 15 or more headache days a month, each lasting four hours or more. Botox on a botchaline mtoxin A prevents headaches and adults with chronic migraine. It's not for those who have 14 or fewer headache days a month. Prescription Botox is injected by your doctor. Effects of Botox may spread hours to weeks after injection causing serious symptoms. A lurcher doctor right away as difficulty swallowing, speaking, breathing,

19:40eye problems or muscle weakness can be signs of a life-threatening condition. Patients with these conditions before injection are at highest risk. Side effects may include allergic reactions, neck and injection site pain, fatigue and headache. A lurcher reactions can include rash, welts, asthma symptoms and dizziness. Don't receive Botox if there's a skin infection. Tell your doctor your medical history, muscle or nerve conditions, including ALS Lugeric's disease, myasthenia-gravis or Lambert Eaton syndrome, and medications, including botchaline mtoxins, as these may increase the risk of serious side effects. Why wait? Ask your doctor. Visit BotoxCronicMigraine.com or call 1-800-44 Botox to learn more. You have this organic community-driven resilience happening at the street level, driven entirely by shared survival. But under T provides a look at the exact opposite end of the spectrum, with the private sector disaster economy. This is a completely different world operating in the exact same physical space. He outlines how companies mobilize security, lodging and debris response with a sort of no-contract, no-problem attitude.

20:42These are private security and logistics contractors, looking for immediate work in the disaster zone. On or mentions they are OTA ready. Meaning they use online travel agencies to instantly book massive blocks of hotel rooms just outside the damaged zone to house their mercenary workforce. They monitor the AI weather models, predict where the worst damage will be, and position their teams to secure contracts from wealthy homeowners or corporations who need immediate infrastructure protection. They arrive ready to work for whoever writes the biggest check. The contrast is jerry. You have the cage in navy, which is often just local residents with flat bottom fishing boats, going out to rescue people entirely out of alchewism. And on the other hand, you have private contractors deploying critical infrastructure protection for profit. Both groups are navigating the exact same float waters, relying on the exact same AI weather models. But their motivations and methods are completely opposed. You zoom out from this and you realize the rest of the business world is continuing its normal

21:42operations as if nothing is happening. You have life and death stakes on the Gulf Coast. People are boarding up homes, staging rescue boats, preparing to lose everything. And then you look at the broader tech industry, the built the AI tracking the storm. And they are living in a completely separate reality. Google DeepMind is accurately tracking a deadly storm and providing the data that directs these emergency responses. At the exact same time, Google is aggressively wooing businesses with its Gemini workplace agent, trying to sell AI features for writing emails and summarizing meetings. The juxtaposition is bizarre. Open AI is currently under fire over how they calculate the revenue, and they are dealing with high profile research or exits. Netflix is looming over potential layoffs. A HubSpot founder is taking immense heat for a tone def layoff post. The tech industry is wrapped up in its own insular drama, while its products are literally deciding the fate of coastal cities. You have a technology industry capable of marshalling unbelievable computing power.

22:43They can model compound flooding on a global grid in 80 seconds. They can predict the chaotic fluid dynamics of a hurricane better than any human institution in history. Yet this exact same industry completely struggles with basic human communication, corporate governance, and ethical leadership. It highlights a strange maturity gap in the tech sector. You have an industry that is highly evolved mathematically, but deeply flawed structurally. They can map the future of a deadly storm with pinpoint accuracy, but they can't manage their own internal corporate messaging without causing a scandal. We are trusting the lives of coastal residents to machine learning algorithms built by organizations that are fundamentally chaotic. The anatomy of a modern natural disasters no longer just wind and rain. It is an incredibly complex marriage between machine learning algorithms and heavy duty human logistics.