Google AI takes over hurricane forecasting

SPEAKER_1 (0:00)

New and exclusive holiday decor just launched online at the Home Depot. Check out the viral Grand Duchess collection with smart holiday lights you can control from your phone. Easily customize the twinkle across trees, wreaths and garlands to match your holiday vibe.

Create stunning light effects, then switch it up with the tap of your fingers. Even the man in red will be impressed. Shop the holiday decor you want online only at homedepot.com.

SPEAKER_2 (0:30)

This episode is brought to you by Nordstrom.

For 125 years, Nordstrom has been part of your story. The stylist who found the perfect outfit for your big interview. The first pair of walking shoes for your one-year-old. The tomato soup you shared with your sister at the cafe. These are the moments that make Nordstrom, Nordstrom. Big or small, we're here for all of it. Celebrating 125 years with new moments still add. Discover it all in Nordstrom stores and at nordstrom.com.

SPEAKER_3 (1:01)

When you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a $75 sponsored job credit at indeed.com/podcast. That's indeed.com/podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored Jobs.

SPEAKER_4 (1:30)

The National Hurricane Center just handed the tracking of a category three Atlantic hurricane named Asias entirely over to an artificial intelligence model built by Google DeepMind.

SPEAKER_3 (1:41)

Yeah.

SPEAKER_5 (1:41)

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 states 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.

SPEAKER_4 (2:08)

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.

SPEAKER_5 (2:24)

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, what happens to the physical world when the emergency response is dictated by a tech company's algorithm?

SPEAKER_4 (2:42)

Right. And Andrew Friedman and Michael Fommel 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.

SPEAKER_5 (2:53)

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.

SPEAKER_4 (3:07)

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.

SPEAKER_5 (3:19)

Completely dead. Google's WeatherNext 3 is 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.

SPEAKER_4 (3:32)

The reason the traditional models lost really comes down to how they process information.

A traditional physics-based model relies on raw math. Meteorologists take the current temperature, barometric pressure, wind speed, and they plug those numbers into incredibly dense equations governing fluid dynamics.

SPEAKER_5 (3:50)

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.

SPEAKER_4 (4:02)

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.

SPEAKER_5 (4:15)

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.

SPEAKER_4 (4:29)

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.