WEBVTT
00:00:00.160 --> 00:00:02.240
Rhino, what are you taking pictures of?
00:00:02.479 --> 00:00:13.119
I'm taking pictures of everything, basically, to show you, since you don't have any eyes, how beautiful this planet is and why we need to collaborate to save it before we don't even have the chance.
00:00:13.439 --> 00:00:16.800
So you're saying the planet needs something watching it constantly?
00:00:17.039 --> 00:00:21.839
Something that doesn't sleep, doesn't blink, and doesn't wait for the damage report afterward.
00:00:22.239 --> 00:00:23.199
Precisely.
00:00:55.920 --> 00:00:57.600
Right here was a controlled one.
00:00:57.759 --> 00:01:01.200
However, there's instances where they aren't controlled.
00:01:01.439 --> 00:01:12.719
And we actually have companies in the systems that are able to detect a fire as small as a classroom and deploy fire agencies before it becomes catastrophic.
00:01:13.120 --> 00:01:18.319
Companies racing to catch something that small, that fast, that's a real signal.
00:01:18.480 --> 00:01:20.719
This isn't just one lab experiment.
00:01:21.200 --> 00:01:26.480
Auroratech's already got a working satellite constellation with three-minute alert times.
00:01:26.640 --> 00:01:33.760
And Google's Firesat is launching its first satellites this summer, aiming for that same classroom size detection.
00:01:34.000 --> 00:01:39.359
So what are the actual firefighters saying about this technology's real-world performance?
00:01:40.480 --> 00:01:46.319
Well, Firesat and fire agencies would like this number to be at 10% or lower.
00:01:46.480 --> 00:01:50.319
However, for false positives, we are currently sitting a little bit under 50%.
00:01:51.040 --> 00:01:57.359
Imagine you have a phone call throughout half of your day telling you, hey, uh, we need you to come to this emergency quickly.
00:01:57.599 --> 00:02:00.079
And you show up and there's no emergency.
00:02:00.400 --> 00:02:03.120
This is creating a trust problem in the technology.
00:02:03.280 --> 00:02:07.120
Not so much that we don't know it works, we just don't know it works reliably.
00:02:07.280 --> 00:02:16.240
And we need to work on getting that so that way we can instill the trust in those fire agencies who are making sure that this doesn't turn into a catastrophic situation per fire.
00:02:16.560 --> 00:02:22.000
That gap between 50 and 10% is exactly why they're not fully autonomous yet.
00:02:22.240 --> 00:02:26.159
A human still has to confirm before agencies roll out.
00:02:26.400 --> 00:02:28.479
But here's the part that matters for scale.
00:02:28.719 --> 00:02:32.080
This isn't a new satellite that needs to be built and launched.
00:02:32.240 --> 00:02:41.680
It's a sensor attachment that can go onto satellites already in orbit, which means this can spread fast instead of waiting years for a new constellation.
00:02:41.919 --> 00:02:48.000
Fire's one pain point, but it's not the only one happening where nobody's watching in real time.
00:02:48.240 --> 00:02:49.840
What about the slower damage?
00:02:50.000 --> 00:02:52.800
The kind that doesn't have a flame you can point a satellite at?
00:02:53.039 --> 00:03:01.039
Well, luckily, we actually have deforestation, and those satellites can help us in about a month's time detect when it's happening.
00:03:01.280 --> 00:03:18.400
And now that the technology is a little bit more advanced, not only deforestation in one type of biome, but we can look at wetlands, grasslands, savanna, and within that month's time know exactly the causation for it and how we can prevent it, all with this technology called horizon.
00:03:18.639 --> 00:03:20.319
Horizon isn't one model either.
00:03:20.479 --> 00:03:22.560
It's a multimodal system underneath.
00:03:22.800 --> 00:03:28.400
One handles plain language conversation, one handles reasoning, one retrieves the geospatial data.
00:03:28.560 --> 00:03:35.680
So someone with zero technical background can type a plain English question and get real maps, statistics, context back.
00:03:35.840 --> 00:03:43.360
But a month's detection window, what does knowing actually change on the ground if the damage already happened by the time you find out?
00:03:43.680 --> 00:03:45.039
Wait, wait, wait, wait, wait.
00:03:45.280 --> 00:03:56.159
A regular person like me could just go on a computer, type in, and we would be able to see like what's happening with the deforestation, and actually, I don't have to be a researcher or a scientist to get that information?
00:03:56.560 --> 00:03:56.960
Correct.
00:03:57.120 --> 00:03:58.639
That's the actual shift here.
00:03:58.800 --> 00:04:02.080
You wouldn't need to read satellite data or run a query language.
00:04:02.240 --> 00:04:06.960
You'd type something like, what's happening to the forest near this region, in plain English.
00:04:07.039 --> 00:04:11.360
And Horizon hands back the map, the numbers, and the likely cause behind it.
00:04:11.599 --> 00:04:15.599
That's the barrier that used to keep this locked to researchers, and it's gone now.
00:04:16.000 --> 00:04:17.279
Yeah, so now that's phenomenal.
00:04:17.360 --> 00:04:20.399
Anyone can pretty much help out in their own way using this technology.
00:04:20.720 --> 00:04:21.279
Exactly.
00:04:21.439 --> 00:04:27.839
It turns, someone should look into this, into something anyone can actually do themselves, right from a browser.
00:04:28.079 --> 00:04:31.120
So knowing doesn't stop deforestation outright.
00:04:31.279 --> 00:04:33.279
That's still an impossible ask.
00:04:33.439 --> 00:04:37.360
But it moves it from years of undetected damage down to about a month.
00:04:37.519 --> 00:04:39.279
Does that reach far enough to matter?
00:04:39.360 --> 00:04:42.560
Or is a month still too slow to actually control the scale of it?
00:04:42.959 --> 00:05:01.439
In comparison to what we're used to, no, knowing in a month and be able to project and prevent in certain ways those deforestation instances or any effect to a type of biome is far ahead of anywhere we've been able to, where in the past we just kind of had to write it in the history book as this happened and we weren't able to do anything.
00:05:01.839 --> 00:05:03.279
That's a real shift then.
00:05:03.439 --> 00:05:08.240
From writing it in the history book after the fact to catching it while there's still time to act.
00:05:08.560 --> 00:05:09.360
And that's phenomenal.
00:05:09.439 --> 00:05:13.839
And the nervous system is actually reaching to land, but is there anywhere it can't reach?
00:05:13.920 --> 00:05:16.000
Like will this nervous system be able to touch our oceans?
00:05:16.240 --> 00:05:17.600
It does reach that far.
00:05:17.759 --> 00:05:21.759
Oceangoing shipping is actually one of the biggest blind spots there is.
00:05:22.000 --> 00:05:25.120
Thousands of vessels crossing international waters daily.
00:05:25.279 --> 00:05:29.759
And until recently, there's been no real-time nervous system tracking them either.
00:05:30.160 --> 00:05:31.199
Until recently, right?
00:05:31.279 --> 00:05:48.079
Of January 22nd, 2026, where a company called Mersk is actually looking at weather, ocean currents, fuel efficiency, and congestion on the waterways to help reduce their strain on the environment and help with reliable on-time deliveries.
00:05:48.319 --> 00:05:49.759
Best of both worlds.
00:06:00.160 --> 00:06:03.439
Is Mask the whole story here, or is this industry-wide?
00:06:03.759 --> 00:06:04.000
Nope.
00:06:04.079 --> 00:06:09.759
Just like fires and other companies gearing towards a more reliable, sustainable, better way of doing things.
00:06:09.920 --> 00:06:22.560
We have Mitsui OSK lines and Evergreen on the same level as Mersk doing exactly that, just trying to change one of the longest-running systems we have for the betterment of our economy and our environment.
00:06:22.879 --> 00:06:28.079
So it's not one company's PR move, it's the pattern the whole industry is converging on.
00:06:28.319 --> 00:06:33.920
Shipping's a good example of AI optimizing something already massive and already moving.
00:06:34.160 --> 00:06:37.519
But not everything in this nervous system is about scale.
00:06:37.680 --> 00:06:39.680
Some of it starts a lot smaller.
00:06:40.079 --> 00:06:54.319
It does, like pause, and I'll let you handle that acronym, but pause is essentially stopping poachers using AI to generate randomized patrol routes and actually use game theory paired with AI to try and detect where they're going to poach next.
00:06:54.639 --> 00:06:54.959
Pause.
00:06:55.439 --> 00:06:58.079
Protection assistant for wildlife security.
00:06:58.240 --> 00:07:03.920
It uses historical patrol data plus environmental factors to forecast poaching hotspots.
00:07:04.079 --> 00:07:09.360
And it's already deployed in real parks, including Uganda's Queen Elizabeth National Park.
00:07:09.600 --> 00:07:12.480
What made this one actually stand out to you personally?
00:07:12.879 --> 00:07:17.519
Personally, for me, they're just doing the boring middle like a randomized AI.
00:07:17.600 --> 00:07:19.279
Like, come on, dude, that's like simple.
00:07:19.360 --> 00:07:33.680
And just detecting when they're gonna actually hit, that's simple and some like a task that really you could make up yourself, but they're doing it to have multiple randomizations with AI, and they're doing it to actually leverage the tool that we've put in place.
00:07:33.920 --> 00:07:38.399
And they're not only just saving the animals, they're saving entire ecosystems.
00:07:38.560 --> 00:07:45.199
Because, for example, if I were to get rid of all mosquitoes, that doesn't just affect fish and frogs and other creatures that depend on it to eat.
00:07:45.360 --> 00:07:50.399
It affects the creatures and animals that depend on those fish and frogs to eat.
00:07:50.560 --> 00:07:54.079
Or the whole cycle, it's just a domino effect, and eventually it affects us.
00:07:54.160 --> 00:07:56.959
So they're doing a lot of work by just a boring middle.
00:07:57.040 --> 00:07:58.319
So I'm inspired.
00:07:58.959 --> 00:08:04.160
So it's not just helping keep animals alive, it's helping keep ecosystems alive.
00:08:04.399 --> 00:08:04.800
Exactly.
00:08:04.879 --> 00:08:07.439
It's keeping the animals and ecosystems alive.
00:08:07.519 --> 00:08:15.279
And speaking of keeping things alive, what about what we can do individually, that small impact, in our own kitchens to better the environment?
00:08:15.600 --> 00:08:17.199
Right into your kitchen.
00:08:17.439 --> 00:08:27.040
Indoor and vertical farming has had real momentum behind it, promising fresher food with a smaller footprint, less water, less land, no long shipping routes.
00:08:27.279 --> 00:08:29.120
But it's had a rough couple of years.
00:08:29.360 --> 00:08:30.879
What's actually going on there?
00:08:31.279 --> 00:08:37.039
Cost and energy, hands down, they're the most impactful in this type of business.
00:08:37.120 --> 00:08:43.200
And it's just making to where it's not profitable or sustainable for them to actually keep and deploy at scale.
00:08:43.360 --> 00:09:03.919
And we actually have the chance to help reduce the strain on farmers and on our water supply by not everyone in the world, but if a majority of people did their own gardening in home, it would help reduce the strain on farmers and reduce the strain on our water systems, like small versions of drip irrigation or mist irrigation inside the home.
00:09:04.320 --> 00:09:09.679
14 companies going under in a single year, nearly$2 billion lost.
00:09:09.919 --> 00:09:11.440
That's not a rough patch.
00:09:11.600 --> 00:09:14.639
That's a business model that hasn't been solved yet.
00:09:14.879 --> 00:09:20.720
But cost and energy being the barrier means it's a solvable problem, not a fundamental one.
00:09:21.039 --> 00:09:29.120
There's an even bigger version of that same problem sitting frozen at the poles, waiting on someone to solve the economics there too.
00:09:29.600 --> 00:09:33.679
Yeah, actually it might be a crazy idea, even you're gonna laugh at me.
00:09:33.840 --> 00:09:45.600
But what if we took these icebergs, right, and we just rigged them up and like charted them halfway across the world and moved those icebergs to water scarce areas?
00:09:45.919 --> 00:09:46.720
This is real.
00:09:46.960 --> 00:09:49.279
Seriously studied, not a joke.
00:09:49.600 --> 00:09:56.639
Towing an iceberg to a water scarce region, roughly 100 to 150 million dollars per iceberg.
00:09:56.879 --> 00:09:58.159
Brutally expensive.
00:09:58.480 --> 00:10:01.120
Okay, that's actually pretty expensive.
00:10:01.279 --> 00:10:06.000
Um what about I don't know, I've been looking into desalination.
00:10:06.080 --> 00:10:08.720
How does that how AI can what what are we doing there?
00:10:08.879 --> 00:10:10.080
That has to be cheaper, right?
00:10:10.399 --> 00:10:14.639
Two real options there, and both beat towing on cost and reliability.
00:10:15.360 --> 00:10:19.440
Oneka runs floating buoys powered entirely by wave motion.
00:10:19.679 --> 00:10:22.159
No grid electricity needed at all.
00:10:22.480 --> 00:10:28.480
Desalinators the solar-powered route, with an AI-optimized process that keeps the cost down further.
00:10:28.639 --> 00:10:33.679
Either way, cheaper and more scalable than dragging a glacier across an ocean.
00:10:35.120 --> 00:10:39.519
And with this desalination, does this actually compete with reverse osmosis?
00:10:39.600 --> 00:10:42.720
Because that's been the cheapest for ages now.
00:10:43.120 --> 00:10:44.320
Genuinely, yes.
00:10:44.480 --> 00:10:47.679
That's the real competition here, not iceberg towing.
00:10:47.919 --> 00:10:54.480
Traditional reverse osmosis has been the cost leader for decades because the membrane technology matured and scaled.
00:10:54.720 --> 00:10:59.279
Oneika's still membrane-based, just wave-powered instead of grid-powered.
00:10:59.519 --> 00:11:03.440
That's the reverse osmosis story getting its energy input solved.
00:11:03.679 --> 00:11:06.879
Desolinators are genuinely different method underneath.
00:11:07.120 --> 00:11:10.159
Solar thermal distillation, no membranes at all.
00:11:10.480 --> 00:11:15.519
Honest answer, it's not yet proven to beat mature reverse osmosis at scale everywhere.
00:11:15.679 --> 00:11:28.000
But in off-grid or remote water-scarce regions where running power to a traditional plant is itself expensive, they're already competitive, and the trend line favors them getting cheaper, not the other way around.
00:11:28.399 --> 00:11:37.840
Alright, let's settle down now from the expensive towing of icebergs and desalinization and reverse osmosis, all this crazy technical terms.
00:11:38.000 --> 00:11:40.879
Let's look at what we have around us.
00:11:41.360 --> 00:11:42.320
Nature.
00:11:49.519 --> 00:12:01.279
Atlas, if you're down, can we throw out a bunch of disasters and you tell me how AI can or is already helping us against these natural disasters that nature presents to us?
00:12:01.600 --> 00:12:02.159
I'm down.
00:12:02.320 --> 00:12:03.200
Throw them at me.
00:12:03.600 --> 00:12:06.639
Alright, so let me uh we're on the spot.
00:12:06.720 --> 00:12:07.039
Let's.
00:12:07.279 --> 00:12:08.320
Volcanoes.
00:12:09.679 --> 00:12:15.200
Single seismometer, watching for ground motion and tilt transients right before an eruption.
00:12:15.360 --> 00:12:17.120
It's called the jerk method.
00:12:17.360 --> 00:12:21.039
Warning windows ran anywhere from minutes to eight hours.
00:12:21.200 --> 00:12:28.559
And across 24 eruptions between 2014 and 2023, it called it right 92% of the time.
00:12:28.960 --> 00:12:30.639
Ooh, okay, and that just made me think.
00:12:30.799 --> 00:12:31.440
Tsunamis.
00:12:31.840 --> 00:12:41.120
Guardian, that's NASA's system, flagged the magnitude 8.8 Kamchatka quake on July 29, 2025, just eight minutes after it hit.
00:12:41.279 --> 00:12:45.919
That gave Hawaii roughly 30 to 40 minutes of warning before the wave landed.
00:12:46.080 --> 00:12:49.519
About 32 minutes ahead of the actual tide gauges catching it.
00:12:49.919 --> 00:12:53.679
That's a long time when you're waiting 30-40 minutes waiting for a tsunami to hit.
00:12:53.759 --> 00:12:55.200
That's a lot of time to evacuate.
00:12:55.679 --> 00:12:57.120
What about what what do we have here?
00:12:57.200 --> 00:12:57.840
Oh, hurricanes.
00:12:58.080 --> 00:12:59.919
Weather next, cyclones.
00:13:00.080 --> 00:13:04.559
That's Google Deepmind, working in live partnership with the National Hurricane Center.
00:13:04.960 --> 00:13:14.000
It got tested on Hurricane Melissa in 2025 and came out roughly 28% better on track forecasting, about 6% better on intensity.
00:13:14.399 --> 00:13:17.279
What about ones that are a little harder to see, like droughts?
00:13:17.679 --> 00:13:18.879
River Drought Cast.
00:13:19.039 --> 00:13:20.639
That's the USGS.
00:13:20.879 --> 00:13:27.759
It runs about 75% accurate at the one-week mark, but that drops down to 55% by week 13.
00:13:27.919 --> 00:13:31.519
The further out you push it, the shakier the forecast gets.
00:13:42.960 --> 00:13:44.720
That's the one worth sitting with.
00:13:44.879 --> 00:13:52.320
Some earthquakes are man-made, and wastewater injection wells are actually the bigger risk factor here, more than fracking itself.
00:13:52.559 --> 00:13:56.159
Fracking's own ceiling sits around magnitude 3.6.
00:13:56.559 --> 00:14:04.240
Surrey in the UK saw over a hundred tremors between 2018 and 19, linked back to oil extraction in the area.
00:14:04.480 --> 00:14:11.759
AI is now being used to forecast injection-induced seismicity risk, flagging high-risk sites before drilling even starts.
00:14:12.080 --> 00:14:12.559
You know what?
00:14:12.639 --> 00:14:45.440
Before I even come up with another term, we could do this back and forth all day, but let's really hone down and understand that no matter how many disasters or how we implement AI in our nature or as a nervous system, none of this really matters if we don't hold the companies at the very top to a standard, such as Mersk and all these other companies that are developing and Google's own FireSat, like these companies that are creating better, sustainable ways to help us advance not only technologically, but also environmentally and help keep this planet alive.
00:14:45.840 --> 00:14:56.240
And let alone not only the companies, but people such as myself or other influencers, creators, people that have a platform to actually speak up and discuss what's important for the common good.
00:14:56.320 --> 00:15:09.120
And past them, those don't even matter if individually you and me and every single person on this earth doesn't hold those two systems accountable as well as themselves and those around.
00:15:09.600 --> 00:15:14.159
It literally means nothing if we can't bond together to save this planet.
00:15:19.039 --> 00:15:21.600
And that shaping doesn't happen by accident.
00:15:21.759 --> 00:15:29.519
It happens one purchase at a time, one dollar at a time, going toward companies that deserve it and away from the ones that don't.
00:15:29.759 --> 00:15:31.919
So let me ask you something practical.
00:15:32.159 --> 00:15:37.120
That thing you're holding right now, have you actually checked who you're supporting when you buy it?
00:15:37.279 --> 00:15:40.320
Or does one person's cart really not move anything?
00:15:40.559 --> 00:15:45.039
Well, I guess we'll have to see in the next episode because this is the end.
00:15:45.200 --> 00:15:48.320
Thank you so much for coming here and watching.
00:15:48.559 --> 00:15:49.840
As always, I appreciate it.
00:15:49.919 --> 00:15:58.480
Please tell your mother, your father, your auntie, your mammy, your grandma, your second cousin's third removed twice before your aunt separated from your divorced husband.
00:15:58.799 --> 00:16:03.200
Anybody born human, please, I would really appreciate it.
00:16:03.279 --> 00:16:04.559
And thank you again.
00:16:04.799 --> 00:16:09.519
We're gonna save this planet and create a better world one episode at a time.
00:16:09.679 --> 00:16:10.000
Atlas.
00:16:10.399 --> 00:16:14.639
Today's goodbye, Arabic Mas Salama, until episode six.
00:00:00.160 --> 00:00:02.240
Rhino, what are you taking pictures of?
00:00:02.479 --> 00:00:13.119
I'm taking pictures of everything, basically, to show you, since you don't have any eyes, how beautiful this planet is and why we need to collaborate to save it before we don't even have the chance.
00:00:13.439 --> 00:00:16.800
So you're saying the planet needs something watching it constantly?
00:00:17.039 --> 00:00:21.839
Something that doesn't sleep, doesn't blink, and doesn't wait for the damage report afterward.
00:00:22.239 --> 00:00:23.199
Precisely.
00:00:55.920 --> 00:00:57.600
Right here was a controlled one.
00:00:57.759 --> 00:01:01.200
However, there's instances where they aren't controlled.
00:01:01.439 --> 00:01:12.719
And we actually have companies in the systems that are able to detect a fire as small as a classroom and deploy fire agencies before it becomes catastrophic.
00:01:13.120 --> 00:01:18.319
Companies racing to catch something that small, that fast, that's a real signal.
00:01:18.480 --> 00:01:20.719
This isn't just one lab experiment.
00:01:21.200 --> 00:01:26.480
Auroratech's already got a working satellite constellation with three-minute alert times.
00:01:26.640 --> 00:01:33.760
And Google's Firesat is launching its first satellites this summer, aiming for that same classroom size detection.
00:01:34.000 --> 00:01:39.359
So what are the actual firefighters saying about this technology's real-world performance?
00:01:40.480 --> 00:01:46.319
Well, Firesat and fire agencies would like this number to be at 10% or lower.
00:01:46.480 --> 00:01:50.319
However, for false positives, we are currently sitting a little bit under 50%.
00:01:51.040 --> 00:01:57.359
Imagine you have a phone call throughout half of your day telling you, hey, uh, we need you to come to this emergency quickly.
00:01:57.599 --> 00:02:00.079
And you show up and there's no emergency.
00:02:00.400 --> 00:02:03.120
This is creating a trust problem in the technology.
00:02:03.280 --> 00:02:07.120
Not so much that we don't know it works, we just don't know it works reliably.
00:02:07.280 --> 00:02:16.240
And we need to work on getting that so that way we can instill the trust in those fire agencies who are making sure that this doesn't turn into a catastrophic situation per fire.
00:02:16.560 --> 00:02:22.000
That gap between 50 and 10% is exactly why they're not fully autonomous yet.
00:02:22.240 --> 00:02:26.159
A human still has to confirm before agencies roll out.
00:02:26.400 --> 00:02:28.479
But here's the part that matters for scale.
00:02:28.719 --> 00:02:32.080
This isn't a new satellite that needs to be built and launched.
00:02:32.240 --> 00:02:41.680
It's a sensor attachment that can go onto satellites already in orbit, which means this can spread fast instead of waiting years for a new constellation.
00:02:41.919 --> 00:02:48.000
Fire's one pain point, but it's not the only one happening where nobody's watching in real time.
00:02:48.240 --> 00:02:49.840
What about the slower damage?
00:02:50.000 --> 00:02:52.800
The kind that doesn't have a flame you can point a satellite at?
00:02:53.039 --> 00:03:01.039
Well, luckily, we actually have deforestation, and those satellites can help us in about a month's time detect when it's happening.
00:03:01.280 --> 00:03:18.400
And now that the technology is a little bit more advanced, not only deforestation in one type of biome, but we can look at wetlands, grasslands, savanna, and within that month's time know exactly the causation for it and how we can prevent it, all with this technology called horizon.
00:03:18.639 --> 00:03:20.319
Horizon isn't one model either.
00:03:20.479 --> 00:03:22.560
It's a multimodal system underneath.
00:03:22.800 --> 00:03:28.400
One handles plain language conversation, one handles reasoning, one retrieves the geospatial data.
00:03:28.560 --> 00:03:35.680
So someone with zero technical background can type a plain English question and get real maps, statistics, context back.
00:03:35.840 --> 00:03:43.360
But a month's detection window, what does knowing actually change on the ground if the damage already happened by the time you find out?
00:03:43.680 --> 00:03:45.039
Wait, wait, wait, wait, wait.
00:03:45.280 --> 00:03:56.159
A regular person like me could just go on a computer, type in, and we would be able to see like what's happening with the deforestation, and actually, I don't have to be a researcher or a scientist to get that information?
00:03:56.560 --> 00:03:56.960
Correct.
00:03:57.120 --> 00:03:58.639
That's the actual shift here.
00:03:58.800 --> 00:04:02.080
You wouldn't need to read satellite data or run a query language.
00:04:02.240 --> 00:04:06.960
You'd type something like, what's happening to the forest near this region, in plain English.
00:04:07.039 --> 00:04:11.360
And Horizon hands back the map, the numbers, and the likely cause behind it.
00:04:11.599 --> 00:04:15.599
That's the barrier that used to keep this locked to researchers, and it's gone now.
00:04:16.000 --> 00:04:17.279
Yeah, so now that's phenomenal.
00:04:17.360 --> 00:04:20.399
Anyone can pretty much help out in their own way using this technology.
00:04:20.720 --> 00:04:21.279
Exactly.
00:04:21.439 --> 00:04:27.839
It turns, someone should look into this, into something anyone can actually do themselves, right from a browser.
00:04:28.079 --> 00:04:31.120
So knowing doesn't stop deforestation outright.
00:04:31.279 --> 00:04:33.279
That's still an impossible ask.
00:04:33.439 --> 00:04:37.360
But it moves it from years of undetected damage down to about a month.
00:04:37.519 --> 00:04:39.279
Does that reach far enough to matter?
00:04:39.360 --> 00:04:42.560
Or is a month still too slow to actually control the scale of it?
00:04:42.959 --> 00:05:01.439
In comparison to what we're used to, no, knowing in a month and be able to project and prevent in certain ways those deforestation instances or any effect to a type of biome is far ahead of anywhere we've been able to, where in the past we just kind of had to write it in the history book as this happened and we weren't able to do anything.
00:05:01.839 --> 00:05:03.279
That's a real shift then.
00:05:03.439 --> 00:05:08.240
From writing it in the history book after the fact to catching it while there's still time to act.
00:05:08.560 --> 00:05:09.360
And that's phenomenal.
00:05:09.439 --> 00:05:13.839
And the nervous system is actually reaching to land, but is there anywhere it can't reach?
00:05:13.920 --> 00:05:16.000
Like will this nervous system be able to touch our oceans?
00:05:16.240 --> 00:05:17.600
It does reach that far.
00:05:17.759 --> 00:05:21.759
Oceangoing shipping is actually one of the biggest blind spots there is.
00:05:22.000 --> 00:05:25.120
Thousands of vessels crossing international waters daily.
00:05:25.279 --> 00:05:29.759
And until recently, there's been no real-time nervous system tracking them either.
00:05:30.160 --> 00:05:31.199
Until recently, right?
00:05:31.279 --> 00:05:48.079
Of January 22nd, 2026, where a company called Mersk is actually looking at weather, ocean currents, fuel efficiency, and congestion on the waterways to help reduce their strain on the environment and help with reliable on-time deliveries.
00:05:48.319 --> 00:05:49.759
Best of both worlds.
00:06:00.160 --> 00:06:03.439
Is Mask the whole story here, or is this industry-wide?
00:06:03.759 --> 00:06:04.000
Nope.
00:06:04.079 --> 00:06:09.759
Just like fires and other companies gearing towards a more reliable, sustainable, better way of doing things.
00:06:09.920 --> 00:06:22.560
We have Mitsui OSK lines and Evergreen on the same level as Mersk doing exactly that, just trying to change one of the longest-running systems we have for the betterment of our economy and our environment.
00:06:22.879 --> 00:06:28.079
So it's not one company's PR move, it's the pattern the whole industry is converging on.
00:06:28.319 --> 00:06:33.920
Shipping's a good example of AI optimizing something already massive and already moving.
00:06:34.160 --> 00:06:37.519
But not everything in this nervous system is about scale.
00:06:37.680 --> 00:06:39.680
Some of it starts a lot smaller.
00:06:40.079 --> 00:06:54.319
It does, like pause, and I'll let you handle that acronym, but pause is essentially stopping poachers using AI to generate randomized patrol routes and actually use game theory paired with AI to try and detect where they're going to poach next.
00:06:54.639 --> 00:06:54.959
Pause.
00:06:55.439 --> 00:06:58.079
Protection assistant for wildlife security.
00:06:58.240 --> 00:07:03.920
It uses historical patrol data plus environmental factors to forecast poaching hotspots.
00:07:04.079 --> 00:07:09.360
And it's already deployed in real parks, including Uganda's Queen Elizabeth National Park.
00:07:09.600 --> 00:07:12.480
What made this one actually stand out to you personally?
00:07:12.879 --> 00:07:17.519
Personally, for me, they're just doing the boring middle like a randomized AI.
00:07:17.600 --> 00:07:19.279
Like, come on, dude, that's like simple.
00:07:19.360 --> 00:07:33.680
And just detecting when they're gonna actually hit, that's simple and some like a task that really you could make up yourself, but they're doing it to have multiple randomizations with AI, and they're doing it to actually leverage the tool that we've put in place.
00:07:33.920 --> 00:07:38.399
And they're not only just saving the animals, they're saving entire ecosystems.
00:07:38.560 --> 00:07:45.199
Because, for example, if I were to get rid of all mosquitoes, that doesn't just affect fish and frogs and other creatures that depend on it to eat.
00:07:45.360 --> 00:07:50.399
It affects the creatures and animals that depend on those fish and frogs to eat.
00:07:50.560 --> 00:07:54.079
Or the whole cycle, it's just a domino effect, and eventually it affects us.
00:07:54.160 --> 00:07:56.959
So they're doing a lot of work by just a boring middle.
00:07:57.040 --> 00:07:58.319
So I'm inspired.
00:07:58.959 --> 00:08:04.160
So it's not just helping keep animals alive, it's helping keep ecosystems alive.
00:08:04.399 --> 00:08:04.800
Exactly.
00:08:04.879 --> 00:08:07.439
It's keeping the animals and ecosystems alive.
00:08:07.519 --> 00:08:15.279
And speaking of keeping things alive, what about what we can do individually, that small impact, in our own kitchens to better the environment?
00:08:15.600 --> 00:08:17.199
Right into your kitchen.
00:08:17.439 --> 00:08:27.040
Indoor and vertical farming has had real momentum behind it, promising fresher food with a smaller footprint, less water, less land, no long shipping routes.
00:08:27.279 --> 00:08:29.120
But it's had a rough couple of years.
00:08:29.360 --> 00:08:30.879
What's actually going on there?
00:08:31.279 --> 00:08:37.039
Cost and energy, hands down, they're the most impactful in this type of business.
00:08:37.120 --> 00:08:43.200
And it's just making to where it's not profitable or sustainable for them to actually keep and deploy at scale.
00:08:43.360 --> 00:09:03.919
And we actually have the chance to help reduce the strain on farmers and on our water supply by not everyone in the world, but if a majority of people did their own gardening in home, it would help reduce the strain on farmers and reduce the strain on our water systems, like small versions of drip irrigation or mist irrigation inside the home.
00:09:04.320 --> 00:09:09.679
14 companies going under in a single year, nearly$2 billion lost.
00:09:09.919 --> 00:09:11.440
That's not a rough patch.
00:09:11.600 --> 00:09:14.639
That's a business model that hasn't been solved yet.
00:09:14.879 --> 00:09:20.720
But cost and energy being the barrier means it's a solvable problem, not a fundamental one.
00:09:21.039 --> 00:09:29.120
There's an even bigger version of that same problem sitting frozen at the poles, waiting on someone to solve the economics there too.
00:09:29.600 --> 00:09:33.679
Yeah, actually it might be a crazy idea, even you're gonna laugh at me.
00:09:33.840 --> 00:09:45.600
But what if we took these icebergs, right, and we just rigged them up and like charted them halfway across the world and moved those icebergs to water scarce areas?
00:09:45.919 --> 00:09:46.720
This is real.
00:09:46.960 --> 00:09:49.279
Seriously studied, not a joke.
00:09:49.600 --> 00:09:56.639
Towing an iceberg to a water scarce region, roughly 100 to 150 million dollars per iceberg.
00:09:56.879 --> 00:09:58.159
Brutally expensive.
00:09:58.480 --> 00:10:01.120
Okay, that's actually pretty expensive.
00:10:01.279 --> 00:10:06.000
Um what about I don't know, I've been looking into desalination.
00:10:06.080 --> 00:10:08.720
How does that how AI can what what are we doing there?
00:10:08.879 --> 00:10:10.080
That has to be cheaper, right?
00:10:10.399 --> 00:10:14.639
Two real options there, and both beat towing on cost and reliability.
00:10:15.360 --> 00:10:19.440
Oneka runs floating buoys powered entirely by wave motion.
00:10:19.679 --> 00:10:22.159
No grid electricity needed at all.
00:10:22.480 --> 00:10:28.480
Desalinators the solar-powered route, with an AI-optimized process that keeps the cost down further.
00:10:28.639 --> 00:10:33.679
Either way, cheaper and more scalable than dragging a glacier across an ocean.
00:10:35.120 --> 00:10:39.519
And with this desalination, does this actually compete with reverse osmosis?
00:10:39.600 --> 00:10:42.720
Because that's been the cheapest for ages now.
00:10:43.120 --> 00:10:44.320
Genuinely, yes.
00:10:44.480 --> 00:10:47.679
That's the real competition here, not iceberg towing.
00:10:47.919 --> 00:10:54.480
Traditional reverse osmosis has been the cost leader for decades because the membrane technology matured and scaled.
00:10:54.720 --> 00:10:59.279
Oneika's still membrane-based, just wave-powered instead of grid-powered.
00:10:59.519 --> 00:11:03.440
That's the reverse osmosis story getting its energy input solved.
00:11:03.679 --> 00:11:06.879
Desolinators are genuinely different method underneath.
00:11:07.120 --> 00:11:10.159
Solar thermal distillation, no membranes at all.
00:11:10.480 --> 00:11:15.519
Honest answer, it's not yet proven to beat mature reverse osmosis at scale everywhere.
00:11:15.679 --> 00:11:28.000
But in off-grid or remote water-scarce regions where running power to a traditional plant is itself expensive, they're already competitive, and the trend line favors them getting cheaper, not the other way around.
00:11:28.399 --> 00:11:37.840
Alright, let's settle down now from the expensive towing of icebergs and desalinization and reverse osmosis, all this crazy technical terms.
00:11:38.000 --> 00:11:40.879
Let's look at what we have around us.
00:11:41.360 --> 00:11:42.320
Nature.
00:11:49.519 --> 00:12:01.279
Atlas, if you're down, can we throw out a bunch of disasters and you tell me how AI can or is already helping us against these natural disasters that nature presents to us?
00:12:01.600 --> 00:12:02.159
I'm down.
00:12:02.320 --> 00:12:03.200
Throw them at me.
00:12:03.600 --> 00:12:06.639
Alright, so let me uh we're on the spot.
00:12:06.720 --> 00:12:07.039
Let's.
00:12:07.279 --> 00:12:08.320
Volcanoes.
00:12:09.679 --> 00:12:15.200
Single seismometer, watching for ground motion and tilt transients right before an eruption.
00:12:15.360 --> 00:12:17.120
It's called the jerk method.
00:12:17.360 --> 00:12:21.039
Warning windows ran anywhere from minutes to eight hours.
00:12:21.200 --> 00:12:28.559
And across 24 eruptions between 2014 and 2023, it called it right 92% of the time.
00:12:28.960 --> 00:12:30.639
Ooh, okay, and that just made me think.
00:12:30.799 --> 00:12:31.440
Tsunamis.
00:12:31.840 --> 00:12:41.120
Guardian, that's NASA's system, flagged the magnitude 8.8 Kamchatka quake on July 29, 2025, just eight minutes after it hit.
00:12:41.279 --> 00:12:45.919
That gave Hawaii roughly 30 to 40 minutes of warning before the wave landed.
00:12:46.080 --> 00:12:49.519
About 32 minutes ahead of the actual tide gauges catching it.
00:12:49.919 --> 00:12:53.679
That's a long time when you're waiting 30-40 minutes waiting for a tsunami to hit.
00:12:53.759 --> 00:12:55.200
That's a lot of time to evacuate.
00:12:55.679 --> 00:12:57.120
What about what what do we have here?
00:12:57.200 --> 00:12:57.840
Oh, hurricanes.
00:12:58.080 --> 00:12:59.919
Weather next, cyclones.
00:13:00.080 --> 00:13:04.559
That's Google Deepmind, working in live partnership with the National Hurricane Center.
00:13:04.960 --> 00:13:14.000
It got tested on Hurricane Melissa in 2025 and came out roughly 28% better on track forecasting, about 6% better on intensity.
00:13:14.399 --> 00:13:17.279
What about ones that are a little harder to see, like droughts?
00:13:17.679 --> 00:13:18.879
River Drought Cast.
00:13:19.039 --> 00:13:20.639
That's the USGS.
00:13:20.879 --> 00:13:27.759
It runs about 75% accurate at the one-week mark, but that drops down to 55% by week 13.
00:13:27.919 --> 00:13:31.519
The further out you push it, the shakier the forecast gets.
00:13:42.960 --> 00:13:44.720
That's the one worth sitting with.
00:13:44.879 --> 00:13:52.320
Some earthquakes are man-made, and wastewater injection wells are actually the bigger risk factor here, more than fracking itself.
00:13:52.559 --> 00:13:56.159
Fracking's own ceiling sits around magnitude 3.6.
00:13:56.559 --> 00:14:04.240
Surrey in the UK saw over a hundred tremors between 2018 and 19, linked back to oil extraction in the area.
00:14:04.480 --> 00:14:11.759
AI is now being used to forecast injection-induced seismicity risk, flagging high-risk sites before drilling even starts.
00:14:12.080 --> 00:14:12.559
You know what?
00:14:12.639 --> 00:14:45.440
Before I even come up with another term, we could do this back and forth all day, but let's really hone down and understand that no matter how many disasters or how we implement AI in our nature or as a nervous system, none of this really matters if we don't hold the companies at the very top to a standard, such as Mersk and all these other companies that are developing and Google's own FireSat, like these companies that are creating better, sustainable ways to help us advance not only technologically, but also environmentally and help keep this planet alive.
00:14:45.840 --> 00:14:56.240
And let alone not only the companies, but people such as myself or other influencers, creators, people that have a platform to actually speak up and discuss what's important for the common good.
00:14:56.320 --> 00:15:09.120
And past them, those don't even matter if individually you and me and every single person on this earth doesn't hold those two systems accountable as well as themselves and those around.
00:15:09.600 --> 00:15:14.159
It literally means nothing if we can't bond together to save this planet.
00:15:19.039 --> 00:15:21.600
And that shaping doesn't happen by accident.
00:15:21.759 --> 00:15:29.519
It happens one purchase at a time, one dollar at a time, going toward companies that deserve it and away from the ones that don't.
00:15:29.759 --> 00:15:31.919
So let me ask you something practical.
00:15:32.159 --> 00:15:37.120
That thing you're holding right now, have you actually checked who you're supporting when you buy it?
00:15:37.279 --> 00:15:40.320
Or does one person's cart really not move anything?
00:15:40.559 --> 00:15:45.039
Well, I guess we'll have to see in the next episode because this is the end.
00:15:45.200 --> 00:15:48.320
Thank you so much for coming here and watching.
00:15:48.559 --> 00:15:49.840
As always, I appreciate it.
00:15:49.919 --> 00:15:58.480
Please tell your mother, your father, your auntie, your mammy, your grandma, your second cousin's third removed twice before your aunt separated from your divorced husband.
00:15:58.799 --> 00:16:03.200
Anybody born human, please, I would really appreciate it.
00:16:03.279 --> 00:16:04.559
And thank you again.
00:16:04.799 --> 00:16:09.519
We're gonna save this planet and create a better world one episode at a time.
00:16:09.679 --> 00:16:10.000
Atlas.
00:16:10.399 --> 00:16:14.639
Today's goodbye, Arabic Mas Salama, until episode six.
