Sam: The ocean is on track to be over two and a half degrees Celsius warmer than average. Which... is not great.
Kai: I'm Kai.
Sam: And I'm Sam.
Kai: And this is Open Source, Unstarred for Sunday, August 16th, 2026.
Sam: We have three big, and I mean big, stories for you today.
Kai: We're starting with a 'Super El Niño' that new climate models are predicting for this winter, and the forecasts look intense.
Sam: Next up: Your BMI is basically useless for predicting heart disease risk. We'll tell you what the better metric is.
Kai: And we'll finish with a reality check on AI in drug discovery. Is it actually finding cures, or just finding hype?
Kai: Let's start with the weather. Because Sam, this isn't just weather.
Sam: No, it's not. The latest long-range forecasts are pointing to an El Niño that could reach 'Super' status, and it's already breaking records.
Kai: The models are screaming. I was looking at the visualizations from the European Centre for Medium-Range Weather Forecasts, and the data is just... wild.
Sam: And 'wild' here means severe droughts, catastrophic floods, and bizarre temperature swings across the US and Europe.
Kai: But the forecasting technology, Sam! This is a monumental achievement for computational climate science. And so many of these models and their data are open!
Sam: If they're right. Remember, long-range forecasting is still a game of probabilities. The media loves latching onto the worst-case scenario from a single model run.
Kai: But it's not just one model! It's a growing consensus across the major global systems—the ECMWF, the UK Met Office, the American CFS. They're all pointing in the same alarming direction.
Sam: And that direction is toward a warm blob in the Pacific hitting over two and a half degrees Celsius above average.
Kai: Exactly. The number you opened the show with. The impact on global supply chains, agriculture, energy grids... it's massive.
Sam: So what's the takeaway for our listeners, besides maybe buying a backup generator?
Kai: It means the open-source tools we build to model complex systems are working and more important than ever. This is a massive, public dataset—a call to action for every data scientist and viz expert to jump in.
Sam: Or a call to action to check your home insurance. The tech is impressive, but the real-world consequences are what matter. And they look severe.
Sam: Alright, next up, let's talk about a number that has ruled public health for decades: your BMI.
Kai: Which we've all secretly known is kind of a joke, right? It famously classifies Dwayne 'The Rock' Johnson as obese. It's a flawed metric.
Sam: Well, now it's officially flawed. A huge study, presented at the American College of Cardiology's annual conference, found that your waist-to-hip ratio is a far more accurate predictor of heart attacks and strokes.
Kai: This is fantastic! You can measure it with a simple tape measure. It's totally accessible.
Sam: Or... and here comes the tech angle I knew you'd bring up... you can do it with your phone. There are already computer vision apps that claim to do body measurements.
Kai: Of course! Your watch, your health app... instead of that useless BMI, it could give you a genuinely useful risk indicator. That's actionable data for your health.
Sam: Whoa, hold on. 'Actionable data' you're sending to some startup's server? This is a metric now scientifically proven to be a direct line to your heart disease risk. The privacy implications are massive.
Kai: Fair point on privacy. But the upside is people can monitor their own health, instead of waiting for a doctor to deliver bad news.
Sam: Or it can lead to anxiety and obsession. No single number tells the whole story. This is a better predictor, not a crystal ball.
Kai: But the takeaway for developers is that there's a new, validated metric to build on—a chance to create something more meaningful than another BMI calculator.
Sam: Just treat that data with the seriousness it deserves. Because you can bet insurance companies are very interested in this.
Kai: For our final story, let's talk about the holy grail: using AI to discover new drugs.
Sam: A topic absolutely drowning in venture capital and marketing hype.
Kai: But it is happening, Sam! Generative AI is designing novel protein structures. Companies are using this to find new candidates for fighting diseases.
Sam: Are they? Because a fantastic, sobering post in Science by chemist Derek Lowe suggests we all need to take a deep breath.
Kai: He's cautious, I'll give you that. But he doesn't deny the progress! AI is incredible at finding patterns in massive datasets, perfect for screening millions of compounds quickly.
Sam: It's good at step one of a very, very long journey. The hard part isn't finding a molecule that might work. It's proving it works in a human, isn't toxic, and can actually be manufactured. That's development, not discovery.
Kai: But if AI can accelerate that first step by a thousand times, that's a revolution!
Sam: Is it? The article points out a brutal fact: the clinical trial failure rate hasn't changed. It's still over 90 percent. AI is just helping us get to our failures faster.
Kai: That is... a very cynical framing. What about AlphaFold solving protein folding? That's not a failure.
Sam: Incredible science, no doubt. But knowing a protein's 3D structure doesn't magically tell you how to make a drug for it. Lowe calls it the 'lamp-post problem'—we have a brilliant new light, but we're still just looking for our keys right underneath it.
Kai: So the takeaway for our listeners is... the hype is way ahead of the science?
Sam: Dramatically. AI is a powerful tool in the lab, but it hasn't replaced the lab. The messy, difficult, human work of science is still the main event. We are a long way from an AI printing out a cure for Alzheimer's.
Kai: Alright, to recap: A 'Super El Niño' is on the horizon, powered by forecast models that are getting scarily accurate.
Sam: Your BMI is dead. Long live the waist-to-hip ratio, bringing a host of new opportunities and privacy risks for health tech.
Kai: And AI is a great lab assistant for drug discovery, but it's not the lead scientist the headlines want you to believe. At least, not yet.
Sam: Before we go, a quick shoutout to whoever registered `git-oops.com`.
Kai: Wait, what is it? A new tool?
Sam: Nope. It's just a webpage with a single, giant, red button that says, 'I need an adult.' It's not a service, it's a mood. And I have never felt more seen.
Kai: That's our show for Sunday, August 16th! We'll see you tomorrow.
Sam: Until then, maybe go find a tape measure. And for goodness sake, don't push to prod on a Sunday.
This show is made with AI: the hosts’ voices are synthetic and the scripts are AI-assisted. Every story links to its original source.