Aug 19, 2026 · 21 min · 10 segments
In this episode of AI10X, we explore how AI is revolutionizing archaeology by uncovering hidden sites and revealing new insights into ancient civilizations. In this episode, we cover: ✔️ AI-driven…
Yeah.
And you know, the core premise of this document, it really challenges our fundamental perception of machine learning.
Yeah.
Because we are so conditioned, right, to view artificial intelligence as this, um, this predictive engine for the future.
Right.
Like forecasting the stock market or, you know, driving autonomous cars.
Exactly.
Generating tomorrow's code, all of that.
But digging into data presents this completely inverted use case, which I just find incredible.
It details how researchers are deploying these really complex algorithms as, uh, a retroactive lens.
A retroactive lens.
I like that.
Yeah.
They are essentially building a computational time machine to reconstruct the ancient past rather than predict the future.
I am so glad you are joining us for this deep dive because I feel like that inversion is exactly what makes this topic so captivating for anyone listening.
To understand how an algorithm can locate a lost civilization, you know, one that has completely evaded human detection for thousands of years, we really have to look at the exact data they are feeding into these models.
Right.
The inputs are everything here.
And the source material outlines this highly specialized, basically a three-part toolkit, right? We've got historical records, satellite imagery, and ground-penetrating radar data.
Which is fascinating because those three data streams represent entirely different modalities of information, where you have textual data, you have geospatial visual data, and then you have this subsurface spectral data.
It's a lot to process.
It is.
And historically, you know, synthesizing those three streams, that required these completely disjointed teams.
You'd have linguists, geologists, field archaeologists, and they're all working in silos over the course of decades, really.
Yeah.
And you know, the core premise of this document, it really challenges our fundamental perception of machine learning.
Yeah.
Because we are so conditioned, right, to view artificial intelligence as this, um, this predictive engine for the future.
Right.
Like forecasting the stock market or, you know, driving autonomous cars.
Exactly.
Generating tomorrow's code, all of that.
But digging into data presents this completely inverted use case, which I just find incredible.
It details how researchers are deploying these really complex algorithms as, uh, a retroactive lens.
A retroactive lens.
I like that.
Yeah.
They are essentially building a computational time machine to reconstruct the ancient past rather than predict the future.
I am so glad you are joining us for this deep dive because I feel like that inversion is exactly what makes this topic so captivating for anyone listening.
To understand how an algorithm can locate a lost civilization, you know, one that has completely evaded human detection for thousands of years, we really have to look at the exact data they are feeding into these models.
Right.
The inputs are everything here.
And the source material outlines this highly specialized, basically a three-part toolkit, right? We've got historical records, satellite imagery, and ground-penetrating radar data.
Which is fascinating because those three data streams represent entirely different modalities of information, where you have textual data, you have geospatial visual data, and then you have this subsurface spectral data.
It's a lot to process.
It is.
And historically, you know, synthesizing those three streams, that required these completely disjointed teams.
You'd have linguists, geologists, field archaeologists, and they're all working in silos over the course of decades, really.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.