All Segments
a16z Podcast | Exploding the Map
33 segments available
In this episode, Wei Luo, founding COO of DeepMap -- who build HD maps for autonomous vehicles -- and David Rumsey, founder of the David Rumsey Map Collection (one of the largest paper private map collections in the world, now at Stanford University, and *the* largest digital online private collection in the world, at 80,000 + maps) talk with a16z's Hanne Tidnam about how maps -- and mapmaking tools -- are changing in the age of autonomous vehicles. New ways of mapping the world have always led to profound changes. In the Renaissance -- another golden age of mapmaking -- mapmakers used tools such as sextants to measure distance to the stars and compasses to navigate the world around them. Cartography is undergoing yet another major paradigm shift as it now evolves into HD mapping. So what kinds of data and information do maps now need to contain in order to allow cars (and other autonomous robots of all kinds) to navigate the world around them, down to only a few centimeters of accuracy? How will the nature of maps fundamentally change when they are made by self-driving cars, for self-driving cars, in the era of HD mapping?
Segments Timeline
The Evolution of Cartography
In this segment, Hanne Tidnam introduces the podcast and discusses the evolution of cartography with Wei Luo and David Rumsey. They explore how mapping tools have transitioned from sextants to modern technologies like computer vision and lidar, and question whether the nature of mapping has fundamentally changed or simply evolved.
"hi and welcome to the a 16z podcast I'm Hannah and we're here today talking about the evolution of cartography how map making is fundamentally changing in the age of autonomous vehicles with whale oh ..."
Mapping's Golden Era
David Rumsey reflects on the golden era of mapping during Columbus's time, highlighting the desire to map newly discovered lands. He discusses the significance of historical globes, such as Martin Behaim's globe from 1492, which lacked North and South America, illustrating how maps encapsulate the knowledge of their time.
"referred to sextants as primitive the lidar of the day sextants opened up the whole world and created the ability to make scale maps and to map essentially beyond where you were when would you say the..."
The Backbone of Modern Mapping
The conversation shifts to the current state of mapping, with Wei Luo explaining the technological advancements that have transformed how maps are created. He emphasizes the role of sensors, satellites, and data collection in producing digital maps, while noting that the process of map creation is still largely traditional.
"really powerful well it's interesting because it's capturing an entire time to write it's a representation of our all our knowledge at the time well that's the history and that globe has information t..."
Dynamic vs. Static Maps
Wei Luo contrasts the static nature of paper maps with the dynamic capabilities of digital maps. He discusses how digital maps can change in real-time, adapting to new information, and how this evolution represents a significant shift in the mapping landscape.
"they need to predict into areas that they cannot visually see and today you know back to your question you know how what are the backbones of today's digital maps well first of all we have a lot of se..."
The Democratization of Maps
The discussion continues with David Rumsey explaining how access to maps has changed over time, from exclusive ownership by the wealthy to widespread availability. They explore how the depiction of space remains consistent, even as the methods of map production have evolved.
"are maps that you have on your phone there are maps that are probably in your car navigation system as well those maps were not created from the GPS unit that you have on your car or the camera system..."
The Need for Precision in Mapping
Wei Luo emphasizes the importance of precision in modern mapping, particularly for autonomous vehicles. He discusses the need for centimeter-level accuracy in HD maps, which are crucial for safe navigation in complex environments.
"maps I do collect the tools of map making oh yeah and for mapping the West there would be you know typically officers in the Army 1870 with a we have a big wooden board that they would keep on their a..."
The Future of Autonomous Mapping
The conversation shifts to the future of mapping for autonomous vehicles. Wei Luo explains the concept of high-definition maps designed specifically for robotics, highlighting the need for these maps to provide precise information for safe navigation.
"would you say know one thing that has changed drastically from paper maps is that maps are dynamic now mm-hmm maps are changing you know with paper maps you have a sense of scale the scale is printed ..."
Challenges in Autonomous Navigation
The segment discusses the challenges faced by autonomous vehicles in interpreting their environment. Wei Luo points out that while robots can outperform humans in certain tasks, they struggle with complex decision-making that humans handle intuitively.
"eventually digital distribution over the web so maps have become ubiquitous so the access to the max the access has been the most profound change the actual depiction of space I think are the goal to ..."
High-Definition Mapping Explained
Wei Luo elaborates on high-definition maps, which require extreme precision for autonomous vehicles. He explains the critical information these maps must contain, such as lane boundaries and curb heights, to ensure safe navigation in real-world scenarios.
"looking at say Google Maps or you know Apple maps ways and and so on these are maps purposefully built to be first of all very easy to interpret by human beings and easy to use for navigation purpose ..."
The One-to-One Mapping Concept
The discussion culminates in the concept of one-to-one mapping, where maps represent the real world with high fidelity. Wei Luo reflects on the philosophical implications of HD mapping, likening it to a story about a cartographer who created an impossibly detailed map.
"mapping to serve that we need maps that are easy to interpret by robots deep map is focusing on self-driving cars but you can't generalize it to any robot that needs to roam around the physical world ..."
High-Definition Maps: A New Standard
The conversation shifts to the specifics of high-definition (HD) maps, which are essential for self-driving vehicles. These maps require extreme precision, detailing not just lanes and road boundaries but also curbs and other critical features. The speakers discuss the importance of HD mapping in ensuring safety, especially in high-risk areas like roads with cliffs.
"vehicles the maps that are purposively field for self-driving purpose are usually called high-definition maps or height or HD map for short they specifically refer to the maps that have extra that hav..."
The Holy Grail of Mapping: One-to-One Accuracy
This segment explores the concept of one-to-one mapping, where the map reflects the real world with unparalleled accuracy. The speakers reference a fictional story about a cartographer who created a one-to-one map, drawing parallels to the aspirations of HD mapping in the context of autonomous vehicles. They discuss how this level of detail transforms the way maps are perceived and utilized.
"there absolutely I mean people tend to think why do I need to say 5 centimeter or 10 centimeter accuracy when I'm driving down the road in most cases you know the tolerance for error might be higher t..."
From Simplification to Comprehensive Mapping
The speakers delve into the evolution of maps from simple representations to comprehensive tools that capture extensive details. They discuss how modern maps must include a multitude of elements, such as speed limits and lane restrictions, to provide autonomous vehicles with the necessary information for safe navigation.
"information that again humans may take for granted so not only we need to know where the lanes are where the road boundaries are we also want to know where the curbs are how high the curbs are if it's..."
The Historical Context of Mapping
In this segment, the discussion touches on the history of mapping, referencing the first atlas published in 1570 and its significance in capturing global information. The speakers highlight the importance of sharing mapping information and how historical practices have shaped modern mapping techniques, emphasizing the value of open content in advancing accuracy.
"and like taking one element of that information and showing and simplifying one INT down into one picture of like you know borders or topography or land versus sea but now you're talking about actuall..."
The New Frontier of Mapping Technology
The conversation shifts to the technological advancements in mapping, particularly in the 21st century. The speakers discuss how self-driving cars share information with each other, enhancing the accuracy of maps. They reflect on the historical reluctance to share mapping information and how modern practices favor collaboration for better mapping outcomes.
"because it it isn't just about scale it's also about comprehensiveness so the first atlas was published in 1570 by abraham ortelius and it was actually the best-selling book in Europe years and years ..."
Sensors and Data Collection in Autonomous Vehicles
This segment focuses on the various sensors used in self-driving cars, such as cameras, lidar, and GPS, and their roles in map creation and updates. The speakers discuss the importance of utilizing multiple sensors to ensure safety and accuracy, highlighting the need for real-time data processing in autonomous navigation.
"areal map now you are you're really just exploding what the map is the map becomes reality well it's also an immersive map so what kind of tools allow you to do that how are we actually starting to ex..."
Triangulation and Precision in Mapping
The discussion continues with a focus on the methods of triangulation and the importance of precision in mapping. The speakers compare historical mapping techniques with modern approaches, emphasizing how advancements in technology allow for instant and accurate measurements, which are crucial for the functionality of autonomous vehicles.
"which is a unit that tracks the movement of the car and the radars as well are you using everything you possibly can or have you made strategic choices about what the best combination and why those to..."
Creating and Consuming Maps Simultaneously
In this segment, the speakers explore the concept of self-driving cars creating and consuming maps in real-time. They discuss the challenges of maintaining up-to-date maps and the necessity of aggregating data from multiple vehicles to enhance mapping accuracy and reliability.
"precision measurement cameras oftentimes it's just not enough and lasers gives you a precise measurement of depth or distance in the 3d space if we're talking about sensors that are mounted on the sel..."
The Challenge of Data Management
The conversation shifts to the challenges of managing vast amounts of data generated by autonomous vehicles. The speakers discuss strategies for categorizing information to optimize real-time sharing and the importance of balancing data transmission with efficiency to ensure safe driving.
"driving at the high speed we suppose the cameras the lidar the radars and all the other sensors do you have a little like map making bubble basically that you're traveling around the world in also rea..."
The Value of Open Information in Mapping
This segment emphasizes the historical significance of sharing mapping information and its impact on accuracy and business models. The speakers reflect on how open content has proven to be more valuable in the long term compared to proprietary information, drawing lessons from historical mapping practices.
"history of mapping has been really important in moving accuracy ahead and not every state in the world believed in sharing information for instance at one time in the period of explorations the Spanis..."
The Future of Mapping: Aggregated Data Insights
The final segment discusses the future of mapping in the context of autonomous vehicles, focusing on the need for aggregated data from multiple sources. The speakers highlight the importance of collaboration and data sharing in creating accurate and reliable maps for safe navigation.
"value as sharing so I think that's one very important themes in terms of the history of Car Talk I love that so we talked a little bit about the sensors are there ways that you gather the information ..."
Mapping the World: Collective Data Gathering
Luo explains the necessity of multiple vehicles contributing to map creation. He argues that a single car cannot map the entire world effectively; instead, an aggregated approach using data from many cars enhances the quality and accuracy of HD maps, reflecting real-time changes on the roads.
"mean is it just one car driving through one time do it is enough one soldier on a horse I don't think one soldier on the horse is sufficient a single car cannot map the entire world the reason is that..."
Challenges of Real-Time Data Processing
The conversation shifts to the challenges of processing vast amounts of data collected by autonomous vehicles. Luo highlights the need to categorize information for real-time sharing, balancing the urgency of critical updates, like accidents, against less immediate data, such as environmental changes.
"on the road or the more data that we collect from the higher quality the map will be right a constellation a constellation yes many bubbles see you're dealing with enormous amounts of data that you ha..."
The Living Map: Dynamic Updates
Luo describes how HD maps are evolving into dynamic entities that continuously update based on real-world changes. He discusses the role of cloud infrastructure and edge computing in maintaining map accuracy, allowing vehicles to make independent decisions based on the most current data.
"know accident on a road right or maybe you know you know ranked last rainy season they're adding a barrier there were a lot of landslides not trees falling down on the road how about faster things lik..."
Components of Self-Driving Software
The segment outlines the four key components of self-driving software: perception, localization, planning, and mapping. Luo explains how these components interact to ensure vehicles can navigate safely and effectively, emphasizing the importance of accurate mapping in this ecosystem.
"then you distribute that information more broadly to all the cars it's so interesting that it's sort of exploding the time dimension of map to you know it's like becoming a living organism it's a very..."
Decision-Making in Autonomous Vehicles
Luo elaborates on the decision-making process within self-driving cars, detailing how they detect changes in their environment and determine whether to share this information in real-time. He emphasizes the separation of mapping functions from decision-making to enhance operational efficiency.
"term people tend to use is called edge computing essentially each car will carry some storage and some computing power so that it can make its own decision independently right if you know it's complet..."
The Evolution of Mapping: A New Paradigm
In this segment, Luo reflects on the revolutionary changes in mapping technology, likening it to a living organism. He discusses how HD maps not only represent current conditions but also maintain a historical context, allowing for a deeper understanding of changes over time.
"the decision-making is separate from the map it's completely two separate functions so on a very high level the self-driving sulfurous that consists of four components the first component is called th..."
Visualizing Data for Humans and Robots
The discussion shifts to the visualization of mapping data for both humans and autonomous systems. Luo explains the need for human moderators in the map creation process and how aesthetic choices in map design can enhance usability while prioritizing functionality.
"anything like it and of course I want to collect it well let's what is it like what does it look like to the car does the car see something does a passenger see anything has David seen what this map l..."
Legal Implications of Mapping
Luo and Rumsey delve into the legal complexities surrounding maps, discussing how maps have historically held power and influence. They explore the challenges of ensuring accuracy and authenticity in mapping, especially in a digital age where misinformation can easily spread.
"ways there are so many aesthetic choices that have go into map making which I'm sure part of why you love maps so much it's an art so what kind of aesthetic choices did you guys make or how are you yo..."
The Future of Data Ownership
The segment delves into the question of data ownership in the context of HD mapping. Luo and Rumsey discuss who should own the data generated by autonomous vehicles and the implications for privacy and governance. They reflect on historical perspectives of map ownership and the evolving nature of data rights in the digital age.
"atlases on the ship it's gonna be a big challenge for HD mapping because as of today even government regulations are not clear even though digital maps are widely available today but there are countri..."
Capturing History Through Maps
Rumsey emphasizes the historical value of maps as they capture moments in time. He draws parallels between traditional atlases and modern HD maps, suggesting that the latter can provide a dynamic representation of our world. The conversation highlights the importance of preserving data for future generations and understanding historical contexts.
"regulation side what the government is willing to do for instance who should own the data as the driver of the car or as the private homeowner do I own the data around my private Hortonville hard make..."
Empowering Citizens as Mapmakers
The discussion explores the democratization of mapping through technology. Luo suggests that as more individuals contribute to mapping efforts via autonomous vehicles, they become active participants in knowledge creation. This shift could empower citizens and historians alike, providing unprecedented access to historical data and patterns.
"before we had our conversation I to be honest I didn't because you know when it comes to the value is to have it in the present just always about freshest it should be that's exactly what you should b..."
The Paradigm Shift in Mapping
Luo and Rumsey conclude by discussing the transformative potential of HD mapping. They envision a future where mapping is not only more accurate but also more accessible, creating a living map that reflects real-time changes in the environment. This segment encapsulates the idea that everyone can become a mapper, fundamentally altering our relationship with geography.
"we're creating a digital infrastructure of a cities road network and they do try to create a digital infrastructure today because almost every city in the US has a GIS department they send out a surve..."