newsfilter.io
Fireside Chat, Panel, Conference Presentation

a16z Podcast | Self-Driving Cars — Where Are We, Really?

  • Autonomy is projected to emerge through incremental advances occurring annually or biennially rather than as a sudden event, with specific predictions that major OEMs in Europe, the US, and Japan will begin shipping Level 4 vehicles in 2020 and 2021.
  • Small-scale Level 4 operations are expected to appear within the next couple of years, with full consumer availability at dealerships predicted in approximately 10 years, while widespread capability for all is anticipated to arrive significantly sooner than 13 years.
  • Major barriers to large-scale deployment include sensor suites that are currently not economically feasible, unfinalized software infrastructure, and a talent bottleneck of robotic experts and computer science engineers over the next few years.
  • Technical challenges remain significant for achieving Level 4 or 5 across all weather conditions, necessitating teleoperations to handle unknown scenarios, though concerns regarding latency and security vectors for remote takeover persist.
  • Future consumer expectations will include personalization where fleets remember user history and the ability to repurpose commute times, such as the 30 to 40 minutes to Palo Alto, for entertainment or connectivity.
  • Regulatory frameworks must evolve to address daily over-the-air software updates and the certification of intelligent robots without stifling innovation, with a need for government cooperation to prevent a patchwork of standards and potentially build digital infrastructure like HD maps.
  • Societal impacts include a likely drop in crime rates due to vehicles acting as moving sensors, changes in housing prices as parking space becomes available, and a prediction that car ownership will shift from utility to recreational use without disappearing entirely.
  • The industry landscape is expected to see a "bloodbath" concentrated at the tier-one supplier level rather than among manufacturers, with massive opportunities in software applications and startup potential regarding security and data ownership.
  • Security risks are anticipated to mirror those of the web and mobile web, with potential for massive data breaches and concerns over high-precision centimeter-level data, while the industry aims to reach the state of advancement held by forerunners like Waymo.
  • A future scenario with no steering wheel or brakes is predicted in approximately 20 years, though human drivers will remain necessary for rural regions and tough roads due to insufficient data, and governments may need to define boundaries for self-driving technologies.