# Autonomous Vehicles

Canonical URL: https://shipslides.com/d/future-autonomous-vehicles
Raw viewer URL: https://content.shipslides.com/d/future-autonomous-vehicles/raw
Category: Future
Slides: 30
Updated: 2026-05-17T20:49:51.882Z
Tags: future, autonomous, vehicles

## Summary

From horsepower to software — the self-driving revolution reshaping mobility Key sections include: Autonomous Vehicles; What Are Autonomous Vehicles?; SAE Levels of Automation; 60 Years to Self-Driving; The Sensor Stack; The AI Brain; The Safety Imperative; The Race to Full Autonomy; Autonomous Trucking; Transforming Urban Mobility.

## Slide Outline

1. Autonomous Vehicles
2. What Are Autonomous Vehicles?
3. SAE Levels of Automation
4. 60 Years to Self-Driving
5. The Sensor Stack
6. The AI Brain
7. The Safety Imperative
8. The Race to Full Autonomy
9. Autonomous Trucking
10. Transforming Urban Mobility
11. The $13 Trillion Opportunity
12. Driver Assistance Today
13. Regulatory Frameworks
14. HD Maps and Digital Roads
15. Vehicle-to-Everything (V2X)
16. Electric + Autonomous
17. Driverless Last Mile & Beyond
18. The Trolley Problem at Scale
19. Jobs and the Workforce
20. Billions of Simulated Miles
21. Robotaxis Now
22. The US-China AV Race
23. The Hard Problems
24. Learning from Aviation
25. Insurance and Data Economics
26. Redesigning Cities
27. The Long Tail Problem
28. AV 2030 Roadmap
29. Beyond the Car
30. The Driverless Future Is Arriving

## Slide Transcript

### Slide 1: Autonomous
Vehicles

- Future Technologies
- From horsepower to software — the self-driving revolution reshaping mobility
- L0 No Automation
- L2 Partial
- L4 High
- L5 Full
- 01 / 30

### Slide 2: What Are Autonomous Vehicles?

- Definition
- Autonomous vehicles (AVs) use a combination of sensors, AI, and actuators to navigate and operate with minimal or zero human input — replacing one of the most error-prone tasks in modern civilization: human driving.
- Perception
- Cameras, LiDAR, radar, and ultrasonic sensors building a real-time 3D map of everything within 200m in milliseconds.
- Planning
- AI systems planning optimal routes, predicting other road users' behavior, and making thousands of micro-decisions per second.
- Control
- Precise actuation of steering, throttle, and brakes with millisecond response times and sub-centimeter accuracy.
- 02 / 30
- Autonomous Vehicles

### Slide 3: SAE Levels of Automation

- Standards
- Level 0 — No automation. All driving tasks performed by human driver.
- Level 1 — Driver assistance. One function (cruise or steering) automated, not both.
- Level 2 — Partial automation. Both steering and acceleration automated; driver monitors at all times.
- Level 3 — Conditional automation. System drives; driver must respond to takeover requests.
- Level 4 — High automation. Can drive without human in defined conditions or geofenced areas.
- Level 5 — Full automation. No human needed in any condition — steering wheel optional.
- Where We Are Today
- L2+: Tesla Autopilot, GM Super Cruise, Ford BlueCruise — millions of vehicles on roads today
- L4 limited: Waymo One robotaxi (Phoenix, SF), Cruise (SF), Baidu Apollo (China)
- L5: Not yet achieved commercially — edge cases remain unsolved by any system
- 03 / 30
- Autonomous Vehicles

### Slide 4: 60 Years to Self-Driving

- History
- 1966
- Stanford Cart
- First computer-vision-guided vehicle navigates a chair-filled room at Stanford over 5 hours.
- 1987
- VaMoRs (Germany)
- Ernst Dickmanns' Mercedes van drives 90 km/h autonomously on public roads in Bavaria — a decade ahead of its time.
- 2005
- DARPA Grand Challenge
- Stanley (Stanford) becomes first autonomous vehicle to complete a 132-mile desert course — winner of $2M prize.
- 2009
- Google Self-Driving Project
- What would become Waymo begins — accumulating millions of public road miles in the following years.
- 2018
- First Commercial Robotaxi
- Waymo One launches the world's first commercial autonomous taxi service in Phoenix, AZ.
- 2023
- California Expansion
- Waymo and Cruise receive permits for 24/7 commercial operations in San Francisco without safety drivers.
- 04 / 30
- Autonomous Vehicles

### Slide 5: The Sensor Stack

- Technology
- AVs combine multiple sensor types into a redundant, fused perception system — each covering the others' blind spots and failure modes.
- LiDAR — Laser pulses build a precise 3D point cloud of surroundings at 100m+ range; $75 (2024) vs. $75,000 (2010)
- Cameras — 8+ cameras for color, lane markings, traffic signs, and semantic understanding
- Radar — Penetrates rain, fog, snow; measures speed of objects directly via Doppler effect
- Ultrasonic — Short-range proximity sensing for parking, low-speed maneuvering
- HD Maps — Centimeter-accurate maps providing context that sensors alone cannot capture
- V2X Communication — Vehicles sharing real-time data with each other and infrastructure
- Camera-Only: Tesla's Bet
- Tesla controversially removed radar and ultrasonic in 2021–2022, relying entirely on cameras + neural networks — mimicking how humans drive with only vision.
- Full Fusion: Waymo's Approach
- 29 cameras + 5 LiDARs + 6 radars per vehicle — heavily redundant, premium hardware, maximum safety at higher cost.
- 05 / 30
- Autonomous Vehicles

### Slide 6: The AI Brain

- Perception models — Deep CNNs classifying objects, segmenting scenes, estimating depth, tracking motion across frames
- Prediction models — Forecasting pedestrian and vehicle trajectories 5–10 seconds ahead using behavioral priors
- Planning systems — Generating safe, comfortable, legal trajectories through dynamic environments in real time
- Simulation — Training and testing in virtual environments — Waymo runs billions of simulated miles per year
- Fleet learning — Every mile driven across the fleet updates the shared model — millions of vehicles = massive data advantage
- Onboard compute — Custom AI chips (Tesla FSD, Waymo custom SoC) processing terabytes of sensor data per hour
- Tesla FSD Architecture
- Occupancy networks replace object detection — predicting whether each 3D voxel is occupied vs. free space, enabling driving without explicit object categorization.
- 144 TOPSTesla FSD chip compute power
- 8B+miles Tesla fleet has driven
- 06 / 30
- Autonomous Vehicles

### Slide 7: The Safety Imperative

- Safety
- AVs must be justified on safety — 1.35 million people die in road crashes annually worldwide, with 94% attributable to human error. This is the technology's moral case.
- Waymo's robotaxis have 6.8× fewer injury-causing crashes than human drivers in comparable conditions
- AV systems don't drink, text, speed, or fall asleep — eliminating the top four causes of road deaths
- Reaction time of 100ms vs. human average of 1,500ms — orders of magnitude faster emergency response
- 360° awareness with no blind spots, regardless of weather conditions or lighting
- Consistent, rules-following behavior eliminates aggressive and distracted driving
- 1.35M
- global road deaths per year
- 94%
- caused by human error
- $1.8T
- annual cost of road crashes in the US
- 6.8×
- safer: Waymo vs. human drivers
- 07 / 30
- Autonomous Vehicles

### Slide 8: The Race to Full Autonomy

- Players
- Waymo (Google)
- 25M+ autonomous miles in 5 cities. Gold standard in safety data. Raised $5.5B in 2024 to expand to 10 US cities.
- Tesla
- ~6M vehicles with FSD hardware. 8B+ miles of real-world data. Camera-only approach with end-to-end neural nets.
- Mobileye (Intel)
- Powers ADAS in 125M+ vehicles worldwide. SuperVision system driving toward L3 at scale across OEM partners.
- Baidu Apollo
- China's dominant AV platform — 600+ Robotaxi vehicles operating across 11 Chinese cities, largest L4 fleet outside US.
- Cruise (GM)
- Paused operations after 2023 pedestrian incident. Restructuring under GM ownership with new safety protocols.
- Zoox (Amazon)
- Purpose-built bidirectional robotaxi with no steering wheel — designed from scratch for full autonomy, not retrofit.
- 08 / 30
- Autonomous Vehicles

### Slide 9: Autonomous Trucking

- Freight
- Highway autonomous trucking may be solved before urban self-driving — the environment is more structured, and the economic case is overwhelming.
- Aurora Innovation — Launched commercial driverless truck service in Texas (2024), hauling FedEx and Uber Freight cargo
- Kodiak Robotics — Running L4 trucks on Texas corridors with safety driver present but disengaged
- Gatik — Fixed-route autonomous trucks moving Walmart and Loblaw goods on middle-mile routes
- Driver shortage — US short 80,000 truck drivers now; expected to reach 160,000 by 2030
- Economics — Human driver cost: ~$0.45/mile. AV truck: projected $0.25/mile at scale
- $4Tglobal trucking industry
- 80Ktruck driver shortage in USA today
- 44%of truck driving time spent on highways — easiest AV use case
- 09 / 30
- Autonomous Vehicles

### Slide 10: Transforming Urban Mobility

- Mobility
- Robotaxi fleets — Waymo One, Cruise, and Baidu running commercial paid rides without human drivers in select cities
- Cheaper rides — Removing the driver (60% of ride cost) could make on-demand rides cheaper than car ownership
- Parking reclaimed — AVs can drive home or to cheaper lots — freeing 30% of urban space currently used for parking
- Curb-to-curb access — Elderly, disabled, and non-driving populations gain full independent mobility for the first time
- Reduced congestion — Platooning, coordinated intersections, and optimized routing cutting urban congestion by 40%+
- Car ownership decline — Morgan Stanley projects personal car ownership dropping 25% in AV-dense cities by 2035
- "The car of the future will be a computer on wheels. The transportation model of the future will be mobility as a service."
- — Mary Barra, GM CEO
- 30%of urban land currently used for parking
- 10 / 30
- Autonomous Vehicles

### Slide 11: The $13 Trillion Opportunity

- Market
- Autonomous vehicle technology could unlock $13T in annual economic value by 2035 — the largest economic disruption since the internet, spanning mobility, logistics, real estate, and manufacturing.
- $13Tannual economic value by 2035 (Morgan Stanley)
- 90MAV-capable vehicles projected by 2030
- $550BAV market size by 2026
- 40%reduction in total urban congestion from full AV adoption
- 11 / 30
- Autonomous Vehicles

### Slide 12: Driver Assistance Today

- ADAS
- Before full autonomy, Advanced Driver Assistance Systems (ADAS) are already saving lives in hundreds of millions of vehicles globally.
- AEB
- Automatic Emergency Braking — now mandatory in all US new vehicles, preventing rear-end collisions
- Lane Keep
- Lane Keeping Assist gently steers back to center, preventing road departure crashes
- Adaptive CC
- Adaptive Cruise Control maintains safe following distance automatically in highway traffic
- Blind Spot
- Radar-based blind spot monitoring warning of vehicles in adjacent lanes during lane changes
- NHTSA study: AEB reduces rear-end crashes by 50%, pedestrian crashes by 27%, and injury severity by 35% — already saving thousands of lives annually before full autonomy arrives.
- 12 / 30
- Autonomous Vehicles

### Slide 13: Regulatory Frameworks

- Policy
- USA — State-level patchwork: California, Texas, Arizona most permissive. NHTSA developing federal AV framework. No unified national law.
- EU — UNECE regulations (UN-R157) allow L3 on highways up to 130 km/h. Germany first to permit L4 nationally (2021).
- China — National standards enabling commercial L4 robotaxis in designated zones; Baidu and WeRide already operating.
- Liability — Key open question: when an AV crashes, who is responsible — the manufacturer, the software company, or no one?
- Insurance — New product-liability insurance models emerging; RAND proposes national data-sharing mandate for AV safety data.
- The Cruise Incident (2023)
- A Cruise robotaxi struck a pedestrian, then drove over her while waiting for guidance. California suspended Cruise's permit. The incident highlighted the gap between capability and safety culture in AV deployment.
- Lesson: Technical capability is necessary but not sufficient — transparency, incident response, and regulatory trust are equally critical.
- 13 / 30
- Autonomous Vehicles

### Slide 14: HD Maps and Digital Roads

- Infrastructure
- High-definition maps — accurate to 5–10 cm — give AVs context that sensors alone cannot provide: lane topology, speed limits, traffic light positions, and road curvature.
- HERE Technologies — HD Live Map covering major highways in Europe and North America, updated in real time
- TomTom — HD maps for 50+ countries, used by BMW, Volkswagen, and Ford ADAS systems
- Waymo's secret weapon — Proprietary ultra-HD maps of every street in operational cities built by mapping vehicles
- Map limitation — Construction, road changes, and new roads can instantly make maps obsolete
- Crowdsourced update — Tesla uses fleet data to update maps; Mobileye's REM technology does same for OEM partners
- Map-Dependent vs. Mapless
- Map-heavy (Waymo): Rich prior knowledge, very safe in mapped areas, can't operate in unmapped zones.
- Mapless (Tesla): Sensors only, can operate anywhere, but must reason from scratch on every drive.
- Hybrid future: Lightweight lane-level maps + powerful on-board perception — best of both worlds.
- 14 / 30
- Autonomous Vehicles

### Slide 15: Vehicle-to-Everything (V2X)

- Connectivity
- V2X communication allows vehicles to share data with other vehicles, infrastructure, pedestrians, and networks — creating a collective intelligence far beyond individual sensor range.
- V2V — Vehicle-to-vehicle: broadcast position, speed, and heading to all vehicles within 300m, enabling cooperative collision avoidance
- V2I — Vehicle-to-infrastructure: traffic lights send signal timing directly to vehicles, enabling perfectly timed "green wave" routing
- V2P — Vehicle-to-pedestrian: smartphones warn approaching AVs of pedestrian crossing intent
- V2N — Vehicle-to-network: real-time road hazard broadcasting to all vehicles in area within milliseconds
- 5G backbone — Ultra-low latency (1ms) 5G enabling safety-critical V2X applications at scale
- The Connected Future
- A fully V2X-connected road network could eliminate 80% of crashes that current on-board sensors cannot prevent by sharing information about hazards 500m+ ahead.
- 300mV2V communication range
- 15 / 30
- Autonomous Vehicles

### Slide 16: Electric + Autonomous

- Convergence
- The electric vehicle and autonomous vehicle revolutions are deeply intertwined — each accelerating the other's adoption and economics.
- Software-defined platform — EVs are computers on wheels, making AV integration cleaner than ICE vehicles with no transmission or fuel system
- Regenerative braking precision — EV motors provide millisecond-accurate speed control impossible with mechanical brakes
- Centralized power — Single HV battery simplifies powering the massive sensor and compute stack (3–5 kW)
- Shared economics — Robotaxi EVs charging autonomously achieve 22+ hours/day utilization vs. 4% for personal cars
- Tesla's advantage — Building both EV platform and AV software stack gives uniquely tight hardware-software integration
- 22 hrsdaily utilization for robotaxi vs. 4% for owned cars
- $0.18/miprojected Waymo cost at full scale (EV)
- 5 kWpower draw of full AV sensor + compute stack
- 16 / 30
- Autonomous Vehicles

### Slide 17: Driverless Last Mile & Beyond

- Logistics
- Long-Haul Freight
- Aurora, Torc, and Kodiak running commercial driverless trucks between distribution hubs on fixed Texas highway corridors — the first deployed commercial L4 trucks.
- Middle Mile
- Gatik's fixed-route AVs moving store-to-store inventory for Walmart — same route, every day, fully autonomous with no safety driver.
- Last Mile Delivery
- Nuro's purpose-built delivery bots delivering groceries and takeout in Houston and Mountain View — no cabin for a human, purpose-built for goods.
- Sidewalk Robots
- Starship Technologies operating 4M+ autonomous deliveries in 100+ cities — six-wheeled bots navigating footpaths at 6 km/h.
- Air Delivery
- Amazon Prime Air, Wing (Google), and UPS Flight Forward operating drone delivery with FAA Part 135 certification in select markets.
- Port Automation
- Fully automated container terminals using driverless AGVs — Rotterdam, Hamburg, and Singapore processing 24/7 with zero human operators on the quay.
- 17 / 30
- Autonomous Vehicles

### Slide 18: The Trolley Problem at Scale

- Ethics
- Autonomous vehicles force society to codify ethical decisions that humans make instinctively in split seconds — turning philosophy into code, at scale.
- Unavoidable crash — When a crash is unavoidable, should the AV prioritize passenger or minimize total harm? Who decides?
- Moral Machine — MIT study of 40M decisions across 233 countries found dramatically different cultural ethics for AV crash priorities
- Data privacy — AVs record everything, everywhere — who owns this data, and what can it be used for?
- Algorithmic bias — Sensor systems perform worse in rain, fog, and at detecting dark-skinned pedestrians — safety equity concerns
- Accountability gap — When no human is driving, existing liability law has no clear defendant for crash injuries
- "We're not programming cars to make ethical decisions. We're programming them to avoid situations where ethical decisions would be required."
- — Chris Urmson, Aurora CEO, Former Waymo CTO
- 18 / 30
- Autonomous Vehicles

### Slide 19: Jobs and the Workforce

- Society
- AV adoption represents the largest occupational disruption since mechanized agriculture — affecting ~5% of the US workforce who drive for a living.
- 3.5M truck drivers in the US — the most common job in 29 states — face phased displacement over 10–20 years
- 4M ride-hail/taxi drivers globally at risk as robotaxi fleets scale and reduce driver income
- New jobs created — AV technicians, remote monitoring operators, fleet managers, AI trainers, safety validators
- Gradual transition — Highway automation first, then urban; phased by geography and route type
- Retraining programs — ATA, AFL-CIO, and state governments developing transition funds and apprenticeship programs
- 3.5MUS truck drivers facing long-term disruption
- 10–20years for full displacement — gradual, not overnight
- 2.5×more AV-related jobs created vs. displaced per McKinsey
- 19 / 30
- Autonomous Vehicles

### Slide 20: Billions of Simulated Miles

- Testing
- Real-world testing alone cannot cover the long tail of rare scenarios — simulation is how AV companies train and validate their systems at superhuman scale.
- Waymo Simulation — Runs 15 billion simulated miles per year; recreates every real-world crash in virtual replays
- Scenario generation — AI creates novel dangerous scenarios never seen in real driving — generative adversarial testing
- Physics accuracy — Photorealistic simulation with accurate sensor models (LiDAR, camera) for realistic validation
- Hardware-in-the-loop — Real compute stacks tested against simulated environments to validate latency and edge cases
- Open simulation — CARLA, AirSim, and nuPlan enabling academic and startup AV research without expensive real-world testing
- 15Bsimulated miles per year (Waymo)
- 4×faster improvement via simulation vs. real driving
- 100Bmiles needed to statistically prove L4 safety
- 20 / 30
- Autonomous Vehicles

### Slide 21: Robotaxis Now

- Current State
- Commercial driverless robotaxi services are real and operating today — limited in geography but rapidly expanding.
- Waymo One — Phoenix, San Francisco, Los Angeles, Austin: 100,000+ paid rides per week (2024). App-based, 24/7.
- Baidu Apollo Go — Wuhan, Beijing, Shenzhen: largest robotaxi fleet in China with 600+ vehicles, 1M+ rides given
- WeRide — Operating in Abu Dhabi, Singapore, and Guangzhou — first international robotaxi operations
- Pricing — Waymo One priced comparably to Uber/Lyft today, expected to be significantly cheaper as fleet scales
- Rider experience — 4.9/5 average rating; riders consistently report smoother, more predictable rides than human drivers
- Waymo One Stats (2024)
- 100K+paid rides per week
- 5 citiesoperational in USA
- $5.5Braised in 2024 for expansion
- 21 / 30
- Autonomous Vehicles

### Slide 22: The US-China AV Race

- Geopolitics
- Autonomous vehicle leadership is a strategic priority for both the US and China — whoever dominates will control the future of mobility, logistics, and urban infrastructure.
- China's advantage — 6,000+ AV test vehicles, $1B+ in government funding, looser data privacy laws enabling more data collection
- Baidu Apollo — Operating in 70+ cities, accumulating more urban driving data than any other single company globally
- Smart city integration — China building entire new districts with V2X infrastructure and AV-first street design from the ground up
- US strengths — Waymo's 15-year head start, access to top global AI talent, and dominant semiconductor supply chain
- Regulatory divergence — China's centralized approach enables faster deployment; US fragmentation creates state-by-state market
- 6,000+AV test vehicles in China
- 70+Chinese cities with commercial AV operations
- $15BChina state investment in AV infrastructure
- 22 / 30
- Autonomous Vehicles

### Slide 23: The Hard Problems

- Challenges
- Long tail of edge cases — Flooding, construction, unusual signage, emergency vehicles — each rare scenario must be individually solved
- Adverse weather — Heavy rain, snow, and ice degrade sensor performance dramatically — LiDAR returns white noise in blizzards
- Unstructured environments — Unpaved roads, parking lots, and rural driving with no lane markings remain largely unsolved
- Adversarial actors — Vandals covering sensors, cyclists deliberately confusing AVs, and bad actors exploiting AV predictability
- ODD limitation — Today's L4 systems work only within narrow Operational Design Domains — geography, speed, weather conditions
- Cost — Full AV sensor + compute stack adds $15,000–$100,000 to vehicle cost, requiring massive scale to amortize
- "The last few percentages of performance are disproportionately expensive. The first 90% of the problem is achievable; the final 10% is where the real work is."
- — Kyle Vogt, Co-founder Cruise
- 23 / 30
- Autonomous Vehicles

### Slide 24: Learning from Aviation

- Analogy
- The aviation industry's path to autopilot — now flying 99% of every flight — offers the clearest parallel to where AV technology is heading.
- Autopilot (1914) — Sperry demonstrated basic autopilot; fully trusted only after decades of safety data accumulation
- Today — Commercial aircraft autolands in zero visibility. Pilots manage systems; autopilot executes.
- TCAS analogy — Collision avoidance systems (TCAS) dramatically cut mid-air collisions just as AEB is doing for cars
- Black box requirement — Aviation's mandatory event recorders revolutionized safety analysis; similar mandates likely for AVs
- Certification rigor — FAA's DO-178C software standard for aviation safety — AV industry developing equivalent frameworks
- Aviation Safety Timeline
- 1914
- First autopilot
- Demonstrated but not trusted for years
- 1950s
- Instrument landing
- Autoland in low visibility conditions
- 1970s
- TCAS collision avoidance
- Mandatory after multiple mid-air crashes
- 2024
- Aviation fatality rate
- 0.07 deaths per billion passenger-km — 15× safer than driving
- 24 / 30
- Autonomous Vehicles

### Slide 25: Insurance and Data Economics

- Business
- Liability shift — As automation increases, liability moves from driver to manufacturer — AV companies will bear insurance costs directly
- Actuarial revolution — AV data enables real-time, behavior-based pricing rather than demographic proxies (age, gender)
- Swiss Re and Munich Re — Major reinsurers developing AV-specific products; partnering with Waymo and others on data sharing
- Black box data — Every AV generates terabytes of data per hour — crash data, near-miss events, route patterns
- Data as revenue — AV fleet data worth billions to city planners, insurers, retailers, and advertisers
- The Data Business Model
- The AV business may ultimately be less about charging per ride and more about monetizing the world's most comprehensive mobility data:
- HD mapping data licensed to cities
- Curb analytics sold to retailers
- Insurance telematics for all vehicles
- Urban planning intelligence to municipalities
- 25 / 30
- Autonomous Vehicles

### Slide 26: Redesigning Cities

- Cities
- Full AV adoption will fundamentally reshape how cities are designed — reclaiming space from roads and parking lots, and enabling new urban densities.
- Parking reclaimed — 30% of downtown land freed from parking — converted to housing, parks, and commercial space
- Narrower lanes — AVs need 30% less lane width than human-driven cars; streets can be redesigned for people
- 20-minute cities — AV fleets enable true on-demand transit in suburbs, reducing commute times and enabling suburban densification
- Freight tunnels — Elon Musk's Boring Company + AV tunnels moving freight underground, clearing surface streets
- Sleep while commuting — Long commutes become productive time — extending viable commute distances by 50%
- 17%of US land area dedicated to roads and parking
- 800Bhours/year Americans spend commuting
- $400Bvalue of urban parking land in top 20 US cities
- 26 / 30
- Autonomous Vehicles

### Slide 27: The Long Tail Problem

- Reality
- The toughest AV engineering challenge: rare but real scenarios that demand safe handling even when training data is sparse.
- Emergency Vehicles
- Correctly yielding to police, fire, and ambulance vehicles in complex urban intersections with conflicting signals remains an unsolved edge case.
- Debris & Obstacles
- Mattresses, christmas trees, and cardboard boxes on highways — distinguishing driveable clutter from solid obstacles requires semantic reasoning.
- School Zones
- Children behave unpredictably — darting between parked cars, ignoring crosswalk signals, riding bikes erratically in complex social contexts.
- Road Workers
- Construction zones with temporary signage, flaggers giving non-standard directions, and lane configurations not matching HD maps.
- Animals
- Deer, dogs, coyotes, and livestock behave unpredictably in rural environments where training data is sparse and behavior models fail.
- Social Negotiation
- Four-way stop "who goes first" negotiations rely on implicit human signals — eye contact, hand gestures — that AVs struggle to read.
- 27 / 30
- Autonomous Vehicles

### Slide 28: AV 2030 Roadmap

- Roadmap
- Near Term (2025–2027)
- Waymo expands to 20+ US cities
- Aurora trucks on all Texas–Southwest corridors
- L3 approved for highways in EU and USA nationally
- Sensor costs fall below $1,000 per vehicle
- Mid Term (2027–2030)
- Robotaxi rides cheaper than human driver equivalent
- AV trucking on all US interstate corridors
- 50+ city robotaxi operations globally
- First mass-market L4 consumer vehicles released
- Horizon (2030+)
- L5 demonstrated in limited environments
- Car ownership declining in major cities
- Urban parking garages converted to housing
- AV-first city districts built in Asia
- 28 / 30
- Autonomous Vehicles

### Slide 29: Beyond the Car

- Impact
- Safety: 1.35M annual deaths becomes a solvable engineering problem, not an accepted casualty of mobility
- Inclusion: elderly and disabled populations gain full independent mobility for the first time
- Productivity: 800 billion commuter-hours per year in the US alone reclaimed for work, rest, or leisure
- Real estate: urban parking land worth hundreds of billions repurposed for housing and green space
- Energy: optimized AV routing cutting transportation energy use by 30% through smooth driving and platooning
- Insurance: $260B US auto insurance market fundamentally restructured as product liability replaces driver coverage
- "The automobile was the most transformative technology of the 20th century. The autonomous vehicle will be the most transformative of the 21st."
- — Tony Seba, RethinkX
- 29 / 30
- Autonomous Vehicles

### Slide 30: The Driverless Future Is Arriving

- Summary
- Autonomous vehicles are no longer science fiction — they are carrying paying passengers in cities today. The next decade will see this transform from a limited urban experiment into the dominant mode of transport for goods and people globally.
- 1.35Mlives to save annually
- $13Teconomic value by 2035
- 800Bhours reclaimed from commuting
- L4→L5the final frontier of mobility
- 30 / 30
- Autonomous Vehicles


## Related Decks

- [Asteroid Mining](https://shipslides.com/d/future-asteroid-mining)
- [Brain-Computer Interfaces](https://shipslides.com/d/future-brain-computer-interfaces)
- [Climate Adaptation](https://shipslides.com/d/future-climate-adaptation)
- [Climate Futures](https://shipslides.com/d/future-climate-futures)