Andrej Karpathy, former Tesla Autopilot Director, offers a fascinating perspective on Artificial General Intelligence (AGI) by drawing parallels with the evolution of self-driving technology. As the world grapples with the implications of AGI, Karpathy suggests that the trajectory of autonomous vehicles provides valuable insights into what AGI might entail and its impact on society.
The blog post has since been deleted, but you can still check it out here; Karpathy defines AGI as an autonomous system exceeding human capabilities in most economically valuable tasks. This definition hinges on two criteria: complete autonomy and broad economic applicability. According to Karpathy, the journey of self-driving technology exemplifies the societal dynamics of increasing automation and, by extension, AGI's potential evolution.
The Gradual Rise of Automation
Self-driving technology stands out due to its high visibility, significant economic footprint, large human workforce, and the complex challenge of automating driving. Unlike other sectors that have seen automation, the path to self-driving is a prime example of AGI's characteristics: accessibility, economic importance, workforce impact, and technical challenge.
Karpathy outlines the gradual development of driving automation. Initially, vehicles featured Level 2 driver assistance, where AI collaborates with humans in navigation, handling many low-level driving aspects while allowing human intervention. This partial automation is analogous to AI tools in various industries, like GitHub Copilot in programming, highlighting the incremental nature of AI advancement.
The leap to full automation, as seen in Waymo's driverless cars, marks a significant milestone. In cities like San Francisco, Waymo offers autonomous rides in a small, geo-fenced area, however, it helps showcase a future where AI will surpass human driving abilities. The transition to full autonomy will depend on public awareness, trust, preferences, and supply constraints in creating a large automated fleet.
Global Expansion: Challenges and Opportunities
The globalization of full automation, Karpathy notes, is a gradual, resource-intensive process. Waymo's current limitations to specific cities illustrate the challenges of expanding automated services, including adapting to local conditions and regulations. This expansion mirrors the broader trajectory of AGI deployment across various sectors, where scalability is both achievable and gradual.
Society's reaction to self-driving technology parallels AGI's potential impact. Despite significant advancements, public awareness and acceptance vary. Some view autonomous vehicles with curiosity and skepticism, while others adapt quickly. This range of responses suggests how society might adapt to AGI in various industries.
Economically, self-driving technology has both eliminated and created jobs. While driver roles are phased out, new positions in data labeling, remote support, fleet maintenance, and sensor technology emerge. This transformation reflects the broader economic implications of AGI, where work is not merely eliminated but refactored and reshaped.
The competitive landscape in self-driving technology, with companies like Waymo, Tesla, and others, mirrors the expected consolidation in AGI-related industries. As with self-driving, only a few companies may dominate the AGI space after an initial burst of growth and competition.
Karpathy envisions AGI as a gradual, society-involved evolution rather than a sudden, uncontrollable leap. Just as self-driving technology is transforming transportation, making it safer and more efficient, AGI promises to reshape various sectors.
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
The feature is now possible thanks to what Wes Morrill, Cybertruck’s lead engineering, points out is a huge rewrite of Tesla’s Trip Planner, which not only makes it much more accurate but will now allow users to choose their arrival charge percentage.
More Accurate Trip Planner
When you navigate to a destination, your Tesla will automatically calculate when, where, and how much you need to charge. While the process seems straightforward, Tesla deserves a lot of credit for creating a simple user experience because a lot goes into accurately determining this information.
Tesla has to calculate many moving pieces in order to accurately predict when and where you should stop. First, it needs to consider your driving efficiency and wind direction, terrain elevation, traffic, vehicle speed, and ambient temperature. It also needs to predict the best Superchargers to stop at, taking into account congestion and charger speed.
Now, according to Wes’ post this morning, Tesla has made backend improvements to the Trip Planner, thanks to a rewrite by several engineers, that bring even more accurate predictions to Trip Planner.
It sounds like these changes are mostly server-side, so thanks to their OTA connection, more accurate predictions should be available to all vehicles—no vehicle update is required.
Arrival State of Charge
Along with Tesla’s improvements to Trip Planner, Wes also stated that these improvements allow for a popular feature request — the ability to select your desired charge level at arrival.
Up until now, Tesla’s Trip Planner tried to get you to your destination as quickly as possible, which usually meant arriving with a low state of charge.
While this was fine if you have a charger at your destination, it’s not great if you don’t, or it could be even worse if there are no chargers nearby.
Max de Zegher said on X that he has heard the requests for a selectable arrival state of charge. Wes later clarified this by saying that these improvements to Tesla’s Trip Planner now allow for additional features to be added, such as “desired arrival charge.”
Actually trip planner got a huge rewrite on the back end. I had a great conversation with one of the engineers working on this over a morning run a few weeks back, it's pretty neat. The rewrite also unlocks additional features, like desired arrival charge.
Given that Max de Zegher’s comments came last night and Wes Morrill commented this morning, this feature request likely won’t arrive with the upcoming Holiday Update. Tesla actually hinted at such a feature being added in their last app update, so it does seem like they’ve already planned for it.
Although ‘Arrival State of Charge’ was on our wishlist for the Holiday Update, it looks like it may arrive soon after the holidays.
As Tesla update 2024.44.3 continues to roll out, we’re seeing more features in this update. While the update doesn’t include many new feature, it either improves existing features or rolls out a feature to new regions. Actually Smart Summon is rolling out to Europe, although with more strict restrictions. AutoPark is also rolling out to the UK and several other countries for the first time, and we’re now seeing improvements to Autopark arriving in North America and Europe.
As Tesla’s 2024.44.3 update continues to roll out, more features in the update are being revealed. While this update doesn’t introduce many entirely new features, it does refine existing features or expand the reach of others.
Notably, Actually Smart Summon is now rolling out in Europe, albeit with stricter regulations. Vision-based AutoPark is debuting in the UK and several other countries, while North America and Europe are receiving updates that further improve Autopark’s functionality.
Improvements to Autopark
As part of the same update, Tesla is making a batch of improvements to Autopark in Europe and North America. In the release notes, Tesla states that Autopark is receiving performance and visualization improvements. Unfortunately, Tesla doesn’t go into more detail here, but it sounds like these improvements could be the faster and more accurate Autopark enhancements Ashok Elluswamy talked about earlier this year.
Those improvements are expected to make Autopark more reliable, let it pick spots faster, move faster into them, and also shift between forward and reverse faster. The improvements would also let the vehicle park in tighter spots than before, with more accurate vision.
We’ll have to wait and see how this improved Autopark compares to the current version.
Your vehicle's Autopark performance and visualizatons have been improved.
Autopark in the UK and Other Countries
Countries in Europe that previously didn’t have access to Vision-Based Autopark, including the UK, will now have access to the new Autopark for the first time with update 2024.44.3 and later. This includes countries such as the UK, Ireland and Malaysia.
While these countries are receiving Autopark, it’s not clear whether it includes the Autopark improvements that other regions are receiving. The release notes don’t include the Autopark Improvements section in these countries, but that could be due to Tesla simplifying the release notes for these countries that are receiving Autopark for the first time.
With Autopark finally introduced in these regions, it’ll be the first time vehicles without ultrasonic sensors are capable of Autopark. Vision-Based Autopark is far more reliable, faster, and easier to use in more situations than the older USS-based solution.
_300w.png)
