Tesla's Full Self-Driving (FSD) v14 update represents a significant improvement in autonomous driving technology, showcasing enhanced safety, smoothness, and responsiveness, and is considered a major step towards the widespread adoption of driverless vehicles
Questions to inspire discussion
FSD v14 Features and Modes
🚗 Q: What is "Mad Max mode" in FSD v14?
A: Mad Max mode is an aggressive and very fast driving style that stays in the right lane, unlike the previous version's hurry mode, but may not be suitable for areas with strict speed limit enforcement.
🐢 Q: How does "Sloth mode" perform in FSD v14?
A: Sloth mode sticks strictly to the speed limit, making it unsuitable for Texas highways with 80-85 mph limits, potentially causing accidents in places like Houston.
😌 Q: What characterizes "Chill mode" in FSD v14?
A: Chill mode is similar to Sloth mode, staying in the right lane and going only 2-3 mph above the speed limit, while Standard mode goes 7-8 mph over and uses the right two lanes.
⏩ Q: How does "Hurry mode" in FSD v14 differ from v13?
A: Hurry mode is now geography-based, staying with traffic flow and avoiding left lane camping in places like Houston, but may cause issues in areas with strict speed enforcement.
FSD v14 Improvements
👁️ Q: What is the main improvement in FSD v14's perception?
A: FSD v14 has a larger context window, allowing it to detect objects 3-5 seconds earlier and respond faster to situations like narrowing lanes and upcoming exits.
🛣️ Q: How has lane-keeping improved in FSD v14?
A: Lane-keeping is smoother and more stable, with no brake stabbing or rolling motion at speed limits, making it more comfortable for passengers and "ready for Uber".
User Experience and Safety
🚕 Q: How do Uber passengers react to Tesla's FSD in Houston?
A: 100% of Uber passengers in Houston, where Tesla penetration is low, are "blown away" by FSD's capabilities, as reported by a driver with over 1,200 passenger experiences.
🛡️ Q: How much safer is FSD v14 compared to human drivers?
A: According to extensive testing, FSD v14 is claimed to be 2 times safer than human drivers and 100-1,000 times better than version 13.
Future Developments
🚘 Q: What is Tesla's CyberCab?
A: CyberCab is planned to be a fully autonomous vehicle with steer-by-wire and brake-by-wire capabilities, designed without a steering wheel and pedals.
🏭 Q: What production limitation might CyberCab face?
A: Due to federal regulations, CyberCab production may be limited to 2,500 units without steering wheels and pedals, unless Tesla requests an exemption.
AI in Business
🍋 Q: How has Lemonade implemented AI in their business processes?
A: Lemonade uses Agentic AI for automation, saving 150 hours over 3 years and improving product quality by reducing errors.
📝 Q: How can OCR accuracy be improved using AI?
A: OCR accuracy can be enhanced by using an LLM to correct handwritten OCR output, focusing on correcting words that don't make sense without adding new ones.
Insurance and FSD Integration
💰 Q: How could Tesla's FSD miles data benefit insurance companies?
A: Tesla's FSD mileage data through their API could enable companies like Lemonade to offer nearly free, pay-per-mile insurance to Tesla customers.
💵 Q: How much could Tesla owners save on insurance with FSD?
A: Tesla owners using FSD could potentially save $100-200 per month on insurance, potentially making FSD pay for itself through lower rates.
FSD v14 Overall Assessment
🏆 Q: How significant is the improvement in FSD v14?
A: FSD v14 is described as "so good" that previously noted issues are no longer apparent, with improvements in detection time and overall performance.
🌐 Q: When is FSD v14 expected to be widely released?
A: Based on extensive testing, it's anticipated that Tesla's version 14 FSD will be widely released soon, though no specific date is provided.
Key Insights
{KeyInsights}
#Tesla #FSDV14
XMentions: @Tesla @HabitatsDigital @JohnChr08117285 @MatchasmMatt @Farzyness @bradsferguson
WatchUrl: https://www.youtube.com/watch?v=7itP7iG7wOM
Clips
-
00:00 🚗 Tesla's FSD v14 update brings aggressive driving modes, including Mad Max, with improved lane changing and speed handling, differing from previous versions.
- JC shares his experience with Tesla's FSD v14, running version 14.1.3, which he received yesterday, to help understand the update since the hosts haven't received it yet.
- The speaker downloaded FSD v14 over home Wi-Fi, which took 2 hours, and then immediately tested it on the streets of Houston from 9:15 to 11:15 PM.
- FSD v14 in Mad Max mode on the highway felt extremely aggressive, maintaining high speeds and avoiding slower traffic by changing lanes efficiently, almost as if it was "looking for a way out" of slower traffic.
- FSD v14's Mad Max mode is noticeably different from Hurry mode, with users who normally use Hurry mode likely switching to Mad Max permanently, and exhibiting improved behavior, such as changing lanes when necessary.
- The Full Self-Driving (FSD) v14 modes behave differently in Houston, Texas, with Sloth mode following the speed limit, Chill mode slightly exceeding it, Standard mode going 7-8 miles above, and Hurry mode staying with traffic and avoiding left lane camping.
- FSD v14 drives 2-8 miles per hour over the speed limit, adapting to geography, and seeks the path of least resistance rather than strictly following lane markings.
-
08:07 🚗 Tesla's Full Self-Driving v14 shows significant improvements in safety, smoothness, and responsiveness, with enhanced vision-based decision-making and earlier response to road conditions.
- A hardcore Uber driver with 1,200 rides on version 13 notes minimal differences so far with version 14, having had only one navigational intervention and no critical interventions on either version.
- FSD v14's vision-based system can override navigation inputs to make safer decisions, such as taking a U-turn instead of trying to make an uncomfortable or dangerous maneuver.
- FSD v14 seems to see and respond to road conditions 3-5 seconds sooner than v13, handling complex maneuvers like lane changes and exit preparation more effectively.
- The Full Self-Driving (FSD) v14 update has significantly improved its performance, with a user reporting smooth lane keeping, acceleration, and braking, and no instances of "brake stabbing" or abrupt movements.
- The Full Self-Driving (FSD) v14 is ready for use on Uber, but the speaker is hesitant to use it in "Mad Max" mode once deployed, as they expect to have to tone down their driving style.
- The speaker discusses their experience with Tesla's Full Self-Driving mode, preferring "chill" mode for Uber driving, but planning to test "sloth" and "standard" modes to balance comfort and traffic flow.
-
22:16 🚗 Tesla's FSD v14 impresses nearly 100% of passengers with its significant improvements, making people more receptive to driverless robo-taxis after experiencing it firsthand.
- Nearly 100% of 1,200 passengers in a low EV-penetration area like Houston experience a full self-driving vehicle, specifically Tesla's FSD, for the first time and are positively blown away.
- Witnessing people's reactions to experiencing Full Self-Driving (FSD) v14 for the first time is like showing them an objective truth they never knew existed, with their surprise and curiosity comparable to a child taking their first steps.
- People will be less resistant to driverless robo-taxis once they experience the technology firsthand, unlike online critics who don't want to try it.
- Tesla's FSD v14 is 100-1000 times better than v13, enabling it to see and react to situations 3-5 seconds earlier, and is now safe enough for the speaker to consider riding in the backseat unsupervised.
- FSD v14 is like a cake mix that has just been started, showing promise but still requiring time and development to reach its full potential.
-
29:08 🚗 Tesla's FSD v14 impresses with exceptional performance, handling various scenarios, and is considered ready for wide release.
- The speaker finds FSD v14 to be exceptionally good, with no major issues, which is why they post unedited, lengthy videos showcasing their experience.
- JC, a YouTube and X creator, shares his experiences with Tesla's self-driving technology in Houston, including reactions from people like oil executives and a German national unfamiliar with Tesla's capabilities.
- The speaker thanks Bradford for letting him ride in a robo taxi and appreciates his hospitality.
- Tesla's FSD v14 struggled in heavy rain, with the Whimo ride bailing multiple times, but still managing to get to the destination, while two human drivers got into an accident on the same road.
- The speaker is extremely confident in the capabilities of FSD v14, having had a fun and adventurous experience with Robo Taxi and Whimo, and believes it's ready for wide release.
- The hosts thank everyone and bid farewell, then plan to discuss additional topics privately.
-
35:52 🤖 Tesla's Cybercab may need temporary steering wheel & pedals due to regulatory limits, but company considers designing steer-by-wire systems for easy conversion to robo-taxi fleets.
- Farzad claims Cyber Cab won't need pedals or a steering wheel at launch, but federal safety standards currently allow autonomous vehicles without these only up to a 2500 unit production limit.
- Tesla's production capacity for Cybercabs far exceeds regulatory limits, potentially forcing the company to either seek exemptions or modify vehicles with steering wheels and pedals.
- Tesla should consider designing steer-by-wire and brake-by-wire systems for their vehicles to facilitate easy conversion to robo-taxi fleets without physical steering wheels and pedals.
- Regulations requiring a temporary fix like a steering wheel and pedal in the Cybercab don't significantly impact Tesla's vehicle production ramp or quantity by 2026.
- New AI models are up to 10x more efficient than previous ones, rendering older models' compute as waste.
- Farzad gets the win, which the speaker reluctantly agrees with.
-
44:47 🤖 AI advancements like FSD v14 bring massive efficiency gains, will be deflationary, and require businesses to transform operations to avoid being left behind.
- Massive efficiency gains and optimization in AI are driving down costs, with staggering order-of-magnitude drops in cost per token over time.
- AI advancements, such as FSD v14, will be massively deflationary and require businesses to radically transform their operations, or risk being left behind.
- Current AI models, like LLMs, need improvement in harnessing deterministic processes to produce accurate and truthful outputs, as they often provide probabilistic and potentially hallucinatory results.
- A user experienced a hallucination with Grock, an AI model, when it provided a false fact about woolly mammoth bones being discovered by ancient Romans in modern-day Kentucky.
- A project that took 2 days to develop will save 150 hours over 3 years, exemplifying the efficiency gains from investing in automation.
- Using AI to make processes deterministic, rather than probabilistic, is key to automating solutions, such as in customer service, to produce a clear and specific outcome.
-
52:42 🤖 Tesla's FSD v14 and AI implementation, as well as its potential integration with other companies like Lemonade, showcase innovative approaches to scaling efficiently and improving customer experiences.
- DeepSeek's research found that directly interpreting pixel strings yields better results than OCR and translation for extracting text from PDF documents.
- Lemonade, an insurance company, is a prime example of effective AI implementation, achieving breakthroughs in efficiency and customer experience by integrating LLM with deterministic processes.
- Tesla's aggressive approach to utilizing AI, specifically with FSD v14, serves as a blueprint for companies to scale efficiently at low costs, pushing the boundaries of innovation.
- Using LLMs can improve accuracy of imperfect OCR outputs by correcting errors while preserving original text, and can also be applied to automate and improve repetitive and boring tasks.
- Lemonade's access to the Tesla API could enable better pricing for Tesla customers and potentially make it a go-to insurer for those with Full Self-Driving (FSD) capabilities.
- Tesla's API could enable Lemonade to offer pay-per-mile insurance and potentially discounts to Tesla owners using Full Self-Driving (FSD), making insurance cheaper and more tailored to individual usage.
-
01:02:39 👋 The creators discuss future plans, including a separate video on "her and power" and invite viewers to check out their channel, Rebellion.
-------------------------------------
Duration: 1:3:30
Publication Date: 2025-10-21T23:47:06Z
-------------------------------------