Featured Post

This is the Kodak Moment for the Auto Industry

Plug-In Drivers Not Missin' the Piston Electric vehicles are here to stay. Their market acceptance is currently small but growing...

Tuesday, December 1, 2020

Modern EV Era Celebrates 10 Years

By my accounting, the modern electric vehicle (EV) era started 10 years ago this month. It started in December of 2010 when both the 2011 Nissan Leaf and 2011 Chevy Volt rolled onto showroom floors. Since this unveiling, EVs have been making slow but steady progress. More models have been introduced and range and capabilities have continued to increase. Given the historic importance of this milestone, I think it's important to have a little perspective. EVs have tried to become the transportation of choice before and failed. What happened then and will this time be different?


Thomas Edison shows
off a 1914 Detroit Electric

Early 1900s

At the dawn of the automobile era, at the turn of the 20th century, there were several contenders to be the fuel of choice; there were steam-powered cars, electric cars, and gasoline-powered cars. It was not clear if all 3 of these fuel sources would co-exist or if one would dominate. 

Henry Ford's wife, Clara, choose to drive an electric car, rather than a gas-powered car from her husband's company. Gas cars were loud, dirty, and you could break an arm or wrist while trying to crank-start them. Whereas, EVs were clean and quiet and much more suited to a high-class lady of the time.  

Despite Clara's preference (and that of many people like her), after the self-starter engine was invented and the wrist breaking hand crank was removed, the gas-powered-mobile won out and began its century-long domination of transportation.

I find it ironic that the biggest hurdle to the gas engine's adoption (hand cranking) was solved by an electric motor. With this problem solved, gasoline rose up and became the dominant fuel, leaving electric vehicles relegated to golf carts, milk floats, and niche low-speed vehicles.



1974 Electric Prototype

1970s

An oil crisis began in 1973 when the Organization of Arab Petroleum Exporting Countries (OPEC) proclaimed an oil embargo. The embargo targeted several nations. The price of oil more than quadrupled in the US. The embargo caused an "oil shock" that had lasting effects on the global economy and politics, but not significantly on personal transportation.

The crisis created a demand for fuel-efficient cars and alternative fuels. Many electric prototypes were created and DIY electric conversions grew in popularity. Although, battery technology had not advanced significantly from the lead-acid batteries that were used at the turn of the century.

After negotiations, the embargo was lifted in March of 1974. Be it from arrogance or deference, the world seemed to forget about the need for alternative fuels and continued their dependency on a single primary fuel source and we would pay for this again and again with another oil shock at the end of the decade and more to follow in each of the next three the decades ahead. 

None of the various EV prototypes from this decade ever made it into mass production but one ray of light that survived from this era was that the home conversion hobbyists persisted with groups like the Electric Auto Association.*  

 

GM EV1

1990s

The 1990s were the next attempt at an electric revolution. The GM EV1 rolled out in 1996 and was the vanguard of this wave. The EV1 and other EVs of this era had owners that loved them. Regenerative breaking helped extend the range and batteries had had their first major breakthrough with the Nickel-Metal Hydride (NiMH) chemistry, although most buyers (leasers actually) still opted for the cheaper lead-acid option. 

Unfortunately, this attempted revolution was also put down. Automakers sued California to eliminate the state's Zero Emission Vehicle (ZEV) Mandate. The automakers eventually won and most of the EVs were collected and crushed when their leases ended. To no avail, drivers held a mock funeral and candle-lit vigils to try to save their cars; some were even arrested while blocking the trucks carrying their cars to the crusher. Many of the drivers that held vigil would go on to found PlugInAmerica to promote EVs, public awareness, and better EV policy.

 You can see the entire intriguing story of this era in the documentary Who Killed The Electric Car? 

The people that would later go on to found Tesla, noted the devotion that owners had to a compelling EV (more on this later).

This era didn't populate the world with EVs, but it did demonstrate that there's a market of passionate drivers that want EVs which the existing automakers were unwilling to satisfy.



2011 Nissan Leaf SL

2010s

This finally brings us to the modern era of EVs. This era kicked off when both the Nissan Leaf and the Chevy Volt started selling in December of 2010; soon followed by the Tesla Model S in June of 2012. This generation of EVs had a few things that the previous revolution attempts didn't have: 
  • Lithium-ion Batteries
  • Major Automaker Support
  • Tesla
2011 Nissan LEAF Battery Pack

Lithium-ion Batteries

Lithium-ion batteries power most of our modern mobile electronics from smartphones and tablets to smartwatches and earbuds. This has meant that a lot of money was pouring into battery R&D for longer runtime for these devices. EVs were the unintended beneficiary of this mobile digital revolution. All of the EVs coming out in this era are currently powered by Li-ion cells.

Batteries are the most important component in an EV. They are, by far, the most expensive part of the vehicle, they're one of the biggest determiners of range, and avoiding battery degradation is often the limiting factor to performance and recharge time.  

In this era, EVs have finally made it to a production level that they are no longer just dependent on battery advances from the consumer electronics realm. Automakers are funding battery research, partnering with battery companies to build out capacity, and designing custom form factors and chemistries to better meet the demanding cycle-life that EVs require. Advancements here enables lower prices, better performance, longer range, faster charging, longer lifespans... Batteries are the crux.

In the 1990s, NiMH was the advanced battery tech of the time but there was a problem. This battery chemistry was covered by one primary patent and that patent fell into the hands of an oil company. They restricted the license such that NiMH batteries could only be used in hybrid vehicles and not in pure EVs. 

Li-ion batteries have no such patent encumberment. Because Li-ion was used in so many different types of consumer electronics, dozens of companies had patents for various improvements and in many cases, due to legal spats or partnerships, these patents were cross-licensed. Additionally, Li-ion was invented in 1985. This means that many of the initial patents have been long since expired. 

Sidebar: The inventors of the Li-ion battery; John B. Goodenough, M. Stanley Whittingham, and Akira Yoshino; received a Nobel Prize in Chemistry in 2019. 

Major Automaker Support

In the 1990s and aughties, when automakers were required to make EVs, these were generally "compliance cars" made in limited quantity with just enough range to meet state mandates. The limited production made them hard to find even within California and if you lived outside of the state, these cars were nearly impossible to find. However, the Leaf and Volt were being sold nationwide (or worldwide) and not in limited quantities. These were the first EVs that you could walk into a dealership, purchase, and drive off in an EV. That's assuming the dealership didn't try to steer you into the gas car they had on special that week. Dealerships were (and in some cases still are) an obstacle to EV adoption (but that's another story). 

A variety of factors from consumer demand to regional climate goals have pushed automakers to make EVs. Nearly all of the legacy automakers either have EVs on the market or have plans to have them out soon. Here are some of the currently announced plans: 
  • Audi - 20 EV models by 2025 
  • BMW - 25 electrified* models by 2025
  • Daimler / Mercedes - Plug-in option of every offering by 2022
  • GM - 20 EVs by 2023 
  • Fiat-Chrysler -  30 electrified* models by 2022 
  • Ford - 40 EV models by 2022
  • Hyundai - 44 EV models by 2025 
  • Renault-Nissan-Mitsubishi - 12 EV models by 2022 with annual volumes of over 1 million units per year
  • Toyota - 50% of sales in 2028 will be electric
  • VW - 70 EV models by 2028 
  • Volvo - Polestar brand will be 100% electric
* "electrified" in automaker-speak often includes hybrid vehicles (the non-plug-in type). 

Certainly, some of these more aggressive goals will be missed or delayed, but the direction is clear. Automakers are going electric.



Tesla 

Tesla, or a company like it, might seem like an obvious inevitability today, but that was not the case when they started. The prevailing logic was that EVs had been tried in the past and there was no market, that California was not the place for a car company HQ, and that Tesla's $100k+ sports car would sell all of 5 to California billionaires and then the demand would dry up. Variations of this demand narrative continue to this day despite being proven wrong year after year.

Despite the naysayers, Tesla overhauled the way that people perceived EVs. EVs were considered slow. Tesla's cars were really fast. EVs were often weird looking little things. Teslas were sexy and (other than Roadster) were large vehicles. EVs generally has less than 100 miles of range, Teslas had 200+ miles of range. EVs were restricted to a radius of travel equal to about half of their range. Teslas had a vast fast-charging network. Tesla changed the public perception of what an EV could be. Tesla made EVs fun and exciting.

Tesla's business model seemed to be to knock down every objection to the adoption of EVs. And knock them down they did. Arguably, with one exception, affordability. This is one area where they have made great progress going from the $100,000+ Roadster, to the ~$35,000 Model 3 and it's a goal they continue to strive toward. As Tesla ramped their volume, they consistently reduced the price. A $25,000 vehicle was recently announced and expected in 2024. 

There is no question that Tesla has changed the game. People that had never bought a new car were doing the "Tesla Stretch" and buying a new Tesla that was often twice as expensive or more than any car they had purchased previously. People were drawn to Tesla for the performance, the tech, the fun, and the zero-emission factor was almost secondary. Tesla demonstrated that if you made a compelling EV, even an expensive one, there was a market for it.


The Tesla Wake-up Call 

Legacy automakers initially dismissed Tesla as a low-volume niche automaker. As Tesla has continued to make inroads into new markets, the legacy automakers have finally started to take notice. US automakers could initially dismiss Tesla. Volumes started small and Model S mostly ate into the luxury sedan market dominated by German brands such as Audi, Mercedes, & BMW. Next, Model 3 ate into the fuel-efficient sedan market and mostly impacting the Japanese automakers. But Tesla has plans for a truck, a Cybertruck. Trucks are the heart of the American auto industry. The top-selling vehicles in the US are trucks. Tesla can no longer be ignored by GM and Ford. Tesla's products, technologies, and plans are now closely examined by the worldwide auto industry. 


Viva La REVolución

(What Makes This Time Different?)
EVs have been here before, they've gained traction with an enthusiastic early adopter niche, but they were never able to go mainstream; never able to cross the chasm. Several things make this time different. 

With Li-Ion, there is a battery technology that allows for long-range and fast recharge and it's continuing to get better. There's support (to varying degrees) by the major automakers. There's a standard-bearer in Tesla that shows that EVs can be great and there's a swath of start-ups trying to follow in Tesla's footsteps (or even trying to leapfrog them). These start-ups are well funded by venture capital and pre-revenue SPAC-mania from investors with FOMO on the next Tesla. The genie cannot be put back in the bottle.

This time, the EV revolution will not be crushed! 

What a Difference 10 Years Can Make

Looking back on these 10 years, it's amazing how much EVs have improved. Compare the 2011 Nissan Leaf with the upcoming Nissan Ariya, the Ariya has more than 4 times the range. Or comparing the initial Tesla Model S (which won Car of the Year) to the recently announced Plaid Model S. The improvements in tech, performance, and range are amazing. If this trend continues, there will be no reason to even consider a gas car by 2025 for 99% of drivers. We'll move beyond oil.  

I've wondered aloud how good EVs would be today if the automakers had continued their 1990s efforts (or even better, their 1970s efforts). Taking this to its extreme, what if Clara Ford had won the day 100 years ago. Over the last 100 years, we've made great strides in internal combustion technology, we were just working on the wrong problem. What would that alternative history of battery advancement and transportation look like? Would we have battery-powered transcontinental flight by now? 


Disclosure: I am long Tesla
* I'm a member of my local EAA chapter, the OEVA

Tuesday, November 24, 2020

Tesla Investing: Bean Counters vs Visionaries

"Bean Counter vs Visionarie" commissioned image by @lastly_the_squirrel_thing

Since its IPO, Tesla has been a controversial stock. It has attracted bulls and bears. It has been one of the most-watched names out there and Tesla has become the most valuable auto company in the world by market cap. This is despite the fact that Toyota sells about 25 times more cars annually than Tesla. This dichotomy has been a defining feature of the stock and perhaps the company in general. Evaluating the company as an investment, there are two primary camps: the bean counters and the visionaries.

The bean counters look at the balance sheet. They see that (prior to Q3 2019) Tesla lost money in nearly every quarter. They see Tesla's volumes are minuscule compared to other automakers. They see that the company has had to raise capital year after year to fund operations. 

The visionaries see that Tesla has a world-changing vision that could lead to significant growth and profit. They see that Tesla has products that their customers love. They see that Tesla is on the forefront of several trends that will change the way we get from A to B.

I'm not going to pretend to be unbiased here. I'm a longtime Tesla customer and even longer TSLA shareholder. I've appeared on a podcast discussing (arguing) Tesla's quarterly results with (against) 'Montana Skeptic'. 

Why Auto Analysts Got It Wrong 

When many brokerages initiated coverage of Tesla, they looked at the product (likely a Model S at that time) saw 4 wheels and assigned an automotive analyst. As rational as that sounds, Tesla is not a car company, sure they make cars, but they are a technology company and cars are one of their technology delivery 'vehicles' (more on this in the next section). 

Tesla is a "Story Stock." This means that you are not going to find the value of the company by looking at the financial results from past quarters. Instead, it means that the company has big ambitions and the value of the company rests in their ability to deliver (or not) on that vision. This is a radically different method of evaluating a company than looking at fundamentals, comps, price to book, dividend yield, or discounted cash flow and it requires a radically different skillset. 

Tesla Is Not A Car Company 

I assert that Tesla is not a car company. Yes, they make cars, but that is a result of their larger mission, not the ends in itself. So, if Tesla is not a car company, what are they? They are a technology company, they are an end-to-end renewable energy company, they are an engineering company that is not afraid to tackle hard problems. 

To varying degrees, Tesla consists of:
  • A battery pack manufacturing company - Powerwalls, Powerpacks, Megapacks, car battery packs... Tesla purchased ATW Automation in 2020. The company made battery modules and packs for the auto industry.
  • A battery pack management provider - All of the above battery packs need control systems for charging and discharging, heating and cooling
  • A battery cell manufacturing company - Tesla purchased Hibar Systems in 2019 for their battery cell automation solutions. Tesla's current "pilot line" for new cell types qualifies as one of the top 10 cell plants in the world for kWh output. Tesla partners with Panasonic, CATL, LG, and others, but they also have their own cell production efforts. 
  • A battery research and design company - Tesla purchased Maxwell Technologies for their dry electrode innovations in 2019. The battery cell is a fundamental building block for many of Tesla's products. Advancements here leads to better products, more revenue, less capex, and more margin.
  • A manufacturing automation design company - Tesla purchased automation expert Grohmann Engineering in 2016, Compass Automation in 2017, and Perbix, a maker of highly automated manufacturing equipment, in 2017. They don't want to just make things the same way Toyota, GM, and the others do it. They want to rethink manufacturing. This means they cannot just buy the stamps and presses that are available. They have to invent the machines to make their products.
  • An AI hardware company - Tesla developed its own AI inference engine to comprehend the world around the car from the streams of data coming from the cameras and sensors. 
  • An AI software company - Tesla hired Andrej Karpathy away from OpenAI in 2017 and purchased DeepScale Inc. in 2019. DeepScale was known for their innovative energy-efficient deep neural network AI computer vision system.
  • An energy arbitrage company - Tesla's AutoBidder software allows Tesla to buy and sell energy based on grid supply and demand. This is currently used for their industrial-sized battery systems, but someday we might see residential Powerwall ganged together to form a virtual power plant that can earn credits for their owners. 
  • A solar retailer and installer - Tesla purchased SolarCity in 2016
  • A solar cell reacher company - SolarCity bought Silevo Inc. in 2014. 
  • An insurance company
  • A vehicle recharging company that sells home charging equipment
  • A vehicle "refueling" company that deploys and operates a network of fast-charging stations and destination charging
  • A guerrilla marketing company - whether it is launching a car into space, selling short-shorts, smashing windows, or selling Tesla Tequilla, Musk and co. know how to generate buzz in our modern social-media-driven news cycle
  • A glass developer - Tesla has develops specialized glass for their vehicles and solar roof products. As I write this they have a job opening on their website for a Glass Studio Specialist to "work closely with design team on the manufacturing of prototype glass and product development associated with advanced glass compositions and inner layers."
  • A worldwide car "dealership and Service Network" - Unlike most auto manufacturers, Tesla does not have a network of independent dealerships to sell and support their vehicles. This means that Tesla has to provide this service themselves in every region where the vehicles are sold. 
Viewing a company superficially only by its most visible products is a recipe for an incomplete analysis. This type of thinking would have led you to only see Apple as only a computer company, Amazon as only an online bookstore, Google as only a search engine... And today, if you are viewing Tesla as only an automaker, you are missing the majority of the value of the company.

Tesla The Outlier


Tesla stands in stark contrast to traditional car companies. Tesla vertically integrates. They make their own seats, infotainment system, navigation software... sure they still have hundreds of suppliers, but they are component suppliers for Tesla's solutions. Whereas legacy automakers have expertise in drivetrains and manufacturing, and little else. Any hard problems encountered by most automakers are farmed out to their suppliers. The automaker may get the result they are looking for, but they do not develop the in-house skills to further optimize for the next release nor to tackle similar problems that will occur in the future. And whatever solution that supplier designed immediately becomes available to every other competitor that also contracts with that supplier.

The auto industry is in a period of disruption. This means that past results are no longer a valid predictor of future results. The companies that are willing to sacrifice short term gains, to establish a long term portfolio of compelling products will be the long term winner. However, when management is 'graded' only on their short term quarterly results, you are not going to get a viable long term plan.

Moving The Goalposts

Critics have consistently moved the goalposts with regards to Tesla. They said there was _no_ market for EVs. However, when the Roadsters sold, they dismissed this and said there is only a niche market for rich treehuggers; this is the only market and Tesla will remain a bespoke automaker at best. Then Model S came out. It won award after award. It was a reimagining of what a car could be. It will never sell, the touchscreen is too big and people like knobs and buttons, they said. Tesla does not have the supplier agreements to produce more than a hundred per year, they said. Tesla ramped production to 100,000 per year and brought out Model X. With that achieved, the critics conceded the high-end luxury market but said that Tesla could never deliver a high-volume vehicle. Well, Model 3 ranks as the world's all-time best-selling plug-in electric car, with more than 500,000 delivered and Tesla's cars accounted for 81% of the battery electric vehicles sold in the United States in the first half of 2020.

One of the drums the critics of Tesla bang the loudest is Tesla's acquisition of SolarCity. If you view Tesla as just an automaker, the purchase makes no sense. However, if you look at Tesla as an end-to-end energy provider, it's a great fit. Tesla already has stores in malls and other retail locations to sell their cars. Someone buying an EV is far more likely to be interested in solar than the average car buyer. Once you have solar, you might be interested in a Powerwall to store the energy. And at the industrial scale, Tesla could sell massive solar installations and the Powerpacks or Megapacks to store the energy these solar panels produce. The SolarCity acquisition was an investment. It allowed for better utilization of their retail space, it allowed for cross-shopping and up-sales. It was an investment in Tesla's energy storage business.

Now the critics point out that competition is coming. I have two replies to this. One, the legacy automakers don't have the right skillset or engineering talent to make a competitive product. They have a long history of making gas-powered vehicles, they love the rumble and noise; it's a long and difficult path to change the culture at a large company. The half-hearted efforts that they've made to-date are evidence of this. Two, recall that Tesla is not a car company. When a competitor comes out with a car and a potential buyer asks, "Can you sell me solar panels to charge it? And have my solar and charging data all in one app? Does it have over-the-air updates? Can you sell me insurance? Can you sell me a battery pack to mount in my garage to store solar energy and to use during a winter storm black-out? Can I drive coast-to-coast on a fast-charging network?..." A point product is not the same as a full solution.

As of late October 2020, Tesla short-sellers have lost $27B betting against Tesla this year. Perhaps it is time they stop moving the goalposts and admit that they didn't understand the company.

Tesla's Success Is Self-Evident (now)


As I was writing this, Tesla achieved an incredible milestone, they were inducted into the S&P500. This is a major accomplishment for any company. With this milestone, I assert that Tesla has moved through all three stages of truth:

Stage 1: They were a niche player making Roadsters for rich Hollywood types. They were, at best, ignored and, at worse, ridiculed.

Stage 2: Tesla was the target of a massive Fear, Uncertainty, & Doubt (FUD) campaign. As just one example, the Koch brothers (who are deeply embedded in the fossil fuel industry) financed a multimillion-dollar misinformation offensive against electric vehicles. Tesla was the primary target for much of this FUD. These erroneous messages were often prominently placed in search results and occasionally picked up by media outlets. If you believed things like this, EVs polluted as much, if not more, than gas-powered cars, were nearly certain to burst into flames, and had to have their "toxic" batteries replaced every 3 to 5 years.

Stage 3: Now that the UK, California, and other places have put a date certain moratorium on gas and diesel car sales and Tesla has shown that EVs can be made GAAP profitably, there is a wave of fast-follower EV startups and the legacy carmakers are rushing to bring a number of EVs to market within the decade.

Perhaps we've finally come out on the other side of this. After years of the mainstream media predicting Tesla's demise, now we see articles with titles like, "Why Tesla stock could go to $1,000, according to a Wedbush analyst" on Fortune.

Conclusion 

The 'value' of a company cannot always be found in a spreadsheet.

Tesla has shown that there is a demand for EVs and specifically for their EVs. They've shown that they can sustain GAAP profitably and they are being inducted into the S&P 500. These achievements should allow many of the bean counters to now see the value of Tesla. There is a megatrend toward transportation electrification and Tesla is the leader in the space. Add Tesla's ambitions for energy, semi-trucks, autonomous vehicles, robotaxis, and more and you can see why the stock trades at a premium. 

For me, as an investor, when I purchased a Model X in 2016, it became obvious that Tesla was on the path to success and the pack was nowhere to be seen. This was my 3rd EV and I had been following the industry press closely. Tesla had the lead and no one was even trying to catch them. 

At the start of this article, I said that Tesla is a story stock and the value of the company rests in their ability to deliver (or not) on that vision. Tesla has the swagger to attack the best engineering talent. Talented engineers are key to achieving their goals. Combine that with the history of accomplishing things that no other company has been able to bring to market and the future looks bright for Tesla.

Disclosure: I'm long Tesla
This is not investment advice. Do your own research, consult a professional. 

Sunday, September 20, 2020

4 Years of Tesla Model X Ownership

Four years ago, I bought a Tesla Model X. I've written an annual "Owner's Report" each year; you can see previous years' reports here: 1, 2, 3. In these reports, I talk about our adventures in the X, like the time we took it to a drive-through safaritowing our camper, or the hack we used to haul home a tree on the glass roof. Here's the year 4 report:  

2020 has been anything other than a typical year. The pandemic has meant that I'm now working from home. I'm lucky that I have a job that allows me to work from home. This means that the Model X is not getting as many miles as it used to, but we did have a few trips worth mentioning. 

Vista Ridge at Mt Hood Meadows

Mt Hood

During the 2019/2020 season, we took the X to Mt Hood for several skiing day trips. A quick stop at the Sandy, Oregon Supercharger gave us plenty of juice to crank the heater and defroster as we climbed the mountain and zipped through Silent Rock with the surefooted all-wheel-drive of our 90D. On one of our trips, when we should have stayed at the Supercharger a few more minutes, we stopped at Ski Bowl on our way home and charged up at the West Coast Electric Highway station there while we had dinner after a long day on the slopes. 


The Astoria Column

Fire Dancer at Fort George in Astoria, OR


Festival of The Dark Arts

Our last pre-COVID trip was to Astoria, Oregon for the Festival of The Dark Arts. This "Carnival of Stouts with over 70 rare and unique offerings from 50+ breweries" is at Fort George. We stayed the night in Astoria and toured around the city the next day. It was a fun time and sadly, given the pandemic, I doubt the event will be held in 2021. We charged at Seaside on the way there and drove straight home on the way back. Unfortunately, we could not charge overnight at our hotel (not even on a lowly 120V outlet). But the onboard nav said that we'd make it home with a 9% charge. We arrived home with 7%. We arrived home a little lower than the nav estimated likely because we had our snow tires on, but we made it with plenty of buffer to avoid any range-anxiety. As a backup plan, we could have stopped at any of the various CHAdeMO stations in Banks or Hillsboro if we needed a few extra Watt-hours.



Comet Neowise

One dark night, we drove out to a friend's farm, far from the light pollution of the city to get a good view of Comet Neowise. In addition to a great view of the comet, we came home with some farm fresh eggs. This was an easy there and back drive with no need to charge in-route. 


Rockaway Beach

Soon after the Astoria trip, we went into lock-down mode; no more commuting, trips to the gym, tipping a pint with the lads, or dining out... The only miles added were for the occasional take-out food or a grocery run. Our only lockdown exception was a socially-distanced end-of-summer trip to the beach. We chose Rockaway Beach as our destination. There are no Superchargers along this route, but there are a few destination chargers in Tillamook, so we timed our trip for a lunch stop at the Blue Heron Cheese Company.

A peacock and an old tractor at Blue Heron Cheese Co.

As we pulled into the Blue Heron parking lot, we saw that the destination charger was occupied by a Model 3. Luckily there was a J1772 available next to it. We grabbed a quick lunch and sat at their outdoor seating area as we ate and charged up. Soon the 3 left and we were able to move over to the faster destination charging station. We took a stroll around their facility seeing a white peacock, chickens, rooster, goats, llama, and more; then we were back on the road. 

At Rockaway, we hiked the Cedar Wetlands Preserve on a nice boardwalk over the wetlands to a giant cedar. If you ever plan on going there, here's a quick tip: the trailhead is right off of Hwy. 101 at the "Welcome to Rockaway" sign at the south end of the city. This is not where the onboard nav (or Google maps) sent us. After our short hike, we headed to the beach and spent the rest of the afternoon playing in the sand and surf. We had dinner at a little diner near the shore and headed home; arriving home with 28% charge remaining.

Boardwalk to the Giant Cedar Tree 

Wild Fires 

Soon after our beach trip, the entire West Coast seemed to go up in flames. There were at least a dozen fires in various parts of Oregon. Some areas were evacuated, some Oregon towns like Talent and Detroit were severely damaged. The skies over the western part of the state darkened with smoke. The smoke was so bad that flights in and out of the region were canceled. The masks that we'd been wearing for COVID, now served double duty if you had to be outside. 

Why do I mention this? Because the bio-hazard defense mode air filter in the Model X came in handy when we had to go out during the ~10 days when the levels of PM2.5 and PM10 were unhealthy. 

Smoky Skies in Oregon


Year 4 Stats 

Miles added: 5,602 (9016 km) - my lowest year yet
Total Milage: 38,309 (61,652 km)
This Year's Battery Degradation:  2.6% 
Battery Capacity Remaining: 91% (233 miles) 
Software Upgrades: 13 
Current SW version: v10.2 2020.32.2


Degradation

As I write this, it's a few days before Tesla's Battery Day 2020. I expect that they will announce the "Million-Mile Battery" and much more. The million-mile battery will be a milestone for EV longevity. It will mean that the batteries will far outlast all but the toughest road warriors and even for them, battery replacements will be infrequent, if ever. It will also mean that batteries could have a significant second-life in storage applications or that salvaged batteries could be used in conversion projects and still have significant range and lifespan. 

I, unfortunately, do not have a million-mile battery. I have a 90 kWh pack from 2016. The 90 kWh packs have proven to be Tesla's fastest degrading pack; with some older 85 kWh packs having longer range today than many of the 90s. To be clear though, even Tesla's "worst pack" is far better than the degradation that my 2011 Nissan Leaf suffered

Here's the chart of my 90D's pack degradation:


As you can see, I have lost about 9% range. Looking at charts where other drivers have aggregated their data, it looks like degradation flattens out significantly after ~10% degradation. That is the way the green line is bending in our chart above and I certainly hope that proves to be the case here. With less than 40 thousand miles over 4 years and most of my charging done at home, the batteries are not in an especially taxing situation, yet there's more degradation than I'd like to see. If it flattens out, we'll be fine. If we keep losing 2% per year, road trips will start to become difficult by year 6.


Future of the Model X

Some members of the Tesla community think that Model X is not long for this world. The design is ~5 years old and due for an update or to be discontinued. Musk has described it as the "Faberge egg of vehicles" and the many problems they had getting the falcon-wing doors right for production are well documented.

I'm not in the camp that thinks the X will be discontinued. This is their aspirational or halo vehicle in many ways. This is the vehicle that appears in music videos with the falcon-wing doors swung high. This is the car with the Trans-Siberian Orchestra Wizards in Winter song and dance easter egg. 

Sure the Model Y fills in much of the SUV/CUV market that only the X previously served in Tesla's line-up. As great (and more affordable) as the Y is though, it's not the same type of aspirational vehicle as Model X. (Let me know when you see a Model Y in a rap video) and you can't spell S3XY without an X.

Each Tesla vehicle (announced and in-production) fills a niche. The Roadster is the fast one. The Cybertruck is the badass one. The S is the luxury sedan that's faster than a Porsche 911. The 3 and the Y are the relatively more affordable ones. And all of them are sexy.

Given all this, I hope to see a redesigned Model X with more range, a vertical screen, more towing capacity, faster charging, and other upgrades unveiled as the 'one more thing' at battery day or another Tesla event soon.

I plan on buying another X in ~2025, so they better still be making them ☺☺☺


Parting Thoughts

This has been a strange year in many ways. I'll continue to work from home until at least June of 2021, so I expect year 5 to be a similar low mileage year. When COVID struck, we heard a lot about flattening the curve. For year 5, I'm hoping that my battery pack's degradation curve flattens out too. 

I still love the X and have no regrets. If I had it to do over again, knowing what I know now, maybe I'd have waited a few months and got a 100D with AP2, but if I waited for the 'perfect' Tesla, I might still be waiting for some feature or upgrade that has the Tesla followers abuzz and I'd have missed out on 4 years of incredible fun. And when I do finally upgrade, I'll be getting all this and more. If you see a Tesla that meets your needs and your budget, don't hesitate, grab the opportunity. Will there be other innovations from Tesla? Absolutely. You (and I) can get those in our next Tesla.

Disclosure: I am long Tesla stock. 

http://ts.la/patrick7819

Friday, July 31, 2020

Tesla and The March of Nines to Full Self Driving


“It always seems impossible until it's done.” ~ Nelson Mandela



Tesla is working on full self-driving (FSD) cars. Some have said this is impossible. When it is done, this will be added to the growing list of things that Tesla has achieved that were once branded impossible. These once impossible achievements were not always delivered on the promised timeline, but they, nonetheless, arrived. Trent Eady, (the same person tweeting to Elon Musk in the image above) said it well when they wrote, “If Musk promises you the moon in six months and delivers it in three years, keep things in perspective: you’ve got the moon.” How long will the FSD moon take to be delivered? That's what we'll explore below.

In early July of 2020, at the World Artificial Intelligence Conference Musk said, “I’m extremely confident that Level 5 autonomy, or essentially complete autonomy, will happen, and I think it will happen very quickly. I remain confident that we will have the basic functionality for Level 5 autonomy complete this year.”

There’s a massive amount of work with each order of magnitude of reliability. This is the long 'March of the Nines'.

In the Q2 Financial update later in July, Musk reiterated his confidence in FSD, “The car will seem to have just like a giant improvement. We’ll probably roll it out later this year. [It will] be able to do traffic lights, stops, turns, everything pretty much. Then it will be a long march of nines, essentially. How many nines of reliability are okay? So it’s definitely way better than human, but how much better than human does it need to be? That’s actually going to be the real work. There’s just a massive amount of work with each kind of order of magnitude of reliability.”

What Are Nines

Musk mentioned the "nines of reliability." What are the nines? There are plenty of systems where 99% reliability is sufficient. If a video game crashes occasionally, it might be annoying, but no real damage was done. Whereas, something like a flight control system needs to be 99.999% reliable or better. However, it can be tedious to say, “ninety-nine point nine nine nine," so the verbal shorthand is to ignore the decimal point and just say the number of nines, e.g., 99.999% is called five-nines. It would be nice if we had 100% reliable systems, but that is an impossibility. Failures occur, components age, cosmic rays flip bits... so you have a backup, but the backup could fail too, so you have a backup for the backup, but that could fail too... Each layer of backup improves the overall system reliability, but, short of an infinite number of backups, it's not impossible that all of the backups fail at once either coincidentally or due to a common cause.

Why Nines Matter

Here's a simple example of why 99% is not good enough. There are about 150 billion credit card transactions each year totaling about $10 trillion. If these transactions were correct 99% of the time, that would be 1.5 billion transactions (~$100 billion) with errors each year. A system at this scale needs to be better than 99% reliable. Five-nines (99.999%) of reliability would reduce the annual error rate to “only” 15 million errors per year. Seven-nines would reduce it to 150,000 errors (still $10 million in annual errors). This is a system where it literally pays to improve reliability.

What is the March and Why is it So Long?

There are a few ways to look at this and it is different for any effort. Generally speaking, the more complex the system, the more difficult it is to improve its reliability. In a complex system, it can be hard to see the 2nd and 3rd order effects of potential changes.

There are several ways to view this concept; let's look at the 80/20 Rule.

The 80/20 Rule or Pareto Principle has many applications. For our purposes, we'll consider software development and we'll call feature-complete the 80% mark of the effort for a highly reliable application. Let's say that 80% effort took 8 months. That's an average of 10% each month, so the project should be 100% complete in just 2 more months, right? Unfortunately, the last 20% does not scale linearly like the first 80%. This last 20% is where all the hard problems live. These are the bugs that only show up intermittently, in full integration testing, the race-conditions, the new bug fix that would require nearly a complete rewrite, or the scalability problems that only show up at your biggest customer's site... 

That remaining 20% becomes its own 100% effort. An additional eight months later you might have 80% of that 20% done, and the cycle repeats. The number of iterations you go through depends on the level of dependability that your application needs. Let's look at a progression and see how long it would take for this fictional application to reach five-nines and just for fun, let's look at ten-nines too.

Cycle    Reliability %       Nines
180~1
2961
3992
499.8~3
599.973
699.994
799.9995
899.99975
999.999956
1099.999997
1199.999998~8
1299.99999968
1399.99999999
1499.99999998~10

According to our 80/20-rule table, it will take 7 development cycles to hit five-nines. In this example, each cycle was 8 months, so that's 56 months or 4 years, 8 months.

Imagine the conversation where you were 8 months into a project and you were 80% done and then you told your boss or customer that the last 20% will take 4 more years. They might think you're sandbagging them. It's hard to believe that it could take as long to go from 99.99% to 99.999% as it did to go from zero to 80% but this is why this is often referred to as “the long tail."

The 80/20 rule is straightforward, but as I mentioned at the start, no two projects are the same, progress is made in fits and starts and the 80/20 rule is just a rule of thumb and only one possible model.

If the problem that you're tackling has a long tail, then early progress must not be linearly extrapolated to determine a likely completion date.


Another, more academic, method to view the long tail is the Empirical Rule. The empirical rule is also known as the "68–95–99.7 rule." You can find tons of equations in project planning books on this, but we'll keep it simple here. With this method, each iteration is accounting for another standard deviation of input, defects, system behavior... on a normal continuous probability distribution curve.

Cycle   Reliability %    Nines
1680
2951
399.72
499.994
599.99996
699.9999998~9

If our hypothetical application follows the empirical rule, we'd achieve five-nines in just 5 cycles or 3 years, 4 months. Remember when it seemed like we could be done in just 10 months? If the problem that you're tackling has a long tail, then early progress, although great, should not be linearly extrapolated to determine a likely completion date.

How Good Are Human Drivers?

The goal is for an AI driving systems to be better than human drivers. In our article, “AI Driver: Safer Is Not Enough," we discussed why self-driving cars will need to be more than just a little better than human drivers, but let's just look at human drivers and see where that bar is set.

Despite the accident reports that cause mile-long traffic jams that seem to happen all too often, human drivers do a remarkably good job, all things considered. Humans have poor reaction time, are unable to look in multiple directions simultaneously, are distractable, have several blind spots, occasionally fall asleep at the wheel, drink & drive, have medical issues... yet humans are only involved in an injury collision about once every 1 million miles, and a fatal crash only once per 100 million miles or so. This is an injury collision avoidance performance of six-nines and a fatal collision avoidance rating of eight-nines.

Applying the Nines to Tesla Full Self Driving

Musk did not promise FSD by the end of 2020, he said he was confident that they would have “basic functionality for Level 5" by the end of the year, then “the real work" begins. Musk stated, “There’s just a massive amount of work with each kind of order of magnitude of reliability." I think Musk's assessment of the “real work" effort after feature-complete is accurate and an under-appreciated aspect of system development; remember our simple 8 months to feature-complete project that took another 3 to 4 years to reach five-nines. As we see from the human driving data, FSD will need at least six-nines to be as good as a human.

Every Tesla made today has eight cameras, a front-facing radar, and ultrasonic sensors. These sensors are important, but the heart of the system is a deep learning artificial intelligence. All of the various sensor data, GPS info, navigation, speed data, and more are streamed to the AI system where it attempts to make sense of the world around it, make real-time decisions, and get you to your destination without an insurance claim or a hospital visit.

The hard part of a self-driving system is not simply staying in a well-marked lane; it's dealing with all of the edge-cases. Computer systems interacting with each other can have a massive number of edge-cases. Self-driving systems have to interact with the real world, a smorgasbord of edge-cases: occluded signage, rain, snow, dirty cameras, construction, something falling off a truck, potholes, animal crossings, people in costumes, unpredictable human drivers, bicyclists, runners, scooters, skateboarders...

Some have asserted that the tail is so long that it will be impossible for an AI system to drive a car until AI has common sense and understands things like a person looking at their phone is not paying attention and that a person in a costume is still a person. Plus many situations at intersections are resolved with eye-contact and hand waves, how will an AI navigate this? These certainly are difficult problems, but that's what makes engineering interesting. They will be solved, without requiring an AI to be conscious, the only question is when.

When Will Tesla Achieve Level 5? 

To know when you're done with a project, you have to know the goal. Going through this, we've established some of the criteria:
  • Hit feature-complete, so the "real work" can begin
  • Better than a human driver (better than six-nines)
  • Able to handle novel situations safely
Using the methods we've outlined above, we need to need to know how long it took to get to feature-complete. Let's assume Musk is correct and feature-complete will happen in December of 2020. Now we need to know when AutoPilot development started. This is a more complicated question. Musk first mentioned Autopilot publicly in May of '13. Certainly, they had started working on it if it was discussed publically. Using this starting point, it would be 91 months to go from zero to feature complete. However, Tesla initially worked with Mobileye on the Autopilot 1 system. Autopilot 2 started shipping in October of 2016. This is when the sensor suite that's in production now was first seen. Using this starting point, it would be 51 months from start to feature-complete.

Andrej Karpathy became Tesla's director of artificial intelligence in June 2017. With Karpathy's arrival, the direction for Autopilot development was shifted greatly with more operations moved into a unified neural net backbone with multiple heads (dubbed the hydranet). Using Karpathy's arrival as the starting date would yield 42 months.

In April of 2019, Tesla released Hardware 3 and referred to the inference engine hardware as their FSD computer. This is a Tesla-designed custom system-on-a-chip (SoC) to run their neural network. Tesla claims that the new system was 21 times faster than their previous vendor-supplied solution. This is when Tesla said that they had the hardware platform that they required for FSD to be achieved. Based on this date, the time to feature-complete would be 20 months.

Now, which of these dates should we select as our start? I don't want to keep "moving the goalpost" and allow any significant event to be a restart point, yet I don't want to allow false starts or work by suppliers to count against the time either. Given these competing goals, I'm selecting Andrej Karpathy's start date as the legitimate beginning for the current direction for Tesla's FSD direction. (Let me know which date you'd select.)

Given Karpathy's start date and a possible feature-complete of December 2020, that's 42 months from start to feature-complete. So looking at the two models we have above, how long would it take to reach the goal of better than six-nines?

The 80/20 rule would need 9 iterations (8 more after December). That would be 8 * 42 months or 28 years to get to six-nines. By this model, the steering wheel could be deleted from the parts list in December of 2048. Let's look at the other model.

The empirical model gets to six-nines in only 5 iterations (4 more after December). That would be 4 * 42 months or 14 years before you could fall asleep and wake up safe and sound at your destination.

It's possible that we won't see self-driving cars until 2034, but let's use the more recent HW3 date. You could make a case that until this hardware was available, the AI was severely limited and this bottleneck hampered progress. Using this more optimistic date, it was 20 months from power-on to feature complete. And since we're going for the optimistic model, let's use the empirical model. Five iterations (4 more after December) would be 4 * 20 months or 6 years 8 months. That puts the 'sitting in New York and summon your car from LA' date as August of 2027.

Before you assume these models are accurate, let me assure you, they. are. not. These are rules-of-thumb based primarily on people debugging complex systems, not deep learning AI. They are based on projects that occur within a handful of years on a single generation of hardware. AI is still a nascent field, major breakthroughs are still occurring. Moore's Law yields periodic doubling of computing performance. In a mature technology, you don't see a 21x performance boost like Tesla's HW3 effort. And neural nets evolve on a non-linear "S"-shaped sigmoid function which means they can quickly go from incompetent to mastery.


The point of this long entry was to attempt to determine when we might see robo-taxis on the road. Toward this effort, we've generated estimates from 2027 to 2048. This ~20-year window seems large but if you're reading this, it means it is likely to occur within your lifetime. What I can guarantee is that new driver-assist features will continue to roll out and improve each year. And when self-driving cars happen, it will be a step-function in human history. Self-driving cars will join the list of humanity's greatest breakthroughs along with the wheel, electricity, and powered flight; it will save more lives than Penicillin, and yet it will be as taken for granted as quickly as the self-piloting elevator.

Disclosure: I'm long Tesla stock.

Monday, June 22, 2020

Elon's Estimates - Mistaking A Clear View For A Short Distance


Elon Musk is known for many things: Zip2, PayPal, SpaceX, Tesla, Boring Co...
But he is also known for his over-enthusiastic estimates of when a technology can be delivered. Other than Model Y, every one of Tesla's vehicles has been late to market. In December of 2015, Musk said that full self-driving would be available in 2 years; it has made progress, but it is still not here. And more recently on the Joe Rogan podcast, Musk predicted that within 5 to 10 years people will be able to directly communicate thoughts via brain implants rather than using the slow analog process of speech or writing.

For followers of Musk (fans and detractors alike), this is known as MST or Musk Standard Time. Converting from MST to a Gregorian calendar is not an easy task. It involves leap years and slide rules and it is not possible in all instances.

I don't point this out for ridicule; rather it is to ask the question: Why does Musk continue to make bold predictions on unrealistic timelines?

In short, I think that he is falling into the trap that Paul Saffo warned against:

       Never mistake a clear view for a short distance. ~Paul Saffo

Musk has a clear view of his plan. He's well aware that there will be challenges, but he has built teams and achieved many things that were deemed previously impossible. Create a door-to-door driving directions website - Check; Create an internet payment system - Check; Make sexy fast electric cars that blow away gas cars costing 10 times as much - Check; Create giant energy storage systems that change the way energy is bought and sold - Check; Land rockets on autonomous drones ships at sea - Check; Launch the largest network of low Earth orbit satellites that has ever existed to bring internet access to every square millimeter of the planet - Now underway and looking for Beta customers.

Class ½ Impossibilities

So it is not naiveté that brings Musk to these optimistic timelines. Rather, it's a series of successes. You look at the problem and ask yourself if engineering and innovation can achieve it, or would it require magic. If the answer is the former, then it can be accomplished. Cars will be self-driving, the only question is when. Humans will land on Mars, the question is will it be in this generation or another. When done, these will be incredible feats, but it will not be magic that brought them into existence. These accomplishments will be the product of hard-fought breakthroughs. If you have a vision, a roadmap, the ability to raise capital, the ability to attract great talent, and the ability to adapt based on feedback and learnings, the 'impossible' can be achieved. And maybe, just maybe, the people that accomplish it will be called sorcerers.

The Dunning-Kruger Effect is when someone has little skill or expertise in an area and assumes it will be easy for them. Their lack of knowledge gives them undeserved overconfidence. What Musk 'suffers' from is almost the opposite of this effect. He knows it will be Hell; it's just that he's been on the trail through Hell so many times that he could be a tour guide. And the one sure way to fail is to assume it's not achievable.

Musk has been on the trail through Hell so many times that he could be a tour guide.

What is the opposite of the Dunning-Kruger Effect? Would it be The Kruger-Dunning Effect or perhaps the Regurk-Gninnud Effect? 😃 Knowing something will be hard and doing it anyway is how great things are achieved.

The future will not be bound to a timeline. It is fickle and does not give up its secrets easily, but this should never stop the quest for a better tomorrow.

This post started off with a quote from Paul Saffo and I'll end it here with a quote from one of his contemporaries:

        "The best way to predict the future is to invent it." ~Alan Kay 

For more on Musk's Moonshot management style, check out this article.

Friday, June 12, 2020

Prius vs Model 3

The Toyota Prius was a landmark vehicle. At its introduction, it was the biggest advancement in car tech in decades. Worldwide sales of the Prius passed the 1 million milestone in May 2008, jumped the 2 million mark in September 2010, and reached 3 million in June 2013. It was selling well, it was a halo brand for Toyota and branched off many variants: Prius V, Prius C, Prius-Plug-In, and most recently Prius Prime.

Hybrid Technology Never Crossed the Chasm

Prius was the flagbearer hybrid brand in the industry. A hybrid vehicle from any manufacturer was compared to the industry benchmark, Prius. Toyota put hybrid tech into many of their other vehicles too including Lexus brands for a total of 44 different hybrid models sold around the globe.

As I write this in 2020, Toyota has sold over 15 million hybrid electric vehicles. Despite this success, hybrid vehicles have remained a niche product. Hybrid tech has a loyal following, but it has not crossed the chasm to become mainstream.

Will EVs Suffer The Same Fate?

This made me wonder if EVs would suffer the same fate of being relegated to a niche market. As one (far from conclusive) indicator, I decided to compare the sales of the flagship EV (Tesla Model 3) to sales of the flagship hybrid (Toyota Prius). 

Model 3 has been on sale for 11 quarters now, so we put the first 11 quarters of cumulative Prius sales next to the first 11 quarters of cumulative Model 3 sales. Here's that chart:


As you can see, during this time window, Model 3 is selling significantly better than Prius had. This does not guarantee that EVs will go mainstream, but it looks like the technology has a shot and, as we wrote here, and this could be the decade that it happens.

In the final quarter of Model 3 sales, Model Y was included in the date. I would have preferred to have just Model 3, but Tesla lumped Model Y and Model 3 sales together in their Q1 2020 report. Although, Model Y sales have just begun their production ramp, so their volume is not yet significant.

I thought it was important to make this comparison now, since I expect Q2 2020 numbers to be skewed by the pandemic (for Tesla, the rest of the auto industry, and most of the economy).

Will EVs go mainstream? Magic 8-Ball says 'Signs Point To Yes!'

https://en.wikipedia.org/wiki/Toyota_Prius#Sales
https://en.wikipedia.org/wiki/Tesla_Model_3#Deliveries

Thursday, June 4, 2020

2020s The Decade Of The EV


The decade* has not gotten off to a good start: a pandemic, giant killer hornets, racial strife, Ebola outbreak, Michigan dam breaches, Puerto Rico earthquakes, Australian bushfires, Cyclone Amphan, Cyclone Harold, Taal volcano eruption, Brazilian floods & mudslides...

Some of these disasters could leave an indelible mark on this decade; and while I hope that we learn our lessons from these tragedies and improve our society, that's a topic for another forum. This blog is about electric cars and EVs are sure to leave their mark on the 2020s.

Technologies frequently limp along for 10 or 20 years before the stars align and they suddenly become an "overnight success". This decade will be the one where EVs hit this overnight success tipping-point and become the norm. By the end of the decade, new car sales will be dominated by electric vehicles. When you are car shopping in 2029, considering a gas-powered car would be like considering a flip-phone in today's smartphone world.

Source: BloombergNEF
Why do I make this assertion?
  • First, EVs are more fun to drive (they are quieter, smoother, quicker) 
  • Gas prices are volatile and change with the whims of politics, saber-rattling, hurricane refinery outages... Electricity prices are far more stable and you can even generate it yourself from your own roof.
  • Battery prices have and will continue to drop. Batteries are the most expensive component in electric cars today and their price of manufacture has continued to drop. More battery factories are being built today than ever before in history.
  • EVs will be more affordable than gas cars by 2026. Today, if you consider fueling and maintenance, EVs are cheaper from the long term total cost of ownership perspective. However, for many people today, the initial sticker shock drives them away from an EV purchase. Following on the trend of battery costs, the sticker price for EVs will continue to drop. 
  • Charging speeds will increase. As battery tech improves, the causes of battery degradation will be mitigated and batteries will continue to toughen up and become tolerant to higher charging rates and more heat.
  • Ranges will increase. As battery tech improves, more energy will fit in the same space with less weight. This will be driven by both technology improvements and cost reductions.
  • Charging infrastructure will continue to proliferate. Unless you drive an EV, you are likely unaware of all of the charging infrastructure that already exists. Take a look at the map on plugshare.com, there are many places you can plug-in. And as more people start driving EVs, more infrastructure will be deployed at businesses that want to attract EV drivers and by utilities that want to sell electricity.
  • Electric fuel is cheaper. As I write this, gasoline prices are cheaper than they have been in decades. However, even at $1 per gallon, charging overnight at offpeak rates, I'm paying ~70% less per mile than a similar gas-powered car (25MPG @ $1 per gallon compared to $0.05 per kWh @ 4 miles per kWh). 
  • Update: @KennyBSAT pointed out that I forgot to mention the variety of vehicles that will become available during this decade with choices that can "carry more people or a bunch of stuff or tow, all while maintaining range." Good point!

Monday, May 25, 2020

What is Tesla's Project Dojo?


Tesla has made significant investments in artificial intelligence (AI). AI is the key to Tesla's full self-driving (FSD) future. Yet, Elon Musk has also called AI humanity's “biggest existential threat.” How do you reconcile this dichotomy? The answer is simple, Narrow AI vs General AI. A narrow AI is trained for a particular task such as playing a particular game or language processing. These narrow intelligences are not transferable. A narrow chess AI will not know anything about checkers despite the two games sharing a board. Whereas, a General AI (sometimes called Strong AI or Artificial General Intelligence(AGI)) is the hypothetical ability of a system to learn any intellectual task that a human could learn. Skills an AGI learned in one arena could be applied in new areas and an artificial superintelligence could quickly develop. An artificial superintelligence may find humans are irrelevant or worse, a threat. This is the “existential threat” that concerns Musk. 

So Tesla's FSD system will be a narrow AI, able to drive your car and you'll even be able to tell it where you'd like to go. You won't, however, be able to chat with the FSD AI about your day, but at least you'll know it won't decide that the best way to reduce traffic accidents is to kill all humans. 


Tesla's AI investments to date include creating an AI software development and validation team, creating a data labeling team, and creating an FSD hardware team to design their own custom neural network inference engine. Next on Tesla's AI investment list is "Project Dojo."


Project Dojo

We've been given a few hints about Dojo: Musk talked about it in the 2019 financial call and Tesla's Director of Artificial Intelligence and Autopilot, Andrej Karpathy, has talked about it at multiple AI conferences. We'll discuss how neural nets work and then move into some wild speculation; but first, we have to acknowledge the Dad Joke that is the name Project Dojo. We know that Project Dojo is intended to vastly improve the Autopilot Neural Network training. If you want to train, where do you go? A Dojo, of course. 



Before we get into Dojo we need to cover a few basics about neural networks. There are two fundamental phases to neural networks (NN): Training and Inference.

Training

NNs have to be trained. Training is a massive undertaking. This is when the digital ocean of data that is the training dataset must be digested. It takes terabytes of data and exaflops of compute to train a complex NN. Through training the NN forms "weights" for nodes. When the training is complete, the resulting NN is tested. A test dataset that was not part of the training dataset, where the expected results are known, is thrown at the resulting network and if the NN is properly trained, it infers the correct answer for each test. Since Project Dojo is all about training, we'll dig more into this later. Depending on the use case, there may be several stages of simulation and testing before the NN is deployed. Deploying the NN leads us to our next phase, Inference.

Inference

When a neural network receives input, it infers things about the input based on its training; this is known as “inference.” These inferences may or may not be correct. Compared to training, the storage and compute power needed for inference is significantly lower. However, in real-time applications, the inference needs to happen within milliseconds; whereas training can take hours, days, or weeks.

Unlike training, inference doesn't modify the neural network based on the results. So when the NN makes a mistake, it is important that these are captured and fed back to the training phase. This brings us to a third (optional) phase, Feedback.

Feedback

You may have heard the phrase "Data is the new Oil." Nowhere is this more applicable than AI training datasets. If you want an AI that performs well, you have to give it a training set that covers many examples of all of the types of situations that it may encounter. After you have deployed the AI, you have to collect the situations where it did the wrong thing, label it with the expected result, and add this (and perhaps hundreds or thousands of examples like it) to the training dataset. This allows the AI to iteratively improve. However, it means that your training dataset grows with each iteration and so does the amount of computing horsepower needed for training.


Tesla's Autopilot Flywheel 

Now that we've ever so briefly covered AI basics, let's look at how these apply to Tesla's FSD.

Let's start with Deploying the Neural Net. Every car that Tesla makes today is a connected car that receives over-the-air updates. This allows the cars to receive new software versions frequently. When a new version of Autopilot is deployed, Tesla collects data about its performance. The AI makes predictions such as the path of travel, where to stop, et cetera. If Autopilot is driving and you disengage it, this may be because it was doing something incorrectly. These disengagements are reported back to Tesla (assuming you have data sharing enabled). The report could be a small file that only has the data labels and a few details or it could be streams of sensor data and clips of video footage depending on the type of disengagement and the types of situations that Tesla is currently adding to their training set.

Even if Autopilot is not engaged, it is running in "shadow mode." In shadow mode, it is still making predictions and taking note when you, the human driver, don't follow those predictions. For example, if it predicts that the road bends to the left, but you go straight, this would be noted and potentially reported back to the mothership. If Autopilot infers that a traffic light is green but you stop, this data would again likely be noted and potentially reported back.

Tesla has about a million vehicles on the road today collectively driving about 15 billion miles each year. The bulk of these cars are from Tesla's Fremont factory. Tesla now has a second factory, Giga Shanghai, putting cars on the road. Soon Giga Berlin and Giga Austin (or will it be Tulsa?) will join them. All of this will result in a large amount of data for the training dataset.

The bigger the training set, the longer it takes to process. However, with a system like this, the best way to improve it is to quickly iterate (deploy it, collect errors, improve, repeat). If training takes months, this slows down the flywheel. How do you resolve this? With a supercomputer dedicated to AI training. This is Project Dojo: make a training system that can drink in the oceans of data and produce a trained NN in days instead of months.


A Cerebras Wafer Scale Engine

Cerebras

At the start, I promised some speculation. As promised, here it is.

The size of the chips used for AI training has been increasing every year. From 2013 to 2019, AI chips increased by about 50% in size. A startup called Cerebras saw this trend and extrapolated it to its natural conclusion of 1 chip per wafer. For comparison, the Cerebras chip is 56 times bigger than the largest GPU made in 2019, it has 3,000 times more on-chip memory, and it has more than 10,000 times the memory bandwidth.

This wafer-scale chip is an AI training accelerator and my conjecture is that a Cerebras chip will be at the heart of Project Dojo. This wafer-scale chip is the biggest (literally and figuratively) breakthrough in AI chip design in a long time.

There is one (albeit tenuous) thread that connects Tesla and Cerebras, both are part of ARK Invest's disruption portfolio. ARK has investments in both companies and meets with their management teams. When there are two companies that could mutually benefit working together and it would benefit their mutual investor, ARK, you can bet that introductions would be made.