A Cautionary Tale of iRobot and the Future of Tesla

Camera-First Autonomy

The journey toward widespread autonomy, whether in the confined space of a living room or the complex environment of public roads, has hinged on a fundamental debate regarding sensory input. Both iRobot, the creator of the Roomba robot vacuum, and Tesla, the leading manufacturer of EVs, have adopted a strong stance: autonomous navigation should be primarily or exclusively camera-based. This camera data will be processed by sophisticated visual processing rather than relying on auxiliary sensors like LiDAR (Light Detection and Ranging). Although this 'vision-first' approach is philosophically elegant, iRobot’s recent struggles present a cautionary tale for Tesla. Will Tesla suffer a similar fate by rejecting redundant sensor modalities?

The iRobot Misstep: vSLAM Versus LiDAR

iRobot’s early dominance in the robot vacuum market was built on its ingenuity, but its subsequent loss of market share is deeply rooted in its adherence to a vision-only system known as Visual Simultaneous Localization and Mapping, or vSLAM. While vSLAM utilizes an upward-facing camera to track the robot's movement relative to ceiling features and visual landmarks, it faced significant practical limitations. A key problem was its reliance on light; in dimly lit environments, such as under furniture or in dark hallways, the camera’s ability to map and localize the robot was severely compromised, leading to inefficient, frustrating cleaning patterns.

As competition intensified, rivals introduced vacuums utilizing LiDAR. LiDAR sensors rapidly send out laser pulses and measure the time-of-flight to create incredibly precise, three-dimensional maps of the environment. Because LiDAR actively generates its own light source, it operates perfectly regardless of ambient light conditions, offering superior speed, accuracy, and reliability in mapping and path planning. By delaying the adoption of LiDAR in favor of improving its vSLAM system, iRobot allowed competitors, such as Roborock, Ecovacs, and Dreame, to surpass it in core functionality, offering devices that mapped homes faster, cleaned more efficiently, and rarely got stuck. This reluctance to adopt superior sensing technology, despite its obvious practical benefits, was a primary factor that eroded iRobot’s competitive edge, leading to a massive decline in its market share and financial viability.

The Greater Complexity and Higher Stakes of the Road

The environment a car navigates is exponentially more complex than a living room, suggesting that Tesla’s reliance on cameras faces far greater engineering hurdles than iRobot's ever did. A robot vacuum operates on a flat, two-dimensional plane in a structured, relatively static environment; objects are typically furniture, dropped items, or stationary pets. An autonomous vehicle, however, operates in a highly dynamic, three-dimensional world that includes rapid movement, unpredictable human drivers, pedestrians, cyclists, changing traffic signals, highly variable weather conditions, and a full spectrum of lighting situations, from blinding sun to pitch black.

To address this complexity, Tesla’s approach is fundamentally different from iRobot’s. The company utilizes a much larger array of cameras (typically eight), providing a full 360-degree view around the vehicle. More importantly, the artificial intelligence architecture employed by Tesla is vastly more sophisticated. Tesla trains its neural networks on petabytes of real-world driving data, allowing its system to perform complex tasks like object recognition, behavior prediction, and depth estimation based purely on visual cues. The company’s argument is that vision, once adequately processed by powerful, deep-learning algorithms, is the only sensor modality necessary to achieve better than human-level driving capability, as humans themselves navigate primarily through sight.

A Cautionary Tale, Not a Harbinger of Doom

While iRobot's failure to incorporate LiDAR led to its downfall in the hyper-competitive consumer electronics market, this narrative should be seen as a cautionary tale for Tesla, not a prediction of the company's ultimate failure. The core lesson is the danger of relying on a single sensor when a superior, practical, and complementary technology exists. iRobot's vSLAM system was fundamentally weak in low light, a flaw that LiDAR easily mitigated, and cars mitigate via headlights.

Tesla’s challenge is different. Its vision-only strategy, dubbed "Tesla Vision," aims for driving generalization: a system that can see and understand the roadways as well as or better than a human. This approach carries immense risk, especially when competitors like Waymo and Cruise employ multi-sensor stacks that combine cameras with high-resolution LiDAR and radar, providing redundancy and immediate, precise depth information is crucial for safety. As we've pointed out previously, sensor suites have their own risks. However, the sheer computational power and massive scale of the data used by Tesla's AI provide a sophistication that far exceeds the relatively simple algorithms used in robot vacuums. For the moment, Tesla’s bet remains contentious, relying on future AI advancements to overcome the current limitations of vision in edge cases, inclement weather, and sudden low-visibility events. Therefore, iRobot serves as a grim reminder that technological obstinacy can be lethal, but the complexity and cutting-edge AI of automotive autonomy means Tesla is playing a completely different, higher-stakes game.

Carbon Capture Is Sabotaging AI’s Future

 

The Case Against Carbon Capture: Prioritizing Renewables for AI Energy Demands

The rapid growth of AI datacenters has created an urgent need for reliable, scalable, and sustainable energy sources. As electricity demand surges, driven by AI's computational requirements, the debate over how to meet this demand while addressing climate goals has intensified. Carbon capture technologies, such as carbon capture and storage (CCS) and carbon capture, utilization, and storage (CCUS), are often touted as a solution to CO2 emissions from fossil fuel plants. However, evidence suggests that carbon capture is an inefficient and costly approach compared to investing in renewable energy sources like solar, wind, and energy storage, which avoids CO2 generation. This analysis argues that prioritizing renewables over carbon capture is critical for meeting AI datacenter energy needs, reducing CO2 emissions, and maintaining global leadership in AI innovation.

AI Datacenter Energy Demands

AI datacenters are a significant driver of global electricity consumption. According to the International Energy Agency (IEA), datacenters consumed 415 terawatt-hours (TWh) in 2024, representing 1.5% of global electricity use. By 2030, this demand is projected to more than double to 945 TWh, equivalent to the annual electricity consumption of a country like Japan. In the United States, datacenters are expected to account for nearly half of the growth in electricity demand over the next five years. AI-optimized datacenters, with their high computational requirements, are the primary contributors to this surge. Meeting this demand requires energy sources that can be deployed quickly, scaled efficiently, and aligned with global climate goals to reduce greenhouse gas emissions.

Limitations of Carbon Capture

Carbon capture technologies aim to (as the name says) capture CO2 emissions from fossil fuel plants or industrial processes, either storing it underground or repurposing it for other uses. However, research highlights significant drawbacks that undermine their effectiveness and economic viability.

A 2019 Stanford University study found that carbon capture systems capture only 10 to 11% of total CO2 equivalent emissions over a 20-year period when accounting for upstream emissions from energy production and equipment manufacturing. The efficiency of these systems is often overstated, with real-world performance, such as a coal plant achieving 55.4% efficiency over six months, falling far below the projected 85 to 90%. Moreover, due to the energy demand of carbon capture system operations, the net effect can be increased air pollution; this is called the energy penalty. The air pollution includes particulate matter and nitrogen oxides, leading to higher social costs, including health impacts, economic losses, and climate damages. In these situations, it would be better to operate fossil fuel plants without capture than with it. However, the even better solution would be transitioning to renewables with energy storage.

CCS cost is another critical barrier. The Stanford study, as reported by Environment America in February 2025, concluded that investing in carbon capture is 9 to 12 times more expensive than switching to 100% renewable energy when considering energy costs, health impacts, and emissions. The IEA notes that while CCUS may be cost-competitive in hard-to-abate sectors like cement and steel production, it is not a scalable solution for widespread emissions reduction compared to renewables. Diverting funds to carbon capture reduces resources available for more effective solutions, such as solar, wind, and energy storage, which offer greater emissions reductions and faster deployment.

The Case for Solar and Wind with Energy Storage

Solar and wind power are the fastest-growing and most cost-effective energy sources available, making them ideal for meeting AI datacenter energy demands. According to Carbon Brief, solar and wind are the fastest-growing electricity sources in history, with solar power generation increasing more than eightfold and wind power more than doubling in the US over the past decade, as reported by Climate Central in April 2024. In 2023, solar added more new capacity globally than coal, increasing its share of global electricity from 4.6% to 5.5%, while wind held steady at 7.8%.

The cost of renewables has plummeted, with solar and wind costs dropping by 85% and 55%, respectively, between 2010 and 2020, according to Ember. These energy sources can be deployed more quickly than fossil fuel or nuclear plants, which require longer lead times and higher upfront investments. To address the intermittency of solar and wind, industrial-level energy storage systems, such as lithium-ion batteries, pumped hydro, and compressed air storage, provide reliable power delivery. The IEA emphasizes that grid-scale battery storage is critical for integrating renewables into the grid to meet future energy needs and net-zero emissions goals by 2050.

Impact on CO2 Emissions and Fossil Fuel Dependence

Investing in carbon capture risks extending the operational life of fossil fuel plants, which undermines efforts to reduce CO2 emissions. A September 2023 Earthjustice report highlighted concerns that carbon capture is used by the fossil fuel industry to justify the continued operation of polluting facilities, perpetuating reliance on coal and gas. A February 2024 article noted community opposition in Louisiana, where carbon capture projects were seen as prolonging the life of dirty power plants. By contrast, prioritizing renewables allows for the phased retirement of fossil fuel plants, directly reducing CO2 emissions and aligning with climate goals.

The International Institute for Sustainable Development (IISD) reported in November 2023 that some CCS projects produce more emissions than they sequester, particularly when powered by fossil fuels. Redirecting investments to renewables avoids this inefficiency, as solar and wind generate electricity with near-zero emissions during operation. By scaling renewables and storage, countries can meet AI energy demands while accelerating the transition away from fossil fuels, reducing overall greenhouse gas emissions.

Implications for AI Leadership

The ability to meet AI datacenter energy demands is critical for maintaining global leadership in AI innovation. Delays in deploying sufficient clean energy could hinder AI development, as datacenters require consistent and affordable electricity. Solar and wind, supported by energy storage, offer the fastest path to scaling energy capacity, ensuring that AI infrastructure can expand without reliance on fossil fuels. Diverting funds to carbon capture, which is less effective and more costly, risks slowing this progress, potentially ceding technological advantages to countries that prioritize the fast growth that renewables offer.

Conclusion

The evidence suggests that carbon capture is an inefficient and costly approach compared to investing in solar, wind power, and energy storage to meet the growing energy demands of AI datacenters. Carbon capture's low efficiency, high costs, and potential to extend fossil fuel plant lifespans make it a less viable solution for reducing CO2 emissions. In contrast, renewables offer a scalable, cost-effective, and environmentally sustainable path to powering AI infrastructure while supporting climate goals. By prioritizing solar, wind, and energy storage, countries can meet AI energy needs, reduce emissions, and maintain leadership in the global AI race. As of 2025, redirecting resources from carbon capture to renewables is the most strategic approach to achieving these objectives.

An Open Letter to Elon Musk: The Mandate of Fiduciary Duty

Dear Mr. Musk,

As an investor who purchased shares in July of 2010, I extend my congratulations to you and the team on the overwhelming shareholder approval of your 2025 compensation package. This validation of your vision and the resulting, truly incredible stock performance over the past decade and a half has been a great success for all long-term investors.

With the approval of this landmark award secured, the terms of engagement must immediately become more rigorous. The successful passage of this compensation plan demands that we immediately address the assurance, which is the solemn promise the Board publicly cited, that your political involvement must promptly wind down. This commitment, while perhaps structurally absent from the financial text, is now a peremptory corporate governance requirement, a non-negotiable condition for maintaining shareholder trust in the Board’s oversight.

Your fiduciary duty, as the chief executive officer, is to act solely in the sustained best interest of Tesla and all its shareholders. The company’s brand and your public persona are inextricably linked; you are synonymous with Tesla. Consequently, your aggressive social media activity, specifically the inflammatory posts supporting partisan political movements, transcends the realm of personal opinion and becomes a matter of corporate liability.

This behavior is actively counterproductive to Tesla’s mission and constitutes a clear violation of the assurances provided to shareholders. Such polarization does not merely risk alienating significant segments of the global customer base; it actively undermines the essential public trust that fuels our global, consumer-facing technology enterprise. These posts are a direct breach of the expectation that was an essential component of the compensation package’s approval narrative.

To be clear, we recognize and respect your rights to self-expression under the First Amendment. However, in your capacity as the leader of the world’s most valuable automotive and emerging AI company, these rights are balanced against an equally weighty set of responsibilities. We demand that the commitment made to the Board and all shareholders be fulfilled immediately. Your focus must now be singularly and unparalleledly dedicated to achieving the ambitious operational milestones that justify this unprecedented award. Compliance with the spirit of the assurance is not optional; it is the unwavering obligation required to earn the value that shareholders have now overwhelmingly validated.

Sincerely,

A Long-Term Tesla Shareholder

Tesla Model Y Winter Range

Image by OpenAI

Update: This was originally posted on April 13th, 2025. I've reposted it now since winter weather is fast approaching here in the northern hemisphere.

How far can a Tesla drive in the winter? 

We've had our Model Y for two winters, and I've been tracking its driving efficiency year-round since we got it. I've heard it said that EVs have less range in the winter, and I wanted to see what our real-world data shows. 

Winter driving is generally less efficient. The cold air is thicker, energy is used to warm the cabin, and the extra traction of winter tires comes with a higher rolling resistance.

We get some snow here each winter, and we like to go to the mountain occasionally, so we have a set of winter tires. Tires can have a big impact on range, so here's the tire data and other relevant specs: 

Vehicle: 2023 Tesla Model Y Long Range All-Wheel Drive (AWD), an all-electric, mid-size crossover SUV.

All-season Tires: Continental ProContact RX 255/45 R19 104W XL. These tires came on our Y when we picked it up, and we drive on them from mid-March through October.

All-season Wheels: 19" Gemini Wheels, standard on the 2023 Tesla Model Y Long Range AWD, dark grey with plastic covers for improved efficiency. 

Winter Tires: Michelin X-Ice Snow 255 /45 R19 104H XL BSW. These are premium winter tires designed for EVs, crossovers, and SUVs. They're studless and rely on tread and compound for grip, rather than metal studs. They're built to handle severe winter conditions like heavy snow, icy roads, and sub-zero temperatures, while still providing decent dry-road performance. This model is a step up from its predecessor, the X-Ice Xi3, with better hydroplaning resistance and snow traction. We use these tires from November through mid-March.

Winter Wheels: Vision Cross II 19" X8 5-114.30 38 BKMTXX
These are cast aluminum wheels. They are slightly heavier than the OEM Gemini wheels and don't have aero covers. However, the matte black goes well with the modern chrome delete look of Tesla vehicles.

2023 Model Y LR	All-Season	Winter
Wheels	Gemini	Vision Cross II
Tires	Continental ProContact RX	Michelin X-Ice Snow
Miles Driven	7,457	6,182
Ave Temp	71F	50F
Wh/mile	254	278
Rated Range	330 miles	301 miles
Measured Real Range	302 miles	276 miles
Winter Penality		8.6%

The winter range loss is lower than I expected. The tires are not as efficient, the wheels are not as light, there are no aero covers, the heater has to run... and yet, there's only a ~9% range penalty. I was expecting it to be about 30%. 

One of the reasons this winter penalty is low is that Tesla vehicles have a very efficient heat pump. The heat pump includes a liquid-cooled condenser loop and an 8-way octovalve. This enables 3 cooling modes and 12 heating modes, including special modes for temperatures below -10°C. The heat pump uses a compressor that operates on 400 volts and draws power directly from the traction pack.

It's important to note that this data is for my NW Oregon driving. If the seasonal temps are different in your region, the results could be dramatically different.

Referral

If you're interested in a Tesla vehicle or solar, you can use my referral code and we'll both get perks (https://ts.la/patrick7819)

Tesla's TeraFab Gambit: Bluff or Breakthrough?

Tesla's Semiconductor Leap: Bold Strategy or Calculated Bluff?

Introduction

Elon Musk has a knack for turning corporate announcements into global spectacles, and his recent comments at Tesla's annual shareholder meeting on November 6, 2025, were no exception. There, he outlined plans for a massive Tesla chip fabrication plant, dubbed the "TeraFab," to fuel Tesla's AI ambitions. As Tesla eyes billions of Optimus robots and widespread robotaxis, they'll need an unwavering chip supply. The question arises: is this a genuine push into the treacherous world of semiconductor fabrication, or a clever bluff to prod suppliers like TSMC and Samsung into action? In this post, we explore the drivers, challenges, and stakes of Tesla's gambit. Vertical integration here could streamline innovation, but it demands careful navigation of technical and ecological hurdles.

The Surging Demand for Custom Silicon

Tesla's AI hunger is voracious. The company projects needing millions of specialized chips annually to train models for autonomous vehicles and humanoid robots. Musk emphasized that current suppliers cannot meet this scale without compromising other clients, like Apple or Nvidia. A single Terafab, he suggested, would start with 100,000 wafer starts per month and expand to 10 facilities, each churning out enough silicon to power a robot army.

This urgency stems from Tesla's robotics roadmap. Optimus, the company's humanoid bot, is tentatively planned to start rolling off production lines in late 2026. Each unit requires efficient inference chips for real-time decisions, while data centers demand AI training hardware. Without in-house control, Tesla risks delays akin to the 2021 chip shortage that slashed EV output by 30%. By building its own fabs, Tesla aims to secure supply and customize processes, much like it did with batteries. The question remains: is Musk's rhetoric a strategic pressure tactic, designed to extract better terms or (more likely) higher volumes from their foundry partners, or something completely different?

Navigating the Fab Frontier

Semiconductor fabrication is no casual undertaking. It involves etching circuits smaller than viruses in dust-free environments, with upfront costs exceeding $10 billion USD per plant. Construction typically spans three to five years, and yields can plummet from contamination or process flaws. There are also tremendous ongoing costs, as new process nodes must be introduced every 2 to 3 years to stay on the cutting edge. This requires new lithography equipment costing billions in the quest for smaller and smaller transistors.

Apple and Google, titans of tech, remain fabless. They pour resources into design and architecture, outsourcing production to the likes of TSMC. This model avoids the capital sinkholes and talent wars that plague foundries.

At its peak in the early 2000s, there were 22 companies with their own chip foundries. Today, that number has shrunk dramatically to just 3 major foundries (Intel, Samsung, and TSMC), and all but Intel primarily service the fabless chip design companies. 

During the shareholder meeting, Musk floated the idea of a partnership with Intel. Tesla could leverage Intel's US-based expertise and underutilized fabs as an on-ramp to their effort.

Terafab: Bluff or Breakthrough?

There's a more skeptical view of what motivated Musk's Terafab statements. This skeptic angle is that Musk is bluffing. By invoking a "gigantic" Terafab, Musk is hoping to spur TSMC and Samsung to allocate more capacity, echoing his past supply-chain arm-twists. TSMC's latest earnings hinted at reserved slots for Tesla, but no blockbuster deals have surfaced since the meeting. If real, this Terafab venture would mark Tesla's deepest vertical plunge yet, blending automotive grit with silicon precision.

The Third (and Most Likely) Option

So far, we've only examined this as either Tesla making their own fully owned and operated fab or Musk bluffing to gain more capacity from vendors.

The third, and perhaps most likely, path is a partnership similar to the battery cell partnership with Panasonic. Tesla built a dedicated space for Panasonic in GigaNavada to build cells. This partnership works well for Panasonic because it allows them to build cells using their proprietary technology and gives them an on-hand customer for the cells. Additionally, this works out for Tesla because they have a dedicated supply of high-quality cells. 

If Tesla strikes a similar deal with a major chip fabricator for chips, it could work out for both of them. Let's say the deal is structured similarly to the Panasonic deal. Tesla would buy the land, build the structure, and pay for a portion of the equipment costs (via Non-recurring Engineering or NRE payments). In return, all the production capacity of the plant would be dedicated to Tesla. If Tesla didn't need all of the capacity, the IDM would be able to use the surplus capacity for other customers. Because of the equipment and operating costs, it's very important to keep chip fab utilization near full capacity. 

The Dojo Pivot: Lessons in Adaptation

Tesla's chip strategy evolved rapidly this year. This Terafab announcement comes amid a pivot toward next-generation AI5 chips replacing Dojo in Tesla's training cluster. In August 2025, the company disbanded its Dojo team, scrapping the custom supercomputer Musk once hailed as a training powerhouse. He called Dojo an "evolutionary dead end," too niche and costly to scale against Nvidia's GPUs. Dojo resources shifted to AI5 and AI6, versatile chips optimized for both inference and training. These successors build on Dojo's matrix-math innovations but generalize for broader use, with AI5 production slated for 2026.

This pivot underscores Tesla's agility. Dojo's D1 chip, with its wafer-scale design, taught valuable lessons in parallel processing, now infused into AI6's architecture. Musk noted that clustering dozens of these on a board could mimic Dojo's scale, slashing cabling costs by orders of magnitude. The move conserves talent and capital, focusing on chips that power Optimus's dexterity or FSD's navigation without bespoke hardware traps.

Aspect	Dojo (Pre-2025)	AI5/AI6 (Post-Pivot)
Primary Focus	Custom AI training supercomputer	Versatile inference and training
Architecture	Wafer-scale D1 chips	Generalized SoCs, Nvidia-compatible
Production Partners	In-house prototypes	Samsung, TSMC (2026 ramp-up)
Scalability Challenge	High cost, slow iteration	Modular boards for clusters
Projected Output	Limited to prototypes	Millions of units annually (2027)

This table highlights the shift's efficiency gains, positioning Tesla for sustainable growth.

Conclusion

Tesla's flirtation with a Terafab embodies Musk's high-stakes vision: to control the stack and accelerate humanity's autonomous future. Whether it is a bluff or a blueprint, it pressures the industry toward faster scaling. The Dojo cancellation proves Tesla can pivot, channeling setbacks into smarter path selection. And the Panasonic partnership may foreshadow the Terafab plan. As 2026 approaches, watch for groundbreakings or sweetened supplier pacts. In Musk's world, bold bets often pay off, nudging us all toward a more sustainable horizon.

Solar-Powered Heat Pump vs. Gas Furnaces Showdown - Heat Your Home for Pennies!

RUUD Heat Pump and Air Handler

Our furnace and air conditioner are both 30 years old. They are the original equipment installed when the house was built; winter is coming and it's time to replace both of them. The lifespan of equipment like this is generally 15 to 20 years. Ours have exceeded the typical range significantly, but their age is showing, and the annual repair costs are now real.

Since they both need to go, we're considering a heat pump to replace them. Our 4-bedroom home uses methane (natural gas) for the furnace, cooktop, water heater, and (rarely used) fireplace. Soon after we moved in, our water heater needed to be replaced. This was over 20 years ago, so heat pump water heaters were not a viable option yet, so we installed a tankless water heater. It still used methane, but now it's not heating water 24/7, just in case one of us turns on a tap.

Similarly, with this furnace upgrade, I want to reduce our methane use, but it does not have to go to zero since gas is used in other parts of the home. If I were building a new home, it would certainly be all-electric, but this is a retrofit, and I'm happy with steps to reduce fossil fuel usage. Don't let the perfect be the enemy of the good, and all that; but let's see where the costs land.

When replacing your AC and furnace, there are a lot of options to consider. For cooling, if we have an AC unit or a heat pump, the energy usage would be similar, and electric is the only "fuel" option to run it, so let's call that a wash and look into the more complex side, heating. Heating has a lot of options. We could continue to use a furnace (upgrading to a new, more efficient unit), we could use a low-temperature heat pump (getting rid of the furnace completely), or we could do something in-between with a hybrid system that uses a standard heat pump as the primary heating source and a high efficiency furnace to cover the few subfreezing days and nights we have here.

Background and Assumptions

This analysis will compare home heating options for a 4-bedroom house in a Portland, Oregon, westside suburb. Your mileage may vary depending on your location, utility costs, home size, and factors like thermostat settings and insulation levels. These estimates consider the region's mild climate and current energy prices.

Our home is in a temperate climate with winter lows averaging around 34°F. The area is in USDA Climate Zone 8b. This zone is characterized by average annual minimum winter temperatures that do not go below 15°F. We have wet mild winters and warm dry summers, typical of the Pacific Northwest. It also aligns with ASHRAE Climate Zone 4C (cold, humid, marine), which is used for building energy standards, indicating cool winters with significant (2,500–3,000) annual heating degree-days (HDD) and with moderate cooling needs (unless there's a heat dome).

We have insulation typical of a 1990s build, requires an estimated heating load of 50 million BTU annually. Methane prices are set at $1.60 per 100 cubic feet. Regional electricity average of 13 cents per kWh. Methane contains about 1,030 BTU per cubic foot, and we use HSPF and AFUE ratings to convert heating demand into fuel and/or electricity needs.

The hybrid system uses the gas furnace only when temperatures drop to freezing or below. The heat pump will cover many more days per year of heating than the furnace, but the furnace will cover the coldest days (and nights) of the year. This pencils out to the furnace covering about 20% of the heating load, with the heat pump handling the remaining 80%.

Heating Options Overview

• Old Furnace: A 1994 Carrier gas furnace (model 58RAV115-16) with 80% Annual Fuel Utilization Efficiency (AFUE). As covered at the beginning, this is not an option to continue using, but it is included as a baseline.
• New Furnace: A RUUD R962V Endeavor Line Achiever Plus Series Gas Furnace with 96% AFUE.
• Hybrid (Dual Fuel): Combines a RUUD Heat Pump (4 Ton RD17AZ48AJ3NA, ~9.5 HSPF) with the RUUD R962V furnace, using the furnace below freezing.
• Cold Climate Heat Pump: An extended capacity heat pump (10 HSPF, ~3.5 COP at 47°F, ~2.5 COP at 17°F) with no furnace, relying entirely on electricity. May include resistive heating (electric) as a backup source. 

* For completeness, the calculations for each option are included at the end of the article. 

Comparison Table

HEATING OPTIONS COMPARISON

Home in the Greater Portland, Oregon Region

System	Methane Use (feet³)	Electricity Use (kWh/year)	Total Annual Cost (USD)
Old Furnace	603,000	700	$1,056
New Furnace	502,000	600	$881
Hybrid Heat Pump  (Dual Fuel)	100,400	1,434	$346
Cold Climate Heat Pump	0	4,884	$635

carswithcords.net

Key Considerations

The hybrid system is the most cost-effective at $346 annually, leveraging the heat pump’s efficiency in the region's usually mild climate and minimal furnace use.

I admit the "Heat Your Home for Pennies!" portion of the title is clickbaitish, but when I saw that the result was less than $365 annually, that's less than a dollar per day! And I wanted to stress that point. 

The cold climate heat pump, at $635, eliminates gas usage but increases electricity costs due to full electric heating. It takes more work to extract heat from cold air, but this is still a money-saving option compared to either furnace. The new furnace saves more than $150 per year over the old furnace due to higher efficiency, but this option would have the highest carbon footprint, and if we're replacing the AC unit anyway, there's no reason not to put in a heat pump while there are still incentives to do so.

The dual-fuel system gives us energy pricing resilience. It allows us to change the heat pump to furnace switch-over temperature. For example, if electricity rates climb significantly and gas does not, then we could adjust the switch-over point from 32°F to 34 or 35°F. This would use less electricity and more methane for heating during the coldest part of winter (but also increase our carbon footprint).

Even Better With Solar

As regular readers will know, we have solar panels and batteries on our home. Heat pumps, which run on electricity, pair exceptionally well with solar PV systems because they can utilize the clean, renewable energy generated on-site. The batteries allow us to time-shift our solar energy to avoid peak demand electricity rates. This means that when we are using the grid, we're buying energy at the cheapest off-peak rate. This will mean that our heat pump will be running directly from solar, from stored solar, or from off-peak grid energy. This will keep our heat pump running costs low. This heat pump / solar / storage synergy reduces strain on the grid and lowers CO2 emissions by displacing fossil fuel-based energy, especially in regions with coal and/or gas-heavy grids. This trio also helps mitigate HVAC cost volatility; generating your own power insulates you from fluctuating utility rates. If we were only using grid electricity for a new heat pump, I might regret installing it in a year or two if local electricity rates shot up. With solar, we have a guaranteed fixed cost.

When a heat pump is used for heating instead of a fossil gas furnace, renewable energy can directly displace the burning of fossil fuels.

Conclusion

Each heating system presents a different balance of cost, efficiency, and infrastructure requirements. Here is a quick summary:

• Old Furnace: High gas usage and cost, high maintenance costs
• New Furnace: Efficient, but still relies entirely on fossil fuels
• Hybrid System: Excellent performance in mild climates, lower carbon footprint than above. Flexible fuel choice.
• Cold Climate Heat Pump: All-electric, no gas needed, best for decarbonization, higher upfront cost and slightly higher running cost than hybrid

For our home, the hybrid system offers the lowest annual operating cost at $346, followed by the cold climate heat pump at $635. The new furnace ($933) and old furnace ($1,095) are less economical. Heat pumps provide environmental benefits, making them a forward-thinking choice for sustainable heating. I'm placing my order for the hybrid system now. Expect to see an install post coming soon.

Sources: NW Natural, Ruud Products, EIA Degree Days

Option	Annual Energy Usage	Annual Operating Cost
Old Furnace	603,000 feet³ of gas, 700 kWh of electricity	$1,056
New Furnace	502,000 feet³, 600 kWh	$881
Hybrid (Dual Fuel)	100,400 feet³, 1,434 kWh	$346
Cold Climate Heat Pump	0 feet³, 4,884 kWh	$635

* Detailed Calculations

Old Furnace (Carrier 58RAV115-16)

With 80% AFUE, this furnace converts 80% of fuel energy into heat. The annual heating load of 50 million BTU requires an input of 50,000,000 / 0.8 = 62,500,000 BTU. Gas usage is 62,500,000 / 103,675 ≈ 603 CCF or 17,070 cubic meters. Electricity usage for the blower motor is estimated at 700 kWh annually. Operating costs include gas (603 CCF × $1.60 = $964.80) and electricity (700 kWh × $0.13 = $91), totaling approximately $1,095.

New Furnace (RUUD R962V)

The newer RUUD furnace offers a notable improvement in fuel efficiency, reducing gas consumption by over 100,000 cubic feet annually compared to the older unit. This results in yearly fuel savings. The electric blower fan and control systems are slightly more efficient, lowering electricity use as well. This option balances simplicity with better energy performance. The 96% AFUE furnace is more efficient, requiring 50,000,000 / 0.96 ≈ 52,083,333 BTU input. Gas usage is 52,083,333 / 103,675 ≈ 502 CCF, or  ≈14,215 cubic meters. Electricity usage remains at 600 kWh for the blower. Costs include gas (502 CCF × $1.60 = $803.20) and electricity ($78), totaling approximately $881.

Hybrid (Dual Fuel) System

The heat pump covers 80% of the load (40 million BTU) with a 9.5 HSPF (~3.3 COP). Electricity usage is 40,000,000 / (9.5 × 3,412) ≈ 1,234 kWh. The furnace handles 20% of the load (10 million BTU) at 96% AFUE, requiring 10,000,000 / 0.96 ≈ 10,416,667 BTU, or 10,416,667 / 103,675 ≈ 100 CCF, or ≈ 2,832 cubic meters. Total electricity includes 1,234 kWh (heat pump) plus 200 kWh (furnace blower) = 1,434 kWh. Costs are gas (100 CCF × $1.60 = $160) and electricity (1,434 kWh × $0.13 = $186.42), totaling approximately $346.

Cold Climate Heat Pump

With no furnace, this system uses a heat pump with 10 HSPF (~3.0 COP average). The full 50 million BTU load requires 50,000,000 / (3.0 × 3,412) ≈ 4,884 kWh. No gas is used. The operating cost is 4,884 kWh × $0.13 = $635. Depending on electric resistive heating backup usage, this annual electricity usage and cost could be even higher.

Comparison Table

Option	Energy Usage	Annual Operating Cost
Old Furnace	621,090,000 BTUs, 700 kWh	$1,056
New Furnace	517,060,000 BTUs, 600 kWh	$881
Hybrid (Dual Fuel)	103,000,000 BTUs, 1,434 kWh	$346
Cold Climate Heat Pump	0 BTUs, 4,884 kWh	$635

Ω

HW3's Legacy: A Financial and Logistical Analysis of Tesla's FSD Obligation

Calculating Tesla's HW3 Autonomous Obligation

Tesla started selling cars with Full Self-Driving (FSD) computers and cameras, known as Hardware 3 (HW3), in 2019. This was the standard for all of their vehicles through 2022 until HW4 (now known as AI4) supplanted them. During the HW3 window of time, Tesla made approximately 3 million vehicles that were sold as fully FSD-ready, "all that's needed is an over-the-air update, and these cars will be able to drive themselves." Now that we're on the cusp of autonomy, are these HW3 cars an albatross around Tesla's neck or an opportunity? 

At a mere 144 TOPs, it's becoming apparent that HW3 likely does not have the compute horsepower or camera clarity necessary to achieve unsupervised FSD. So what's Tesla going to do if they achieve FSD (on AI4 or 5) but HW3 proves to be insufficient? Will they retrofit the HW3 vehicles? If so, how many cars will need to be upgraded, and how much will it cost Tesla (or you)? 

Introduction

The promise of FSD is a monumental technological endeavor that goes far beyond software alone; it is inherently tied to the hardware underpinning millions of Tesla vehicles already on the road. After working with Mobileye (Autopilot 1) and Nvidia (HW2/2.5), Tesla designed their own custom chip for HW3. This marked a critical milestone in their journey. HW3 was built into all Tesla vehicles for nearly 3 years. The HW3 chip was championed as the final hardware piece necessary for full autonomy. Now, with the vast improvements in subsequent hardware generations (HW4, AI5), some owners are wondering if HW3 will be able to cross the unsupervised finish line or if it will be like trying to run Borderlands 4 on a Commodore 64.

Today, we'll examine the size of the HW3 fleet, its FSD adoption rate, the retrofit costs, and how Tesla may deal with upgrades. 

The HW3 Fleet: Size and Timelines

Tesla began installing HW3 in new vehicles starting in April 2019 (replacing the Nvidia-based HW2.5) and continued to use HW3 in primary production until the shift to HW4 in early 2023. This production window generated a large population of vehicles. Based on cumulative delivery data, there are more than 3 million vehicles in the HW3 fleet. This is a substantial number of cars.

The following graph illustrates the estimated size of this fleet over time, factoring in the deliveries until Q4 2022, and then accounting for vehicle attrition with an age-dependent scrappage rate (4.5% annually for the first 12 years).

The cumulative number of HW3 vehicle production peaked at 3,054,357 in Q4 2022. Additionally, some HW2/HW2.5 vehicles from 2017/2018 were upgraded to HW3. That brings our estimated HW3 fleet peak size to ~3.1 million vehicles.

For this exercise, we'll assume:

1. FSD is solved in 2027;
2. Upgrades will be required for these HW3 vehicles; 
3. Upgrades will start in Q4 of that year.

Accounting for scrappage, the active HW3 fleet size at that time is projected to be approximately 2.8 million vehicles. This does not mean that all 2.8 million vehicles will require immediate retrofits; the upgrades will only be required for owners who have purchased FSD. In 2019 through 2022, FSD was not available for purchase in most parts of the world. Tesla officially stated that at the end of 2022, there were over 285,000 customers in the U.S. and Canada who had purchased FSD. That's a 9% FSD adoption rate. Again, using standard scrappage rates, that's still more than 225,000 vehicles requiring retrofit upgrades in 2027. However, retrofits are not the only possible solution (more on this later).

Comparing HW3, AI4, (and AI5) 

Hardware	Compute Power (TOPS)	Comparison to HW3
HW3	144 TOPS	Baseline (1x)
AI4 / HW4	300–500 TOPS	Approximately 2–3.5x more raw capable (With real-world FSD inference gains typically 3–5x)
AI5	2,000–5,000 TOPS (expected)	Approximately 14–35x more capable than HW3 (based on TOPS; production expected in 2026; Elon Musk describes it as 10–40x better than AI4 overall on FSD specific metrics)
Overall Comparison		AI4 has approximately 3 to 5 times the compute capability of HW3. AI5 is projected to vastly exceed this, enabling advanced unsupervised FSD and Robotaxi features, though exact real-world results are yet to be seen.

Can HW3 Do It?

This post generally assumes that an upgrade will be required for HW3 vehicles to achieve full autonomy, but it's important to acknowledge that Tesla is still trying to squeeze as much as possible out of HW3. During the Q3 2025 Tesla Earnings Call, this issue was brought up directly. Here's the response from Tesla management about the future of this massive installed base.

The company's CFO explicitly stated that Tesla is "not abandoning HW3" and offered a clear assurance to concerned customers: "We will definitely take care of you guys." Adding, "My personal daily driver is a HW3 vehicle." Furthermore, Ashok Elluswamy, VP of AI SW, noted that a lighter-weight version of FSD V14, will be coming in 2026 for HW3. This V14 "Lite" will provide owners with many the latest advancements in the supervised FSD, albeit months behind the AI4 deployment.

So, whether or not Tesla finds a way to cram all of unsupervised FSD into HW3, those vehicle owners will have a path to a fully self-driving vehicle. Next, let's look at what those path options may be.

Upgrade Options & Cost

In 2027 (when FSD is solved in this example), let's assume the cost to retrofit a HW3 vehicle with parts (AI5 computer, cameras) and labor is $2,500. For customers who've already paid for FSD, Tesla has an obligation to upgrade them at no cost. For owners who have not paid for FSD, this can be rolled into the purchase cost. However, look at another option.

Tesla has offered FSD transfers (off and on) for some time now (I've even used it). If Tesla were to offer current HW3-FSD owners a $2000 "upgrade incentive" credit towards a new AI5 vehicle with FSD transfer as an early adopter award, many of them might opt for this. They'd get a newer Tesla, they'd be trading in vehicles that are between 5 and 7 years old, and they'd drive off in a native AI5 vehicle (or more likely have the car drive them).

This might mean that Tesla receives a lower margin on these vehicles, but it would stop their service centers from being overrun with retrofit requests, while being cheaper than retrofits.

Above, we estimated that there would be 225,000 HW3-FSD vehicles on the roads in 2027. If Tesla had to upgrade all of these at $2500 each, the total cost would be $562,500,000. However, let's assume, by then, half of the owners will have already upgraded to newer AI4 or AI5 vehicles, and then another 50% of the remaining customers will take advantage of the upgrade incentive. That twindles the "free" upgrade number to just 56 thousand vehicles and reduces Tesla's cost to $140,625,000. This will be easily affordable for Tesla in a future quarter where "FSD is solved" and vehicle and FSD orders are pouring in.

FSD Adoption and Pricing Post-Solution

Once FSD is truly solved, the perceived value of the software will fundamentally change. FSD will transform from an advanced driver-assistance system into a fully validated eye-off, hands-off system. This will allow you to sleep, play games, watch movies, or just look out the window while your car chauffeurs you. It will also be a robotaxi enabler, allowing owners to put their cars into service to generate revenue. This certainty will dramatically increase both the adoption rate and the FSD sale price.

We can conjecture the following changes:

1. Price Increase: The purchase price of FSD (currently $8,000) will increase significantly. It's been as high as $12,000, but this is the "killer app" for cars, and this price is likely to skyrocket. At least $20,000 is a reasonable estimate. It's even possible that in 2028, to have an older Tesla (like a 2018 Model 3) where the ability to transfer FSD to a new car is worth more than the rest of the vehicle. 
2. Adoption Rate Increase: As soon as FSD is solved, the hesitancy surrounding the low (9 to 12%) adoption rate will evaporate. Customers will recognize FSD as a utility or an investment, pushing the take rate on all new vehicles to 50% in 2028 and eventually above 80% in the years that follow. 
3. Retrofit Demand: Every Tesla built since October 2016 has the camera placements to allow retrofits to a new FSD system. The HW2, 2.5, and 3 fleet comprises ~3,542,000 vehicles. This vast population of existing non-FSD-equipped vehicles would become the target of a massive upgrade demand. The potential revenue generated from these post-2027 sales and upgrades would quickly eclipse the cost for the 56 thousand "free" retrofit vehicles.

Crisitunity

Vehicles	Count
Native HW3	3,054,357
HW2/2.5 upgraded to HW3	70,000
HW3 already with FSD	285,000
HW3 vehicles that Tesla will be obligated to retrofit Due to owners not upgrading to AI4+	55,000
HW2/2.5/3 FSD Potential Adoption	2,740,000

A naïve analysis might look at the 3+ million HW3 vehicles and see an obligation to upgrade these to FSD-ready on Tesla's dime. Millions of cars, costing thousands of dollars each, would be a huge liability for any company. However, the opposite is true. Only a small percentage (~9%) purchased FSD. Of those, most have (or will) upgrade to a vehicle with newer native HW and transfer FSD. This leaves the vast majority of the HW3 fleet (97%) with no obligation for free retrofits. And each of these vehicles in the 97% can be converted to a self-driving car if the owner purchases FSD at the higher 2027 price (which more than pays for the retrofit).

Conclusion

Our analysis of the HW3 fleet highlights a moment of both challenge and opportunity for Tesla. The obligation to provide hardware retrofits to tens of thousands of customers who paid for FSD is a substantial financial undertaking, yet one that management has committed to. The explicit promise made during the Q3 2025 earnings call, "We will definitely take care of you guys," codifies the company's intent to fulfill the original FSD promise. The early adopters paid for the development of the technology, and there are multiple ways to reward them for this leap of faith without being financially ruinous. 

The early adopters helped pay the cost for unsupervised FSD, allowing Tesla entry into the vast market that this technology unlocks. This commitment, coupled with the development of the V14 "Lite" software to retain HW3 owners in the ecosystem until a complete solution is developed, ensures that the HW3 fleet remains an active and valuable part of the effort. The hardware retrofit, while costly, will be swiftly offset by the massive increase in FSD adoption rate and the resultant multiplication of the FSD purchase price. This transforms the HW3 legacy fleet from a liability into a highly valuable, revenue-generating asset that paves the way for a more sustainable future of shared autonomous mobility. The proactive communications from the company, especially regarding their promise not to abandon HW3, are essential for maintaining customer trust throughout this transition period.

Henry Ford’s Tinker’s Damn: Crafting a Future Through Innovation and Adaptability

In 1916, Henry Ford made the following statement to the Chicago Tribune: 

“History is more or less bunk. It’s tradition. We don’t want tradition. We want to live in the present, and the only history that is worth a tinker’s damn is the history that we make today.” 

This statement captures a mindset that resonates in today's high tech world. Ford, the pioneer of the assembly line and mass production, wasn’t dismissing the past out of ignorance. He was challenging the weight of tradition that hinders progress; he was urging us to shape the future through innovation. His words inspire us to embrace change, adopt new technologies, and create a history that reflects the courage to evolve. In a world where advancements like artificial intelligence, renewable energy, and biotechnology are transforming our world, Ford’s philosophy reminds us that stagnation halts growth, and an open mindset is essential for forging a meaningful future.

Embracing change doesn’t mean chasing every new gadget or jumping on every trend. It’s about cultivating a willingness to question the status quo and explore better ways of doing things. Ford’s own life exemplified this approach. His Model T wasn’t the first car, but it revolutionized transportation by making automobiles affordable for the masses. He didn’t invent the wheel. He reimagined how it could roll for everyone. You don’t need to be the first to adopt every new technology, whether it’s a quantum computer or a neural interface, but you must be open to their potential. A closed mind, tethered to “how things have always been,” risks missing the transformative power of what’s possible.

Consider today’s rapid technological landscape. In 2025, we see AI systems that can draft complex documents, analyze vast datasets, and assist in creative arts. EVs, once a niche curiosity, are now mainstream, with companies like Tesla and Rivian echoing Ford’s vision of accessible innovation. Biotechnology is pushing boundaries, from mRNA vaccines to gene-editing tools like CRISPR. These aren’t just tools. They’re invitations to rethink industries, healthcare, and human potential. Their value lies in our willingness to engage, experiment, and adapt. If we cling to tradition, insisting on fossil fuels or outdated manufacturing methods, we risk becoming relics, sidelined by a world that moves forward without us.

Ford’s “tinker’s damn” quote highlights this urgency. He saw history not as a sacred archive to worship but as a living process we shape through action. Tradition, in his view, was a chain unless it served the present. This doesn’t mean erasing the past. Ford himself learned from earlier inventors’ mechanics. It means refusing to let the past dictate the future. Today, we make history by how we respond to these emerging marvels. 3D printing has revolutionized everything from housing to prosthetics. Those who embrace it, experimenting with its applications, are writing tomorrow’s history. Those who dismiss it as a fad fade into the “bunk” Ford scorned.

An open mindset also fuels innovation by encouraging us to seek better ways to solve problems. Ford’s assembly line wasn’t just a technological leap. It was a new way of thinking about production, breaking tasks into efficient steps. Today, innovators follow suit, whether it’s SpaceX rethinking space travel or startups using blockchain to secure supply chains. These advances come from asking, “What if we did this differently?” This question drives progress. It’s why companies like Kodak faltered when they resisted digital photography, while others, like Apple, thrived by embracing the smartphone revolution.

Embracing change doesn’t mean reckless abandon. It’s about calculated openness: testing, learning, and iterating. Ford didn’t build the Model T overnight. He refined it through trial and error. Similarly, adopting new technologies requires discernment. Not every innovation is a game-changer, but dismissing them outright ensures obsolescence. An open mindset means staying curious, asking how a tool might enhance your work or life, and being willing to fail in pursuit of something better.

Ford’s legacy teaches us that history is not a museum piece. It’s what we make today. The marvels on the horizon, from AI and androids to sustainable energy and space exploration, are opportunities to shape a future that reflects our highest aspirations. If we stagnate, clinging to tradition for comfort, we stop growing. But if we embrace change, approach new technologies with curiosity, and dare to innovate, we create a history worth a tinker’s damn, one that drives humanity forward, just as Ford did a century ago.