 |
|
|
|||||
Heat pumps can reduce carbon emissions associated with heating buildings, and many states have set aggressive targets to increase their use in the coming decades. But while heat pumps are often cheaper choices for new buildings, getting homeowners to install them in existing homes isnt so easy. Current energy prices, including the rising cost of electricity, mean that homeowners may experience higher heating bills by replacing their current heating systems with heat pumpsat least in some regions of the country. Heat pumps, which use electricity to move heat from the outside in, are used in only 14% of U.S. households. They are common primarily in warm southern states such as Florida where winter heating needs are relatively low. In the Northeast, where winters are colder and longer, only about 5% of households use a heat pump. In our new study, my coauthor Dan Schrag and I examined how heat pump adoption would change annual heating bills for the average-size household in each county across the U.S. We wanted to understand where heat pumps may already be cost-effective and where other factors may be preventing households from making the switch. Wide variation in home heating Across the U.S., people heat their homes with a range of fuels, mainly because of differences in climate, pricing, and infrastructure. In colder regionsnorthern states and states across the Rocky Mountainsmost people use natural gas or propane to provide reliable winter heating. In California, most households also use natural gas for heating. In warmer, southern states, including Florida and Texas, where electricity prices are cheaper, most households use electricity for heatingeither in electric furnaces, baseboard resistance heating, or to run heat pumps. In the Pacific Northwest, where electricity prices are low due to abundant hydropower, electricity is also a dominant heating fuel. The type of community also affects homes fuel choices. Homes in cities are more likely to use natural gas relative to rural areas, where natural gas distribution networks are not as well developed. In rural areas, homes are more likely to use heating oil and propane, which can be stored on property in tanks. Oil is also more commonly used in the Northeast, where properties are olderparticularly in New England, where a third of households still rely on oil for heating. Why heat pumps? Instead of generating heat by burning fuels such as natural gas that directly emit carbon, heat pumps use electricity to move heat from one place to another. Air-source heat pumps extract the heat of outside air, and ground-source heat pumps, sometimes called geothermal heat pumps, extract heat stored in the ground. Heat pump efficiency depends on the local climate: A heat pump operated in Florida will provide more heat per unit of electricity used than one in colder northern states such as Minnesota or Massachusetts. But they are highly efficient: An air-source heat pump can reduce household heating energy use by roughly 30% to 50% relative to existing fossil-based systems and up to 75% relative to inefficient electric systems such as baseboard heaters. Heat pumps can also reduce emissions of greenhouse gases, although that depends on how their electricity is generatedwhether from fossil fuels or cleaner energy, such as wind and solar. Heat pumps can lower heating bills We found that for households currently using oil, propane, or non-heat pump forms of electric heatingsuch as electric furnaces or baseboard resistive heatersinstalling a heat pump would reduce heating bills across all parts of the country. The amount a household can save on energy costs with a heat pump depends on region and heating type, averaging between $200 and $500 a year for the average-size household currently using propane or oil. However, savings can be significantly greater: We found the greatest opportunity for savings in households using inefficient forms of electric heating in northern regions. High electricity prices in the Northeast, for example, mean that heat pumps can save consumers up to $3,000 a year over what they would pay to heat with an electric furnace or to use baseboard heating. A challenge in converting homes using natural gas Unfortunately for the households that use natural gas in colder, northern regionsmaking up around half of the countrys annual heating needsinstalling a heat pump could raise their annual heating bills. Our analysis shows that bills could increase by as much as $1,200 per year in northern regions, where electricity costs are as much as five times greater than natural gas per kilowatt-hour. Even households that install ground-source heat pumps, the most efficient type of heat pump, would still see bill increases in regions with the highest electricity prices relative to natural gas. Installation costs In parts of the country where households would see their energy costs drop after installing a heat pump, the savings would eventually offset the up-front costs. But those costs can be significant and discourage people from buying. On average, it costs $17,000 to install an air-source heat pump and typically at least $30,000 to install a ground-source heat pump. Some homes may also need upgrades to their electrical systems, which can increase the total installation price even more, by tens of thousands of dollars in some cases, if costly service upgrades are required. In places where air conditioning is typical, homes may be able to offset some costs by using heat pumps to replace their air conditioning units as well as their heating systems. For instance, a new program in California aims to encourage homeowners who are installing central air conditioning or replacing broken AC systems to get energy-efficient heat pumps that provide both heating and cooling. Rising costs of electricity A main finding of our analysis was that the cost of electricity is key to encouraging people to install heat pumps. Electricity prices have risen sharply across the U.S. in recent years, driven by factors such as extreme weather, aging infrastructure, and increasing demand for electric power. New data center demand has added further pressure and raised questions about who bears these costs. Heat pump installations will also increase electricity demand on the grid: The full electrification of home heating across the country would increase peak electricity demand by about 70%. But heat pumpswhen used in concert with other technologies such as hot-water storagecan provide opportunities for grid balancing and be paired with discounted or time-of-use rate structures to reduce overall operating costs. In some states, regulators have ordered utilities to discount electricity costs for homes that use heat pumps. But ultimately, encouraging households to embrace heat pumps and broader economy-wide electrification, including electric vehicles, will require more than just technological fixes and a lot more electricityit will require lower power prices. Roxana Shafiee is an environmental fellow at the Center for the Environment at Harvard Universitys Harvard Kennedy School. This article is republished from The Conversation under a Creative Commons license. Read the original article.
Category:
E-Commerce
AI isnt eliminating human work. Its redistributing human judgment, away from routine tasks and into the narrow zones where ambiguity is high, mistakes are costly, and trust actually matters. This shift helps explain a growing disconnect in the AI conversation. On one hand, models are improving at breathtaking speed. On the other, many ambitious AI deployments stall, scale more slowly than expected, or quietly revert to hybrid workflows. The issue isnt capability. Its trust. The trust gap most AI strategies overlook AI adoption doesnt hinge on whether a system can do a task. It hinges on whether humans are willing to rely on its output without checking it. That gap between performance and reliance, the trust gap, is what ultimately determines where AI replaces work, where it augments it, and where humans remain indispensable. Two factors shape that gap more than anything else: ambiguity and stakes. Ambiguity refers to how much interpretation, context, or judgment a task requires. Stakes refer to what happens if the system gets it wrong: financially, legally, reputationally, or ethically. When ambiguity is low and stakes are low, automation thrives. When both are high, humans must stay firmly in the loop. Most real-world work lives somewhere in between and thats where the future of labor is being renegotiated. A simple way to see where AI fits Think of work along two axes: how ambiguous it is, and how costly errors are. Low ambiguity, low stakes tasks, basic classification, simple tagging, routine routing, are rapidly becoming fully automated. This is where AI quietly replaces human labor, often without much controversy. Low ambiguity but high stakes tasks, such as compliance checks or identity verification, are typically automated but closely monitored. Humans verify, audit, and intervene when something looks off. High ambiguity, low stakes work: creative labeling, sentiment analysis, exploratory research, which tends to use AI as an assistant, with light human oversight. But the most important quadrant is high ambiguity and high stakes. These are the tasks where trust is hardest to earn: fraud edge cases, safety-critical moderation, medical or financial interpretation, and the data decisions that shape how AI models behave in the real world. Here, humans arent disappearing. Theyre becoming more targeted, more specialized, and more on demand. When the human edge actually disappears Interactive voice response systems refine the rule. The stakes were not low, IVR is literally the companys voice to its customers. But ambiguity was. Once synthetic voices became good enough, quality was easy to judge, variance was low, and the trust gap collapsed. That alone was sufficient for AI to take over. When trust keeps humans in the loop Translation followed a different trajectory. Translation is inherently ambiguous, as there are multiple ways to translate a sentence. As a result, machine translation rapidly absorbed casual, low-risk content such as TikTok videos. However, in high-stakes contexts, such as legal contracts, medical instructions, financial reporting, and global brand messaging, trust is never fully transferred to the machine. For these tasks, professional translators are still required to augment the AI’s initial output. Since AI now performs the bulk of the work, full-time translators have become rare. Instead, they increasingly operate within expert networks, deployed just-in-time to fine-tune and verify the process, thereby closing the trust gap. The same shift is now playing out in how data is prepared and validated for AI systems themselves. Early AI training relied on massive, full-time human labeling operations. Today, models increasingly handle routine evaluation. Human expertise is reserved for the most sensitive decisions, the ones that shape how AI behaves under pressure. What this means for the future of work The popular narrative frames AI as a replacement technology: machines versus humans. The reality inside organizations looks very different. AI is becoming the default for scale. Humans are becoming the exception handlers, the source of judgment when context is unclear, consequences are severe, or trust is on the line. This doesnt mean fewer humans overall. It means different human roles: less repetitive labor, more judgment deployed just in time. More experts working across many systems, fewer people locked into single, narrowly defined tasks. The organizations that succeed with AI wont be the ones that automate the most. Theyll be the ones that understand where not to automate, and that design workflows capable of pulling human judgment in at exactly the right moment, at exactly the right level. The future of work isnt humans versus machines. Its AI at scale, plus human judgment delivered through expert networks, not permanent roles. Translation and model validation show the pattern; white-collar work is next. And that, quietly, is what companies are discovering now.
Category:
E-Commerce
AI can do incredible things. So far, though, most of those things have been virtual. If you want a killer article for your bichon frise blog or an expertly crafted letter disputing a parking ticket you probably deserve, chatbots like ChatGPT and Gemini can deliver that. All those things are locked into the nebulous world of information, though. Theyre helpful, but the products of todays large language models (LLMs) and neural networks arent actually doing much of anything. AIs silicon-bound status, however, is beginning to change. The tech is increasingly invading the real world. 2026 is the year that AI gets physical. And that shift has huge implications for the future of the technologyand for the impact when it fails. Call Me a Robot The change started with cars. The idea of a self-driving car goes back to the 1950s. But the technology always felt like it was decades away. Now its here. Robotaxi companies like Waymo and Zoox give more than 450,000 rides per week to paying customers. I ride in Waymo vehicles all the time, and I love calling a robot from an app and having it drive me across town. Self-driving cars finally arrived because of a whole slew of things, including cheap lidar scanners and better batteries. But the rise of deep learning and AI played the most pivotal role. The AI models that power Waymo vehicles are much better at driving than humans. And they can learn and improve on the flyhere in San Francisco where I live, Waymos have gotten more assertive as theyve learned the roads better. Self-driving AI is getting so good that its increasingly able to navigate roads without the need for the fancy (and expensive) sensors you see atop first-generation Waymos. Tesla uses simple cameras, and is getting closer to true self-driving. Fold My Laundry, Siri Self-driving cars are an incredible application of physical AI. But theyre hardly the only one. Driving is a great initial test case for the tech, because it has fairly clear rules and limits. Cars need to stay on the road, recognize red lights, and minimize cat fatalities. Other physical tasks are harder to automate with AI. But they have potentially even bigger upsides. Companies are increasingly pairing artificial intelligence with humanoid robots, teaching the robots artificial brains about the physical world so they can navigate it capably. The ultimate dream is to put these robots to work. They could perform a wide variety of jobs in factories or warehouses, for example. Generally speaking, current industrial robots need to be specifically built for a single task, but an AI powered one could learn multiple onesassembling a product and then placing it on a shelf, for example. But AI-powered robots could also fill gaping holes in the human labor market. Caretaking for the elderly is incredibly important as the world gets older on average. Yet finding enough people for caretaking roles is nearly impossible. Especially in countries like Japan, robots are beginning to fill the gaps. Dexterous, AI-powered robots may soon work well enough for tasks like doing dishes, folding laundry, or even cooking to be automated. These robot companions could help elderly people live on their own more independently. With advanced LLMs, they could even form relationships with their real-world charges, helping with loneliness or reminding a person with memory challenges to take their meds on schedule. The Parable of the Raunchy Bear Of course, all of this comes with risks. When an LLM hallucinates in a virtual space, its annoying but rarely damaging. If your ChatGPT-generated recipe for meatballs sucks, you probably wont die. And if the chatbot writing your blog post confuses a bichon for a poodle, your dog will be very angry with you, but otherwise the consequences are minor. Physical AI is different. Clearly, if Waymos technology goes awry, it could accidentally steer a 5,000-pound object into a building or a bystander. And youve read enough science fiction that I dont need to remind you about robot uprisings. Many of these risks are well understood, though, and thus well controlled. Power outages aside, Waymos rarely run into serious challenges on the road, and industrial robots rarely injure people. The bigger risks start to creep in when AI is applied haphazardly to the physical world without a lot of oversight or planning. As physical AI expands and LLMs get cheaper, this will happen more often. Take the case of an AI teddy bear with a built-in LLM. It was supposed to chat with kids, and perhaps read them bedtime stories. Instead, it started instructing them on BDSM and other raunchy topics, as well as how to pop pills and where to find knives. The bear was quickly pulled from the market. But the lesson is clear: Unlike traditional computer code, LLMs are nondeterministicyou cant predict their outputs from the inputs you feed them. In 2026 and beyond, this will mean cars that avoid accidents better than human drivers, robots that can easily learn work theyve never done before, and AI embedded in physical systems (like power and utility grids) that can instantly respond to damage or outages. But it will also mean lots of failuresand perhaps a few catastrophic ones. LLMs unpredictability is their power. But as AI gets physical, that unpredictability will also lead to a faster, less tractable, more chaotic world.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||