An Excursion into A.I.
Criticism of my writing by a Robot, Claude
I did not intend to go in this direction in my Electrify posts, but there are a number of topics I find myself interested besides the grid; Transportation (because of its GHG impact), taxes (wondering why we tax taxes here in the USA), and Artificial Intelligence (for many reasons).
So the last of these, A.I., recently has led to a lot of reading and arguments with my friends. I realize how much I rely on my calculator, my phone (actually mini-PC), and all the other modern electronic helpers we take for granted now. I am disappointed in how much less I process between my ears, instead relying on my fingers to do so many calculations and retrievals.
One of my best friends insists I have a bad attitude about this. He is using A.I. engines to improve his firm’s service products. He subscribes to my Electrify post and decided to ask his engine (Claude, no last name) what was wrong with my post.
Well I will relieve you of having to read the specific criticisms of my research and writing, and will simply offer you Claude’s version of that same 21st Century Grid Challenges post for fun.
21st Century Grid Challenges: Beyond Simple Electrification
Gerald Glaser as rewritten by Claude
Nov 16
In my previous post, I explored how the electric grid functions and why its operation is remarkable. Now I want to examine why yesterday’s solutions won’t solve tomorrow’s distribution challenges—and propose a more nuanced path forward.
The Scale of Change
From 1900 to 2000, world population grew 3.8-fold while energy use grew 8.5-fold. In the United States specifically, energy consumption increased from 2,700 tWh to 28,633 tWh—a tenfold increase—while population only tripled. This disproportionate growth in energy intensity has defined modern life.
The transformation wasn’t just in quantity but in form. In 1900, primary energy sources were coal and wood, burned directly for heat. By 2000, our energy mix included nuclear (8%), coal, natural gas, petroleum (40%), and renewables (7%, primarily biomass and hydro). But this shift in sources masks a more fundamental transformation: the conversion from primary heat energy to electricity.
The Efficiency Paradox
Throughout the 20th century, when energy primarily meant burning fuel for heat, direct combustion made sense. But as we needed energy for transportation, lighting, and complex machinery, electricity became essential despite its conversion losses.
Converting fossil fuels to electricity wastes substantial energy at each transformation step: chemical bonds to heat, heat to mechanical motion, mechanical to electrical. According to Lawrence Livermore National Lab’s 2023 data, of 93.6 quads of primary energy harvested in the U.S., 61.5 quads (66%) were rejected as waste heat before doing any useful work.
When electricity was a luxury, this inefficiency was tolerable. Today, electricity consumes 22% of primary energy, up from 0.1% a century ago. With building electrification, electric vehicles, and new industrial processes like electric steel production, this percentage will continue rising sharply.
What the Numbers Actually Tell Us
We’ve achieved remarkable efficiency gains: buildings are better insulated, LED lighting is 90% efficient versus 5% for incandescent bulbs, hybrid cars achieve 50 mpg, and aircraft have dramatically improved fuel economy in just twenty years.
Yet total energy consumption—and especially electricity demand—continues growing. This isn’t a failure of efficiency; it’s the Jevons paradox in action: as energy use becomes more efficient and cheaper, we find more applications for it.
Here’s where the scale becomes clear. My household provides one data point: after insulating our home, reducing appliance loads, and converting to all-electric systems including two EVs, our total energy consumption dropped 62%. But our electricity consumption nearly tripled. If this pattern holds broadly—and early data suggests similar trends—the distribution grid faces an unprecedented challenge.
To put this in context: PG&E serves approximately 16 million people. If even a quarter of households follow similar electrification patterns over the next decade, the distribution system would need to handle roughly 50-75% more peak capacity in many neighborhoods. The current grid wasn’t designed for this.
The Supply Question: Largely Solved
Several years ago, I attended a board meeting where our local electric utility presented load forecasts to the California Public Utilities Commission. Their model projected minimal demand growth, assuming efficiency gains would offset new electric applications. That assumption is being rapidly reconsidered.
But can we actually generate enough clean electricity?
The answer appears to be yes. In 2024, the EIA reported that new generating capacity broke down as: 58% solar, 13% wind, with nuclear beginning to ramp up again, while fossil fuel plants represented only 4% of additions. In my region, advanced geothermal facilities are being developed that can be built almost anywhere by tapping deep earth heat.
Total new capacity in 2024 was projected at 62.8 GW—55% more than the 40.4 GW added in 2023. For context, total U.S. generating capacity is approximately 1,200 GW, so we’re adding roughly 5% annually, with renewables dominating. At current growth rates, generation capacity should keep pace with electrification, though regional bottlenecks and timing mismatches will create challenges.
The generation problem has technical and economic solutions. The distribution problem is more complex.
The Cost Question: More Nuanced Than It Appears
A builder I spoke with recently questioned whether the grid can handle growing electric demand. His concern: “Where is all that energy coming from? It’s really expensive compared to existing fossil fuel options.”
Let’s examine this carefully.
California has the second-highest residential electricity rates in the nation at 30.72¢/kWh (after Hawaii), compared to North Dakota’s 7.42¢/kWh and a national average of 15.59¢/kWh. This disparity requires explanation.
Natural gas prices are highly volatile: $9.33/MMBTU in 2022, $1.63/MMBTU in 2024, currently $4.31/MMBTU. PG&E charges $6.70/MMBTU for the gas itself, plus $22.20/MMBTU for delivery. This means delivery costs represent 77% of what consumers pay.
Even if gas prices double, consumers only see a 23% bill increase because delivery dominates the cost structure.
Electricity pricing is more complex, with Time-of-Use schedules, seasonal variations, and zone-based rates. During winter peak periods, 67% of electricity cost is delivery infrastructure; in summer it drops below 50%. California’s high rates stem largely from aging infrastructure and wildfire damage liability—costs that should decrease as utilities underground lines and improve safety systems. PG&E has already reduced residential rates in the past year while promising further reductions.
However, this cost comparison misses critical factors:
Equipment efficiency matters. A heat pump with a coefficient of performance of 3.0 delivers three units of heat per unit of electricity consumed. Even at California’s high rates (30¢/kWh), that’s effectively 10¢ per kWh-equivalent of heat. Natural gas at PG&E’s total delivered price of $28.90/MMBTU, burned in a 95% efficient furnace, costs roughly 10¢ per kWh-equivalent. The economics are already comparable in high-cost electricity markets, and favor heat pumps in most other states.
Total cost of ownership varies dramatically. Installing residential solar and batteries requires $20,000-$40,000 upfront. Not everyone can afford this, and renters, apartment dwellers, and those with poor solar exposure have no access to these solutions. Natural gas infrastructure already exists in many homes.
Long-term price trajectories are uncertain. I claim electricity will become cheaper relative to gas, but this assumes continued renewable deployment, stable policy support, and resolution of grid integration challenges. Gas prices could remain low if fracking continues, or spike if exports increase or regulations tighten. Making 25-year predictions about either fuel’s relative cost is speculative.
My fundamental position: over the 15-25 year lifespan of HVAC equipment and appliances being installed today, electricity will likely be cost-competitive or cheaper than fossil alternatives in most markets. But this isn’t guaranteed—it depends on policy, technology learning curves, and infrastructure investment.
The Question Everyone Should Be Asking
The real challenge isn’t energy generation or even long-term costs. It’s the distribution grid—that final mile from substation to home.
As I explained in my previous post, peak demand determines grid design: conductor sizing, transformer capacity, and branch circuit extent. If households triple their electricity consumption while also changing the nature of peak demand—EVs charging for hours instead of air conditioners spiking for minutes—the distribution system becomes a bottleneck.
Utilities face three options:
Massive infrastructure replacement - Upgrade transformers, conductors, and substations across entire service territories. Cost: hundreds of billions nationally. Timeline: decades. This is economically prohibitive.
Managed demand - Smart thermostats, dynamic pricing, vehicle-to-grid technology, and direct load control can flatten peaks. This helps but requires coordination, customer acceptance, and sophisticated control systems. It’s part of the solution but not sufficient alone.
Distributed generation and storage - Rooftop solar and home batteries can meet local demand, reducing distribution requirements. This is promising but excludes many customers and requires new regulatory frameworks for grid integration.
The realistic path forward combines all three: strategic infrastructure upgrades where needed, aggressive demand management, and distributed resources where feasible.
But here’s what’s missing from most discussions: who pays, and how do we ensure equity?
Wealthy homeowners can install solar and batteries, insulating themselves from high rates and grid constraints. Renters and apartment dwellers cannot. If utilities raise rates to fund infrastructure upgrades, they disproportionately burden those who can’t invest in self-generation. If we rely too heavily on distributed resources, we create a two-tier system.
What Comes Next
We cannot produce or store natural gas at home. We can collect solar energy and store it in batteries. This asymmetry suggests a path forward that utilities and regulators are still grappling with: a hybrid grid model where distribution serves as backup and balancing rather than primary supply for many customers.
In my next post, I’ll examine one approach to this transformation—how residential solar, batteries, and smart controls could reshape the distribution system while addressing both cost and equity concerns. The solution isn’t purely technical; it requires rethinking the utility business model, rate structures, and our definition of grid reliability.
The question isn’t whether we can electrify everything. We can. The question is whether we can do it affordably and equitably, without stranding those who can’t participate in the distributed energy transition.