Hacker News Reader

Comment by pstuart | Hacker News Reader

Going through Three Mile Island and Chernobyl put me in the anti-nuclear camp.

As I got older and understood the technology better, I came around to be conceptually pro nuclear. In fact, if we hadn't had those disasters and had successfully extended our nuclear power fleet we likely wouldn't be this far into climate change.

The pitch is still compelling, but the elephant in the room is cost: renewables (including storage) are cheaper and getting cheaper still.

The antidote to this might be SMRs (Small Modular Reactors) that cut costs by mass producing reactors on an assembly line rather than thesse bespoke behemoths, but even then they apparently can't compete.

We need all the non-carbon energy we can get, and if nuclear can get us there affordably we should embrace that. The economics should be a driving force rather than pure ideology.

View on HN

> The pitch is still compelling, but the elephant in the room is cost: renewables (including storage) are cheaper and getting cheaper still.

We need to ramp up base load capacity so everyone can charge their EVs at night, or run their heat pumps through a cold winter night.

We’re not shooting at a static electric usage target. In our house, we have been electrifying everything and it’s quite surprising how much our electricity usage has gone up. In much of the country, you have design-point winter days where you need 90k BTU that’s currently being supplied by heating oil. Even with an efficient heat pump, that’s 12-13 kW, for everyone at the same time, all night. That’s an insane amount of energy storage.

We need people to charge their EV's when the sun is shining. At least in California. But perhaps not Scotland and Germany.

Given Californians drive 340 billion miles a year at 3.5 kwh per mile and 2200 hours of sunlight a year. I get you'd need 45 GW of solar to power them. California has something like 16-20 GW.

Notable at 35mpg cars use 10 billion gallons of gas. At $4 gallon. $40 billion a year.

Solar to power cars would pay for itself economically in a couple of a years.

Saying renewables are cheap is like saying SDE interns are cheap. The operational instability created by large scale adoption of renewables is an important part of the TCO.

Even places like Australia, which are naturally blessed in terms of wind and sun prevalence, are not able to divest from large scale coal power plants.

Intermittency is a huge problems for grids, and many of the widely deployable energy resources are inherently intermittent (which projects like Xlinks try to solve with other flaky engineering solutions)

> Even places like Australia, which are naturally blessed in terms of wind and sun prevalence, are not able to divest from large scale coal power plants.

This is not true.

In Australia Coal generation peaked at 179135 GWh in 2008 and was down to 126475 GWh in 2024 [1].

The state of Southern Australia generated 71.9% of it's electricity with renewables and only used batteries for 1.1% of generation. It replaced it's coal generation with gas generation, and gas generation is declining.

1. https://explore.openelectricity.org.au/energy/au/?range=all&...

2. https://explore.openelectricity.org.au/energy/sa1/?range=all...

This is absolutely true. NSW energy crises in May and 2022 caused the AEMO to cap prices. The shutdown of the largest power plant in NSW was extended again just a few months ago due to _grid instability_. Onboarding variable production is great. Just like having spot instances for compute is great.

The problem is when enough of your substrate is ephemeral, you open yourself up to a ton of new failure modes. Only "our replication and database failed at the same time" is a lot less bad then "total grid failure".

Nuclear waste storage is a problem to be solved. But that potentially contaminates, at worst, a discrete area of the planet which can be managed around. Climate change contaminates the entire freaking planet.

If you aren't serious about nuclear power, you aren't serious about climate change. We should have nuclearized our grid decades ago. The US Navy has run reactors for over 50 years without incident. It can be done.

Nuclear waste is already an effectively solved problem. Breeder reactors can recycle waste, since roughly 98% of the energy remains in the depleted fuel from light water reactors. This also reduces the volume of waste by a factor of 20X. Long term storage sites already exist for the remainder, e.g. Yucca Mountain.

Decades ago it certainly made sense. Even now if it can be done reasonably cheaply it should be done.

If it can be done cheaper with renewables and batteries, we should do that instead .

Over dependence on renewables and battery storage can be a strategic vulnerability, for example an extended stretch of cloudy and windless days. Freak weather events like hail can wipe out entire solar installations, cold snaps can disable wind turbines, batteries underperform significantly in extreme cold. And if we electrify everything, such as mandating electric stoves and heat pumps, then such an event would be a double whammy, since electricity demand spikes even more and most people won't have natural gas as a separate energy source. The Texas winter storm in 2021 is a cautionary example of what could happen.

Moreover, there's a distinct possibility of future weaponized weather modification, such as an enemy power releasing aerosols to reduce solar production, or saboteurs starting forest fires.

We should be able to model freak weather events and design around them. Having a geographically connected grid helps with this.

Smart grid. Right now there are multiple construction equipment rental facilities in my city. They all seem to have a row of quarter megawatt class towable generators ready for rental. They sit on a gravel parking lot and do absolutely nothing during an outage. Theoretically, if you fired up every unrented construction gen plus every stand by power generator at every large building, I think you could, in a distributed sense, temporarily generate several megawatts, maybe more than ten MW, in my home town. Yes it would be very expensive per KWH and eventually you'd run out of diesel, but for a couple days it would work. No wiring currently exists and no control systems exist although in theory this isn't any "worse", and probably a lot better, than having a couple MW of intermittent solar capacity. We will probably never get away from needing heavy construction equipment and needing backup generators so those will "probably" always be available in perpetuity. Maybe they'll switch from diesel to biodiesel, but it'll be the same idea. FWIW a quarter megawatt towable generator is smaller than you'd think, like 12 foot long trailer, MUCH smaller than a RV, if they were not all painted up as rental generators you'd think they're "landscaper trailers". They are HEAVY and take a dualie truck to tow, like "ten thousand pounds" with full tanks and very long and very large electrical cables, etc. A row of ten of them in a parking lot is not overly impressive looking until you realize that's 2.5 megawatts on tap, and these are industrial rated so they are designed to run full power 24x7 in the worst weather conditions. A year around long term average estimate is maybe 1.5 KW continuous per house so perhaps they could only run maybe ten thousand homes if they shut down every non-essential business and fired up every generator. However my city only has 30K homes so dropping demand by 1/3 would be "pretty helpful" for renewables if they can't quite keep up. If I owned an EV I would be happy to charge it down the road at 15 cents/KWh (towards the high end for my state) and discharge it into a failing grid for maybe $1.50 KWh, if I could get the cycle time down to an hour (very optimistic...) and I moved 100 KWh each time, I could make nearly $150/hr as an "electricity tank driver". Realistically I think a failing grid would bid up prices well over $1.50 so I think this quite reasonable... if every charger could be upgraded to backfeed into the grid then a fleet of EVs would be a huge source of mobile power.

The US Navy runs 50 MWe reactors inside billion dollar submarines. The requirements are almost entirely orthogonal to economically viable power reactors.

Dealing with the comparatively tiny amount of nuclear waste vs. continuing to pump billions of tons of CO2 into the atmosphere seems like an easy tradeoff.

Immobilize it in glass, stick it underground in a remote area.

Reply on news.ycombinator.com