this post was submitted on 22 Jun 2024
55 points (91.0% liked)
Environment
292 readers
1 users here now
Your definitive source for news, information, issues and activism related to the environment.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
So exactly, it's moronic to use it for AI when we could store it or reduce our usage and have less needs or transport it over longer distances for it to be used where they rely on fossil fuel.
By the way, this conversation started with hydro so challenges are different and worst case just letting water through isn't a bad thing as there's ecological impact to just keeping the water in the reservoirs and letting the level rise while there's no water downstream.
Hydro is irrelevant. It has long since peaked. The water sources that can viably produce power are already utilized. Relatively little additional hydro capacity is feasible. Even if there were, hydro is easy to incorporate into any generation structure. It can be used as baseload, peaker, storage, etc. We don't need any special considerations to utilize hydro. There is little point in discussing hydro in this context.
Solar is underutilized. It will always be underutilized. There is more solar power available than humanity could ever hope to use. It is far more abundant than hydro.
Hydro will remain useful no matter what we do, but if we want to be able to effectively utilize solar (or wind), we need to focus on demand shaping. We need more load on the grid, not less. Without additional load, solar rollout will stagnate long before it successfully replaces traditional generation methods.
Like the other projects I mentioned, AI can create an intermittent demand for excess power produced under ideal conditions. The majority of AI power is consumed while learning, training the model. Responding to queries takes very little energy. AI data centers can burn power when it is readily available, and shut down under adverse conditions, conserving power for more essential services.
Without extra demand during optimal conditions, there is no economic incentive to build enough solar capacity to meet demand during suboptimal conditions. The greater the extra load we can add, the worse conditions have to be before solar cannot meet normal demand. We need much more "extra" demand for intermittent generation to economically justify additional solar and wind.
Find me one major AI or crypto company that will accept to shut down their infrastructure when there's high load from households and offices.
You guys are fucking delusional. There's so much efficiency to be gained by stopping all the energy waste, instead you're pushing for us to use more and more energy so we build more infrastructure as if there was no environmental impact to it. Until you're able to pull the material to produce the the panels from your ass, the most efficient and environmentally friendly thing to do it always to reduce first and foremost.
Offer them ridiculously cheap power under normal conditions, but price them high during shortages. They aren't "shutting down their infrastructure". They will still be able to handle requests. They just won't be incorporating new training data to their models until the sun comes back out.
There is much, much, much, much more energy feasibly available when we focus on demand shaping instead of traditional supply shaping models. Ever hear the phrase "penny wise, pound foolish"?
You are either unaware that we are already regularly experiencing "negative rates" or you are not considering the ramifications. You are not considering how drastically "negative rates" are already stunting solar and wind development.
What I am talking about is boosting intermittent demand so that rates don't go negative. We need ways to soak up every available watt when we have more than we would normally use.
Think of it this way: our current, "supply shaping" model requires extensive use of expensiv , inefficient "peaker" plants. Peakers give us the ability to match an unregulated demand with a variable supply.
A "demand shaping" model, if enacted effectively enough, eliminates the need for inefficient peaker plants, leaving us with a moderately efficient baseload plant for overnight, and extremely efficient solar and wind during the day.
Current peaks are higher than they need to be because people are wasting energy, all you want is to increase demand as if that had no environmental impact. I'm done here.
You do realize we are already using demand shaping, but for the traditional baseload/peaker model, right?
Power companies offer steep discounts to industries like aluminum smelters and iron foundries to move their production to a night shift. Doing this increases the base load, which allows a larger percentage of the total power demand to be met by baseload generators instead of peaker plants.
The problem with this should be obvious: demand shaping on the baseload/peaker model drives demand to hours of the day that solar and wind cannot possibly meet.
Current peaks are not nearly as high as they should be. As much night-time demand as possible should be moved to daytime, where it can be met with solar. We need peak demand to correspond to peak generation. We can't move peak solar production; the trick is to shift our demand curve to match the solar production curve. Both peaks need to occur simultaneously.
You're still not comprehending. I'm meeting a far higher percentage of our current energy consumption with solar instead of dirtier and less efficient coal/oil/nuclear baseload generation, pumped storage, battery storage, etc. The "extra" demand I am asking for is 100% met by the excess capacity of our solar generation that arises under optimal conditions. That excess capacity is baked in. When we have enough solar capacity to meet normal demand with overcast skies, optimal conditions will give us massive surpluses.
Because it is met entirely with excess capacity that would otherwise be wasted, the "extra" demand I am calling for has zero additional environmental impact. It monetizes excess capacity that we wouldn't normally be able to utilize.
Give me the benefit of a doubt for a moment, and actually consider what I am saying. Yes, it sounds paradoxical at first glance, but it will make a lot more sense when you realize I'm talking about how to move the overwhelming majority of our electrical production to solar/wind and virtually eliminating peaker plants.
Hey buddy, no one serious thinks the way you do, the industry is using more fossil fuel to meet the increased demand even if there's subsidies for renewable. You're just recycling crypto bros arguments to justify wasting energy.
I can't believe I'm still answering that bullshit.
The only people serious about widespread implementation of solar are, indeed, thinking the way I am. The general concept is commonly referred to as "demand shaping" in the industry. Anyone still focused on supply shaping in 2024 is supporting coal, gas, and nuclear infrastructure. The supply shaping model attempts to resolve the differences in the supply and demand curves through grid level storage: attempting to broadly time-shift generation.
"Demand shaping" understands that storing power is inherently inefficient, and attempts to solve the differences by moving the time of consumption to the time of production.
The industry already has the solar capacity to meet the kind of demand I am talking about. They already have excess solar production that they can't effectively utilize, and we know that they can't effectively utilize it because it is regularly driving generation rates negative.
We are already producing (or capable of producing) the solar energy in question; we are wasting it due to a lack of demand. We are shutting down solar panels in the middle of the day due to a lack of demand. Solar rollout is stalling due to lack of demand for the specific power that solar is capable of producing.
When we create a demand specifically for solar energy, we increase the profitability of our existing solar infrastructure. We make it feasible and profitable to expand that infrastructure, which makes it pick up a bigger share of our normal load as well.