Correct--while the sun sends an enormous amount of free energy our way, it's maddeningly diffuse and requires enormous investments of energy and materials to capture it in a usable and reliable form. The Wynn in Vegas has a 160 acre solar farm in some of the sunniest land in the country and it provides...almost enough energy to power a single large hotel.
None of this is to say that solar power is bad, just that we should have measured expectations.
Dense is a subjective term, and solar is without a doubt a dense enough energy source.
Compare the amount of land needed to power the US with solar to the amount we use to brew a small amount of ethanol as a minor energy sourced. We can completely power the US with a fraction of the land that we currently use as a subsidy for corn farmers, who grow more corn than we can eat.
While I'm less optimistic than you on solar power, you're absolutely right about ethanol, which is easily one of the worst and most inefficient power sources we could have possibly chosen to invest in. It's a significant indictment of the
US political system that so much money was shoveled into it.
Now include the largest uranium mines like inkai in your area calculation. Or alternatively any recently opened coal mines.
Then include that any solar panels feeding variable load can go wherever (like roofs and parking lots) and take no land but instead enhance the other use.
Then include tilting agrivoltaics (which increases yield for heat sensitive crops and makes animals healthier).
Dedicated utility PV in a good area has a higher area energy density than any scalable power source except gas. Include dual use and it has the lowest land impact worldwide. Specific power is pretty damn good too, about 20W/kg (or 5-8W/kg net, so slightly better than nuclear).
Strategic installation of solar panels certainly helps maximize efficiency, and is in general a great idea, especially in some areas. But to your point about the space taken up by uranium and coal mines, that's only a fair comparison if you also calculate the mines needed for things like lithium, copper, etc. that would all go into a solar-oriented power grid.
It's not because they're no longer needed once the panel is produced. That would be analogous to the sand and copper and steel and indium and cadmium and gadolinium and chromium and zirconium and nickel and silver in the nuclear plant (and fuel cycle). All of which except silver outmass minerals of similar rarity and mining impact in the solar panel (and that is quickly changing). Even then it's questionable because most of the nuclear reactor cannot be recycled, but the solar panel legally must have recycling prepaid in many jurisdictions.
You're welcome to include mining for the entire supply chain of solar and exclude the supply chain of the generator for thermal though, it doesn't change how one sided the land use is in solar's favour.
"No longer needed once the panel is produced"
The land used for the mining isn't land that we're just getting back--given the costs of making it useful again for anything else, space used for a lithium mine is in almost all cases essentially going to be gone for good. This is of course true for other mines as well, but your notion that lithium is an exception is a curious one.
"All of which except silver outmass minerals of similar rarity and mining impact in the solar panel (and that is quickly changing)"
Nuclear reactors require relatively small inputs of these metals compared to the metals used in solar panels, and the huge capital invested in making solar panels less resource-intensive could also be applied to nuclear, if we wanted to do so.
"Even then it's questionable because most of the nuclear reactor cannot be recycled, but the solar panel legally must have recycling prepaid in many jurisdictions."
But again, reactors are small compared to the millions of panels that are necessary to be the equivalent of one plant, and furthermore plenty of those materials can be and are recycled, especially in France (the entire history of the American nuclear energy program have created less waste than solar panels have in just a couple decades). And, again, innovations that make solar panels more recyclable (which we absolutely need because right now they mostly just produce massive amounts of toxic waste) could also be invested in nuclear recycling. I did get a good chuckle out of your vague "many jurisdictions" though.
> The land used for the mining isn't land that we're just getting back--given the costs of making it useful again for anything else, space used for a lithium mine is in almost all cases essentially going to be gone for good. This is of course true for other mines as well, but your notion that lithium is an exception is a curious one.
Solar panels aren't made of lithium.
> Nuclear reactors require relatively small inputs of these metals compared to the metals used in solar panels, and the huge capital invested in making solar panels less resource-intensive could also be applied to nuclear, if we wanted to do so.
[Citation needed] Solar panels don't need tonnes of indium per GW or gadoliunium or an ongoing 100kg of copper per MW per year. Modern panels on a modern racking system have a higher capacity weighted specific power than an EPR and a lower metal fraction. The cells (which are still over 90% silicon) are only about 2% of the total mass of a module and weigh less than the raw uranium for an equal energy output, let alone the rest of the reactor and supply chain.
That last is just more lies. All of the solar panels ever produced could fit in the tailings pit of Husab dug out for a single year of operation. Half of the US dragging their feet doesn't discount the fact that most new PV in the civilized world is recyclable and mandatory to do so.
Do you have anything honest to say or just the same slimy lies?
Solar panels aren't made of lithium, but they are made of a wide variety of other things that have to be mined (which, again, uses up space and has to be taken into consideration), and the batteries which will be necessary to make them a viable part of the total transition to renewables will require massive amounts of it--lithium mining is expected to double in the next few years, and renewables are a major driver of that.
Nuclear reactors use less copper overall than renewable energy sources. (https://help.leonardo-energy.org/hc/en-us/articles/360010919...) Uranium, by the way, is also recycleable--most fissile material can be reused, and the amount of waste produced is tiny.
And yes, the total amount of waste produced by nuclear energy is miniscule, and could fit in a much smaller amount of space than is taken up by solar panel waste, which is well over a quarter of a million metric tons. Both are small compared to, say, coal ash, but solar disposal does have significant costs and can and does produce significant waste.
And one last thing--you're hypersensitive and pathetic. I started this discussion agreeing with your main point and trying to observe something straightforwardly relevant and you've been having a tantrum in response the entire time. Log off.
Again, quantify them rather than vague hand waving. 1kW net of PV module has about 100g of metal in it and lasts 25-50 years. 1kW of fuel rod lasts 3-6 years and requires 130g of enriched uranium, a bunch of rare earths, 2kg of copper and steel for handling and requires extracting 1kg of raw uranium. The amount of lithium 'required' is zero, but if you choose to use LFP for diurnal storage, you need about 1kg. Roll together mining impact for battery, PHES, and solar and you've still not covered the uranium mine, let alone all the steel, extra copper, indium, gadolinium, chromium and so on. Inkai is over 460km^2 with a much larger zone in which the ground is too poisonous to inhabit. If you want to see how much copper is actually needed, maybe rather than using an unsourced article from 2018, look at the most recent IRENA or Frauenhofer PV reports. Racking systems have changed completely, modules are higher power and have less metal, and connection voltages are in the kV range now.
> And yes, the total amount of waste produced by nuclear energy is miniscule, and could fit in a much smaller amount of space than is taken up by solar panel waste, which is well over a quarter of a million metric tons. Both are small compared to, say, coal ash, but solar disposal does have significant costs and can and does produce significant waste.
Again. Compare for me the volume of the tailings pit of a typical open pit uranium mine like Husab with the volume of every solar panel ever produced. A few hundred grams of metal encased in glass for an entire lifetime of energy is not 'significant waste'.
> And one last thing--you're hypersensitive and pathetic. I started this discussion agreeing with your main point and trying to observe something straightforwardly relevant and you've been having a tantrum in response the entire time. Log off.
Projection much? You tried to push lies and propaganda, now you're having a tantrum when challenged. Every topic where nuclear shilling isn't immediately banned is full of the same set of talking points that have been debunked between years and decades ago.
What on earth do you mean by 1kw of fuel rod? Your quantifications make no sense and cite no sources. You're going to need to be more specific than "read the latest reports"; this report (https://www.irena.org/-/media/Files/IRENA/Agency/Technical-P...) from 2021 for example makes it very clear that copper is not a trivial concern for renewables at all. If you have serious evidence for your unclear and uncited claims of "how much is actually needed", then go right ahead. none of your other numbers are cited either; your assumption of 1kg for lithium battery looks like you just made it up on the spot the way you did with your bizarre earlier falsehood about how much copper a nuclear reactor consumes--and a transition to a renewable-centric energy system will require significant battery storage capacity, so it's not a question that can just be dismissed.
A fuel rod assembly with 500kg of uranium produces about 200,000,000 kwH over its lifetime (https://www.iaea.org/sites/default/files/publications/magazi...), which comes out to 400k kw per kilogram of uranium, or 400 kw per gram. The amount of rare earths in each solar panel may be relatively low, but it also uses a significant amount of silicon, which as a metalloid has to be mined. Beyond that, none of this even addresses the differences in intensity of production; because nuclear produces its lifetime energy much more quickly than a solar panel does, the relative density of solar that you need in order to provide a comparable level of power over time is considerably unbalanced.
"Again. Compare for me the volume of the tailings pit of a typical open pit uranium mine like Husab with the volume of every solar panel ever produced. A few hundred grams of metal encased in glass for an entire lifetime of energy is not 'significant waste'."
I'm at a loss for how to explain to you that this a fallacy on your part. I'm talking about waste production, and you're trying to compare apples to oranges by bringing up uranium mines instead of comparing waste production of the different methods. That you're so intent on repeating this obvious fallacy suggests to me that it's the best you can do. I'd like to say it doesn't matter, but the anti-rational fanaticism of green energy cultists is going to push the world off a cliff.
You can't even distinguish between energy and power, now you're mixing kw and kWh. A reactor with about 100t of fuel rods produces around 1GW and refuels every 3-6 years. Ergo 100g of refined uranium (plus zirconium and gadolinium, then caesium indium and silver for control rods).
The toxic slurry in a uranium mine is waste from producing energy. It outmasses everything else you are talking about combined by orders of magnitude and is rarely dealt with in any permanent or safe way, see the Indian village of Kadapa or any of the mines in Niger or Uzbekistan for examples.
Then bringing up the most abundant element on earth as if mining it is a relevant impact is another huge stretch. Quartz mining doesn't even need the same grade of sand as concrete. What an utterly intellectually bankrupt claim.
You're the one trying to concern troll over the space and minerals used by PV. Demonstrate that the total land use and mining impact of the entire supply chain is actually higher.
I also note you've cherry picked a report that talks largely about vehicle batteries but tangentially mentions thin film panels (an obsolete technology being abandoned) rather than monocrystalline, and then doesn't quantify it per MW (hint: the numbers in it are nowhere near your claimed 4t/MW). Where is all this copper supposed to be? You're claiming that there's 100kg of copper hiding in this photo https://www.solarpowerworldonline.com/wp-content/uploads/201...
You're very narrowly trying to compare SNF with the entire solar module, ignoring the reactor, the upstream fuel supply, the low level waste, and the waste containment. The sheer stupidity of thinking you have a coherent enough lie is mind boggling.
Significantly less, but not debilitatingly less. For a panel in direct sunlight and adjusted to point directly at the sun, the ground value is 1,050 watts per square meter.
I agree with your perspective on this, but my hope is that it won't have a major negative externality at the end of the day--drum machines and drum software have really solid for well over a decade now, but plenty of people still learn the drums, learn them well, and find opportunities to play them.
This is a really excellent point, both in general and vis a vis the conflict in Ukraine. Something I see happening often is that some people are quick to confuse expertise-as-credential with expertise-as-substance, or to assume that anyone who has above average knowledge of some relevant subject area that relates to some aspect of the war is immune to emotionally biased interpretations.
I think it's an oversimplification to argue that the mind as a whole is like a closet, but I do think memory specifically works a bit in this way--hence the well-established utility of memory palace methods. Giordano Bruno's memory garden idea is something I've been working on, but it does take time to put in place.
This is spot on in my opinion and I wish more people would keep it in mind--it may well be that large language models can eventually become functionally very much like AGI in terms of what they can output, but they are not systems that have anything like a mind or intentionality because they are not designed to have them, and cannot just form it spontaneously out of their current structure.
This very much seems like a "famous last words" scenario.
Go play around with Conway's Game of Life if you think that things cannot just spontaneously appear out of simple processes. Just because we did not "design" these LLM's to have minds does not mean that we will not end up creating a sentient mind, and for you to claim otherwise is the height of arrogance.
It's Pascal's wager. If we make safeguards and there wasn't any reason then we just wasted a few years, no big deal. If we don't make safeguards and then AI gets out of our control, say goodbye to human civilization. Risk / reward here greatly falls on the side of having extremely tight controls on AI.
My response to that would be to point out that these LLM models, complex and intricate as they are, are nowhere near as complex as, for example, the nervous system of a grasshopper. The nervous systems of grasshoppers, as far as we know, do not produce anything like what we're looking for in artificial general intelligence, despite being an order of magnitude more complicated than an LLM codebase. Nor is it likely that they suddenly will one day.
I don't disagree that we should have tight safety controls on AI and in fact I'm open to seriously considering the possibility that we should stop pursuing AI almost entirely (not that enforcing such a thing is likely). But that's not really what my comment was about; LLMs may well present significant dangers, but that's different from asking whether or not they have minds or can produce intentionality.
You forget that nervous systems of living beings have to handle running the bodies themselves in the first place, which is also a very complicated process (think vision, locomotion etc). ChatGPT, on the other hand, is solely doing language processing.
That aside, I also wonder about the source for the "nowhere near as complex" claim. Per Wikipedia, most insects have 100-1000k neurons; another source gives a 400k number for grasshopper specifically. The more interesting figure would be the synapse count, but I couldn't find that.
In most cases there are vastly more synapses than there are neurons, and beyond that the neurons and synapses are not highly rudimentary pieces but are themselves extremely complex.
It's certainly true that nervous systems do quite a bit more than language processing, but AGI would presumably also have to do quite a bit more than just language processing if we want it to be truly general.
I agree with the general point "we are many generations away from AGI". However, I do want to point out that (bringing this thread back to the original context) there is substantial harm that could occur from sub-AGI systems.
In the safety literature one frame that is relevant is "Agents vs. Tools/Oracles". The latter can still do harm, despite being much less complex. Tools/Oracles are unlikely to go Skynet and take over the world, but they could still plausibly do damage.
I'm seeing a common thread here of "ChatGPT doesn't have Agency (intention, mind, understanding, whatever) therefore it is far from AGI therefore it can't do real harm", which I think is a non-sequitur. We're quite surprised by how much language, code, logic a relatively simple Oracle LLM is capable of; it seems prudent to me to widen our confidence intervals on estimates of how much harm they might be capable of, too, if given the capability of interacting directly with the outside world rather than simply emitting text. Specifically, to be clear, when we connect a LLM to `eval()` on a network-attached machine (which seems to be vaguely what OpenAssistant is working towards).
I agree with you that it could be dangerous, but I neither said nor implied at any point that I disagree with that--I don't think the original comment was implying that either. LLM could absolutely be dangerous depending on the capabilities that we give it, but I think that's separate from questions of intentionality or whether or not it is actually AGI as we normally think of it.
I see, the initial reply to my G(G...)P comment, which you said was spot on, was:
> That would only be possible if Sydney were actually intelligent or possessing of will of some sort.
Which I read as claiming that harm is not possible if there is no actual intelligence or intention.
Perhaps this is all just parsing on my casual choice of words "if it was able to make outbound connections it very well might try.", in which case I'm frustrated by the pedantically-literal interpretation, and, suitably admonished, will try to be more precise in future.
For what it's worth, I think whether a LLM can or cannot "try" is about the least interesting question posed by the OP, though not devoid of philosophical significance. I like Dijkstra's quote: "The question of whether machines can think is about as relevant as the question of whether submarines can swim."
Whether or not these systems are "intelligent", what effects are they capable of causing, out there in the world? Right now, not a lot. Very soon, more than we expect.
I don't believe AGI needs to have actual consciousness in order to functionally be AGI, and I personally am not of the view that we will ever make a conscious computer. That said, intentionality could certainly impact the way it operates, so it's something I think is worth keeping in mind for trying to predict its behavior.
Just because they aren't "designed" to have them doesn't mean that they actually do not. Here's a GPT model trained on board game moves - from scratch, without knowing the rules of the game or anything else about it - ended up having an internal representation of the current state of the game board encoded in the layers. In other words, it's actually modelling the game to "just predict the next token", and this functionality emerged spontaneously from the training.
So then why do you believe that ChatGPT doesn't have a model of the outside world? There's no doubt that it's a vastly simpler model than a human would have, but if it exists, how is that not "something like a mind"?
It was not trained to model the game. It was trained to predict the next token based on a sequence of previous tokens, which it wasn't even told are moves in a game, much less how to parse them. And it came up with an internal model of the game based on that that's accurate enough to include the board state. You could say that it "understands" the game at that point, even though it wasn't specifically trained to do that.
I can believe that they've been tracking it for awhile, but it is unlikely that they can easily shoot it down. We probably literally don't have the technical capacity to do so reliably at this time.
Qualia are one of the more interesting subjects in the study of mind because they're inextricable from our perceptions and our experience of everything, but they can't be reduced or quantified in any meaningful way.
(If anyone is tempted to respond to this with anything from Daniel Dennett, please consider citing a serious philosopher instead.)
The question you're responding to is (although of little if any practical importance) a more philosophical one than can be answered by cortex differentiation, which cannot tell us much at all about the experience of qualia.
The quantifiable characteristics of the light spectrum are the same--the experience of the qualia are not necessarily the same. This is common with taste and smell; it could easily be the same story with the experience of color.
This feels like a "god of the gaps" religious argument. Every time we learn more the degree to which two people's perception of color can vary shrinks as we find them functionally identical at every level of our understanding. Will it take the invention of a machine that can perfectly read a human's mind to convince armchair philosophers there are no more shadows for it to hide in?
Your analogy doesn't work and I suspect it's because you may fundamentally misunderstand what qualia actually are--this isn't about gaps; this relates much more to the logical impossibility of reducing qualia to something that can be understood through quantitative methods. You dropped the key in the dark and are looking for it under a beam of light three feet away by trying to claim that any any amount of neuroscience could answer a question of this nature.
That said, much like the problem of induction, its insolubility is not necessarily of much practical importance day-to-day. We know that certain wavelengths will be recognized by people with normal visual faculties as "red"; whether the way they experience that qualia is different or not is not of much practical importance. It's more of a reminder of the limits of our empirical knowledge.
I don't think you can get good at doing calculations without practicing the calculations, but reading books that discuss the higher-level aspects of math and the philosophical underpinnings can help you look at it in a different way that may inspire more interest as well as an easier time grasping the difficult parts.
None of this is to say that solar power is bad, just that we should have measured expectations.