They say $0.01/kWh is the target price, to be reached after some decades.
Don't get me wrong, I am excited about solar power but careful about the economics: the capital cost of solar right now is well over 1$/W (panels+inverters+installation/hookup) and even though it is falling nicely, the amortization schedule needs to be considered.
A rule-of-thumb figure is 1kWh of power per year from 1W nominal installed, so the capital cost will have to be amortized over 100 years to reach $.01/kWh. The installed price has to come down by a factor of 10 for this to work out.
No, $0.01/kWh became possible in 2024 and is easy to achieve in 2026. But you have to follow the Fiberhood procedures based on our simulations and our electronics to actually achieve prices under one dollar cent per kWh. We reach this price in China, most of Europe and Australia but in the US you generally pay thrice as much (in labour cost, tarifs, etc).
You can only get an accurate cost if you create a simulation of every step of the industrial processes of manufacturing the silicon ingots, the glass, aluminum, the silver and labour that goes into making solar panels. Same for batteries, electronic components, etc. You have checks and balances in the simulation, for example you get the cost price of all the material components that you can check against the actual price for sale at the factories, the shipping cost, the wholesale prices on offer, the retail prices in different countries, the installation cost, the underlying loans and their interest rates and labour. But that simulates just the cost of the materials, you have many other factors. For example did the energy used to make the solar cells come from solar or from coal plants? Did you make thin film solar or silicon wafer solar cells. What battery chemistry. How much losses if your solar panel overheats 3 percent of the time. What latitude and longitude did the solar panels operate, at what angle to the sun?
Compared with such accurate cost simulation models calibrated with actual prices paid your claim is very vague and hand-wavy.
The Levelized Cost of Energy (LCOE) is a metric representing the average total cost of building and operating an energy-generating asset over its lifetime, divided by the total energy output produced during that period. It serves as a, "break-even" price per unit of energy (e.g., $/MWh), allowing comparisons between different technologies.
I don't doubt that you can have a low marginal cost, but you still have to invest money to get this system installed. You dismissed my argument as 'vague and hand-wavy', so please give a concrete estimate how much it would cost to install a 10kW system on a typical house in the US midwest, and maybe in Europe or Australia.
25 400W solar panels at $0,08/Wp =$800.00 solar panels bought in bulk
metal poles and brackets $275
10 EVE 334AH 3.2V= 10,680 kWh $70.30 =$700.30 batteries bought in bulk
Enernet Power Router components $290 (I left out 2/3 of components because this is such a small installation and needs no 4x25 Gbps internet)
Cables $230,50
Metal enclosure $61
Install fee 5 hours two people $23 per hour = $230
==============================
$2,587.80
So your "capital cost of solar right now of 1$/W" is 0.26$/W
10000 x $1W nominally installed =10000 kwh per year times 30 years = 3000,000 kWh total power yield
$2587.80/300,000=$0.00862333 per 1 kWh LCOE 30 years
$2587.80/500,000=$0.0051756 per 1 kWh LCOE 50 years
These numbers are of day prices of today. Currently the battery and solar prices are high because of the high oil and gas prices of the US/Israel/Lebanon/Iran/Ukraine/Sudan wars, there is a steep battery and solar panel price rise.
The overal LOCE is still hand-wavy in that several costs might happen in the 50 year of operation:
Drop of 20% solar panel yield by breaking glass, cost of cleaning the panels, cable replacement, accidental electronics wear en tear, extra maintainance labour cost.
Also, I doubt you need a loan to pay for $2582.80. If you did at a compound 3% interest rate with you would pay over 50 years, $11,344.67, more than 4,3 times than if you payed up front (from an online calculator, if you payed off per month than just the interest it would be lower).
But whoever is right, you'll certainly stay under $0,01 per kWh from solar and round trip storage cost in batteries.
If you are lucky, it will be almost half that, $0,005 per kWh.
To heat your apartment and drive all over the American continent with your electric car you'll need more than 10kW for an average US midwestern house.
Thank you, great info. You have great suppliers though: I can get close on batteries (I see four for $355, so $88 instead of your $70) but for panels the best I can get is $220 instead of your $32---how do you go about getting them at those prices?
Also, what Ethernet Power Router for $290? Is that a Fiberhood product? How and where would I order it?
>how do you go about getting them at those prices?
Buy a 40 shipping container with aproximately 770 panels in the factory loading lot, truck to Shenzen harbor, ship to a port in the USA, Belém/Natal/Manaus or Rotterdam, Our own truck to the customer.
Please, where do you get your 4 LifePo4 batteries for $355?
I only have one source for $70, I want more sources.
Edit: Thank you, I just ordered 8400 of your batteries (see h3lp his comment's Google link below) for less than $600,000 (at a discount). Picking them up in two weeks.
My company product, you'll get the datasheets and video as part of the order contract. See my white paper for the old model, its much upgraded since.
>Is that a Fiberhood product?
A Morphle and a Fiberhood product. Morphle is the company mass-producing electronics and chips, including the chargers, dischargers, mppt a.k.a. the Enernet Power Routers. Fiberhood cooperative sells and installs the complete infrastructure systems: Enernet systems and fiber optic internet (since 1987, we were one of the first internet providers), batteries, solar panels, power routers, tiny houses and water tanks.
>How and where would I order it?
With me in Europe or Ukraine or our office in Tucson or Montreal. +31617428596 Signal, Facetime, Whatsapp, Telegram.
You can also bundle your solar panel or battery bulk orders with us, we'll ship them directly to you in the US from the Shenzen factory region or the Poland or US wharehouse. We won't charge you a profit margin, we get a larger joint order and that gets us both a discount, less chipping cost and lower insurance. We have an agent in Shenzen overseeing the loading at the factories.
$ sqlite :memory:
create table t (product,revenue, year);
insert into t values ('a',10,2020),('b',14,2020),('c',24,2020),('a',20,2021),('b',24,2021),('c',34,2021);
select product,sum(revenue) filter (where year=2020) as '2020',sum(revenue) filter (where year=2021) as '2021' from t group by product;
Good insight, but if you discount the visual elements (tabs, buttons, etc), you're limiting TUI to CLI, and I think that's unwarranted. The value proposition of both TUI and GUI is two-fold: you see the available action options, and you see the effect of your actions. So, yes, TUI and GUI _are_ closely related: who cares whether we're displaying pixels or character blocks.
Unfortunately, they are often artificially differentiated by the style of the UX interaction: TUIs promote the keyboard actions, and GUIs prefer mouse without corresponding keyboard shortcuts. Unfortunately for GUIs, their designers are often so enamored with WIMP that they omit the keyboard shortcuts or make them awkward. I hate it when, even if the ACTION button is available by keyboard traversal at all, it requires some unknown number of widget traversals instead of being one tab away.
Since the keyboard is almost always used for the textual data, it makes sense to me to always enable it for command execution. Well designed GUIs and TUIs provide both WIMP and keyboard UX, which sadly is not the norm today, so here's my vote to make them larp for each other more.
Any competent computer engineer can design a much better ISA than RISC-V.
Hello, my fellow bitter old man! I have to respectfully disagree, though. Firstly, RISC-V was actually designed by competent academic designers with four preceding RISC projects under their belt. The tenet of RISC philosophy is that the ISA is designed by careful measurement and simulation: the decisions are not supposed to be based on gut feeling or familiarity, but on optimizing the choices, which they arguably did.
Specifically, about detecting the overflow: the familiar, classic approach of a hardware overflow (V) flag is well known to be suboptimal, because of its effect on speculative and OoO implementations. RISC-V has enough primitives to handle an explicit overflow checking, and they are consistent with performance techniques such as branch prediction and macro fusing, to the point of having asymptotically vanishing cost--there can be no performance penalty. Even more so, the RISC-V code that does NOT care about overflow can completely ignore these checks.
A lot of computer users are domain experts in something like chemistry or physics or material science. Computing to them is just a tool in their field, e.g. simulating molecular dynamics, or radiation transfer. They dot every i and cross every t _in_their_competency_domain_, but the underlying code may be a horrible FORTRAN mess. LLMs potentially can help them write modern code using modern libraries and tooling.
My go-to analogy is assembly language programming: it used to be an essential skill, but now is essentially delegated to compilers outside of some limited specialized cases. I think LLMs will be seen as the compiler technology of the next wave of computing.
The difference is that compilers involve rules we can enumerate, adjust, etc.
Consider calculators: Their consistency and adherence to requirements was necessary for adoption. Nobody would be using them if they gave unpredictable wrong answers, or where calculations involving 420 and 69 somehow keep yielding 5318008. (To be read upside-down, of course.)
But thats the point, an llm is a vastly different object to a calculator. Its a new type of tool for better or worse based on probabilities, distributions.
If you can internalise that fact and look at it like having a probable answer rather than an exact answer it makes sense.
Calculators cant have a stab at writing an entire c compiler. A lot of people cant either or takes a lot of iteration anyway, no one one shotted complicated code before llms either.
I feel discussion shouldnt be about how they work as the fundamental objection, rather the costs and impacts they have.
Greg mentions discipline and vision as determinants of successful software, which is correct but I think he misses another aspect of vision: the ability to attract and crystallize a community around their project. Arguably, most successful softwares thrive in the long term because they have a team of people that inspire each other, fill in with complementary talents, and provide continuity.
Great example, and a lost opportunity in the interview---they should have asked "What are the requirements that would invalidate this answer? and what would you design if the requirements were changed in this way?". Maybe even "how long is the runway for your Progress solution if we consider future scaling up of the requirements"
The alcoholic knows the bad outcomes, and chooses to ignore them. The hapless Android user does not understand the negative consequences of sideloading. I think this makes for a substantial differerence between those two.
> The hapless Android user does not understand the negative consequences of sideloading.
Then make sideloading disabled by default but enable it when the users tap 7 times on whatever settings item. At that time, explain those "negative consequences" to them, explain them real good, don't spare anything and if they still hit "Yes, continue to enable sideloading" you do that immediately in order to avoid increasing their haplessness with other made-up excuses.
I don't see how people are against this. Especially tech-savvy people who browse HN. It really seems to me like everyone here who's on Google's side is just a bot in a botfarm somewhere. they can't possibly be real
Well, power envelope IS the limit in many applications; anyone can build a LOBOS (Lots Of Boxes On Shelves) supercomputer, but data bandwidth and power will limit its usefullness and size.
Everyone has a power budget. For me, it's my desk outlet capacity (1.5kW); for a hyperscaler, it's the capacity of the power plant that feeds their datacenter (1.5GW); we both cannot exceed Pmax * MIPS/W of computation.
Literate programming is not about programming in natural languages: it's about integrating code (i.e. the formal description in some DSL) with the meta-code such as comments, background information, specs, tests, etc.
BTW, one side benefit of LP is freedom from arbitrary structure of DSLs. A standard practice in LP is to declare and define objects in the spot in which they are being used; LP tools will parse them out and distribute to the syntactically correct places.
Well I think the ambition was to have as much as possible in natural language, with macros calling out to ‘hidden’ code intended for machines. So I do think there is a good link with later attempts to write using natural language and make computer languages more human-friendly and he was one of the first to have this idea.
Don't get me wrong, I am excited about solar power but careful about the economics: the capital cost of solar right now is well over 1$/W (panels+inverters+installation/hookup) and even though it is falling nicely, the amortization schedule needs to be considered. A rule-of-thumb figure is 1kWh of power per year from 1W nominal installed, so the capital cost will have to be amortized over 100 years to reach $.01/kWh. The installed price has to come down by a factor of 10 for this to work out.
reply