Project Sunroof | Hacker News Reader

Project Sunroof

sunroof.withgoogle.com

176 pointskaycebasques104 comments

This is a cool project, but it hasn't been updated in YEARS. I think it was put on lifesupport in 2017. https://www.greentechmedia.com/articles/read/googles-project...

Less of a moonshot and more of a sunshot.

That article is only about the lead-generation part of Project Sunroof being shut down. Which makes a ton of sense - it was likely not generating much revenue, but managing a program like that can suck up a ton of time in managing those relationships. It's just easier to kill it and let everyone compete in search results instead.

You're correct in that the design hasn't been updated in years, but... so? From poking around in my city, everything still seems to work the same. (Though I wonder if the estimates are now out of date...)

> You're correct in that the design hasn't been updated in years, but... so?

I checked a house that was built about 4 years ago in a major metro area, and Project Sunroof just showed an empty field and "sorry, we haven't reached that address yet".

I'm fairly sure nothing about this site is being updated. That also means that any structural changes to houses (including additions, remodeling of the roof, or just complete bulldoze-and-rebuild) or the surrounding property (tree growth, trees being cut down, etc) are not reflected in the estimates.

> Though I wonder if the estimates are now out of date

By all appearances, they are. Massively. The older this data gets, the less useful it becomes. It's not really a "so what?" situation if the data isn't being kept up to date.

Ah, interesting. So it sounds like they haven't updated the data in ages, either.

What I always loved about this site is that Google made it SUPER-easy to visualize just how good (or not) any given roof was for solar. At the time it launched, you had to get deep into the sales process with a solar vendor to get that kind of data/understanding.

I just tried a couple of big current-day providers and it looks like that's largely still true. Pick My Solar was the only one who would even try to give an estimate of solar potential without requiring an email address or full-on lead gen.

So at least for the people whose homes haven't changed in the past 5 years, some of the core visualizations will still be useful.

Probably that they were being out-classed by everyone else in the solar lead-gen industry on their own platform. The money involved probably wouldn't move a needle for Google though. I am glad they didn't pull a move and squash the competition by de-ranking competitor sites. Kind of a shifty industry in general, lead-gen, I am surprised Google was playing in it at all.

Damn, thanks for pointing this out. There's a new construction near me that could have a big impact on solar viability of my roof. I figured I'd wait until Sunroof updated to reflect that before considering doing my roof. Guess that's not going to play out the way I hoped.

If you scroll down it asks you to select your average monthly electric bill.

The fact that the maximum value is $500 is telling of how long it has been since this was updated.

Expect this project to be cancelled. Google is in layoff mode.

I don't know why this guy is being downvoted. I thought this already was canceled, as (almost?) all 20% time projects at Google have been killed. Last time I heard about this being relevant was, 2015? ish?

Because it is a repetitive, low-value comment which gets spammed to basically every thread that's even tangential to Google.

Sort of? It's still a valid comment as long as it continues to be true. Google needs to clean house at the top, and go back to engineering-first teams, not engineering-never.

A lot of things can be true, and still not be worth posting thousands of times per year on HN.

Because do you see what's happened here? Whatever interesting discussion there might have been to be had on this project has now been smothered by the top-voted comment being just a tired repetition of the same "Google kills products, hehe, hehe" echo chamber. (I don't know if there really is that much discussion to be had on the actual topic of the post, but it should at least get a chance!)

I tend to prefer an occasional, short, low effort comment more than the long-winded, hall monitor style commentary on hacker news guidelines which so often follows.

It's not true

Source: I work at Google

Can't blame someone for being wary of non-search/ad services offered by Google https://killedbygoogle.com/

Sure, I’m just saying 20% time is still a thing though

> Source: I work at Google

For how much longer? Stadia devs sure thought they were in for the long haul.

If they actually knew it to be cancelled then it would be useful to share that. Saying "it might be cancelled" adds no value. You don't need to know anything or even need look at the web page. Nothing is learned.

Google quiet kills services often.

Nobody at the company is tasked with actually working on it, the creator often no longer works at there, the team that built it often either no longer works there or has moved to another team that does something entirely different.

After a few years, they officially axe it, and it gets added by the KilledByGoogle list; but it was clearly already dead and shuffling around like a zombie.

It's more evidence of how bad Google has been at building diverse revenue streams.

Search is going to be driven to zero profit, and that's going to cripple Google.

Literally everyone is empowered to build better products than Google search. And Microsoft / OpenAI will subsidize it, laughing all the way to the bottom. It'll completely poison 50% of all the money Google can make.

Google had two decades to ensure they were safe, and stuff like Killed By Google / Project Sunroof is all they've got to show for what their product culture can build.

Google is going to try really hard to reinvent themselves, but they haven't accomplished anything along that front in the last two decades. I wouldn't bet on that magic happening overnight. Their founders have been absent and their leadership has been coasting. The picture is clear as day.

Google's done. Google got checkmated hard.

people just get on here and say nonsense like this.

Dude, I'm living in my office working on this stuff 24/7. I quit my $400k/yr TC job because I believe in this stuff so strongly. You have no idea the wave of disruption coming. I'm seeing fundamental breakthroughs every single day. AI Twitter is simultaneously incredible and haunting.

So many people are sleeping on something bigger than the internet itself. This is huge. Pay attention.

Google should be scared to death. They're in a precarious position and ripe for disruption. Even adapting to the new world order puts their primary revenue stream at risk.

Think about it and then think about how you might be able to take advantage of this. Fortunes unlike any before are about to be made.

I can't wait to see this "bigger than the internet."

AGI is bigger than the internet. We're somewhere along that path right now, and I suspect we're closer than we realize.

I personally don't think we'll have AGI in my lifetime.

> You have no idea the wave of disruption coming.

Cool, but can you give some concrete examples so it doesn't sound like you're being extremely hyperbolic?

> Even adapting to the new world order puts their primary revenue stream at risk.

Even if Bing and Google end up having similar LLM chat capabilities with Bing being first to deliver, would it be enough to get Google users to move to Bing?

The current business models need to be more than being able to converse with a chatbot. Ads will still need to be integrated and they might come off a lot worse when integrated into a conversation. It's definitely interesting times but I don't see how it's a new world order yet.

How is Google sleeping on AI? By all accounts, they have had a better LLM than ChatGPT (Lamda), years in advance of ChatGPT.

Google is hampered in that they have a lot to risk. They're a big lawsuit target that can't go around wildly violating copyright and subjecting the world to dangerously, convincingly confident BS.

But make no mistake - they are world leaders in AI, they have their best engineers working on it, they beat everyone to the punch with a high performing LLM amongst a mountain of other AI advances, and they recognize the threat to their business model. You're kidding yourself if you think otherwise.

It doesn't matter how good they are or how much institutional knowledge they've acquired. 50% of their revenue stream is being existentially threatened, and nothing will stick around if that vanishes. Talent will be plucked, internal projects will wither on the vine, and the velocity of the entire company will seize up.

It doesn't matter that Google was the Bell Labs of its time or that it was first and best at AI. Bell Labs also got sunk by disruption despite being the best in the world.

Google was also best at cloud once upon a time (Borg/k8s/spanner/...), and you saw what happened there. This is looking like a repeat.

Let's agree to come back here in 12 months. If Google is dead, I'll buy you (and myself) several beers.

They won't be dead, but pay attention to cash flows and market cap.

I'll also give it 24-36 months to play out.

It's too early for me to start buying puts, but I'm very eager to start. It's not just my investment thesis, but my worldview.

You need to take a break from social media and from your computer.

I think this should be turned around. You should pay attention to what's happening.

99% of what you consume online is just noise, man. You're clearly driving yourself mad. This isn't concern trolling, I have the same problem.

Honest question. Do you think every product reachable from a google URL is staffed?

https://www.google.com/moon/ ?

I will never trust a Google product again for the rest of my life.

I wish this was flamebait hyperbole that I could downvote, especially as an ex-employee, yet here I am now avoiding Google as much as possible. Transitioning to providers like Fastmail and my own home cloud; even switching to Kagi for search. Sad times, but probably for the best

If it's not ad-related, it will probably disappear at some point.[1]

The big worry is Google Cloud Services, which lose huge amounts of money. Google Stadia already shut down.

I just hope Waymo survives. It's close to working, but nowhere near profitability.

[1] https://en.wikipedia.org/wiki/List_of_Google_products#Discon...

This project was launched in 2015. https://spectrum.ieee.org/google-introduces-project-sunroof

I entered an address. It said $14K savings for 20 years. It takes $15K to install solar roof. So break even is like 20 years - which is half the productive life time. It doesn't motivate me to install Solar panels.

You also get 30% back in the form of a tax credit when you purchase solar. So that might mean that the cost is only $10k. I'm also not sure if the Sunroof website is just doing straight math based on today's electricity rates or if it assumes some amount of inflation. If the past predicts the future your rates could be 75%+ higher in 20 years making solar much more lucrative.

I researched historical rates in San Diego (upcoming solar install prework) and prices have increased an average of 5%/year for 10 years running. This felt like a reasonable assumption to me, although to be cautious I used 3%.

Rates will change! Energy doesn’t look like it will get cheaper, especially with EV fleet rollouts.

Accounting for planned 2 EV purchases at 10k miles/year, the largest system I can build on my roof has something like $150k savings against <$50k cost over 20 years.

That gives a lot of room for unexpected issues and feels like a wise bet to take.

Yep, I missed counting energy rate increase. Would the output of solar panel degrade over time? Or is it constant?

It degrades and the amount of degradation is dependent on the PV technology used, but you can find the percentages online. You should still be at 80+% of the original output by the time you retire your panels (20-30yrs later).

It depends on what your electricity price is. For our house, it's way higher than the estimate (and they let you bump up the estimate). The cost to put panels on was pretty spot on to the quotes I've received.

Sounds like it worked, then.

It might be assuming no energy price increases. Given that energy e.g. increased in many german cities from 30ct/kWh to > 60ct/kWh, it might break even much faster.

On the other hand it might however also not include maintenance costs.

In that case, you should probably account for the time value of money and the depreciation of your solar roof. 15,000$ today is worth almost 50,000$ in 20 years at 6% interest rate (long term stock market returns are about 7%). Let's say the solar roof depreciates by half in 20 years and produces 14,000$ worth of electricity. You end up with 14,000$+7,500$=21,500$. Energy cost would have to increase a lot to make the solar roof a wise investment. Also, a quick Google seems to indicate that the inflation adjusted cost of residential energy tends to go down over time (e.g. nominal price goes up, but not as fast as inflation).

Stock market is not guaranteed to gain 7% over the next 20 years, but you can effectively guarantee your electricity usage is probably going to stay the same or go up and can somewhat guarantee the performance of the panels (given the warranties).

It also is a useful form of diversification. If you only had $15k, it might not be the best overall investment, but if you have more money than that lying around it probably isn't a bad place to invest some of your money assuming you plan on using lots of electricity over the next 20 years. At some point having all your money in the stock market isn't the best money allocation either.

Agree it's not the most exciting (if it was, everyone would have solar.)

But there's a bunch of other ways to think about it: you save about a K a year on average. Not a ton but not bad.

You are hedged against electricity pricing going up.

You are somewhat protecting against electricity outages in your area.

I think panels keep your roof cooler.

Plus tax benefits.

I don't think the map sources it uses have been updated in a long time. We've had a tree removed from the front of our house 2 years ago that clearly shows up in the map.

Oh cool, my Master’s thesis, reproduced with the resources available to Google. I’m feeling pretty darn validated right now. Floating on clouds. :)

Link (there’s a level of bravery in sharing this that I didn’t have before. I’m not a great writer): https://uwspace.uwaterloo.ca/handle/10012/7499

Link? Sounds neat.

Updated parent comment with a link.

Cool site, but its been around for years.. not seeing anything new on it today?

I submitted the post. I am a Googler but am not aware of any news about this project. Just thought it was cool and would share it in case it's useful for anyone.

And correct, it has been around for years (2016 to be exact): https://sustainability.google/progress/projects/project-sunr...

Unfortunately, the data has not been updated in years either.

If you're a Googler and you find this project interesting, maybe you can find the right people internally to work with to revitalize this project. It's a really cool concept, just so far out of date that it isn't very useful any more.

This varies quite a bit from the answers my utility company (SMUD) gives. Sunroof says I'll break even in 6 years, my utility estimates 19 years.

The biggest difference between their estimates is that Sunroof thinks the panels can cover 99% of my energy costs, so I'll only need to spend $6,000 over the next 20 years on electricity. My power utility estimates that the panels can only cover 80% of my electricity, and I would still spend $26,000 over those 20 years.

Just in that comparison, I trust my power utility more (99% coverage seems unrealistic). If they're otherwise correct, 19 years to break even is hardly a compelling case for installing solar. And my roof is well-suited for solar too.

FWIW your utility company probably doesn't want you to get solar, so the answer is most likely in the middle of the two.

Had this in my city for a time. Was a partnership between some tech company and the big local energy provider. Got axed after two years and seems like the tech company didn't have any rights to the data since it's no longer accessible. Cool idea but miss my local data :(

Found the company, they still partner with a small amount of cities: https://myheat.ca/

See also the independent company energy sage (https://www.energysage.com/). They provide what seems like a similar service, and have an active referral network in some markets.

I got several quotes through them (though decided to hold off on solar for the time being last fall).

>Enter a U.S. or Puerto Rico address

Enters address in top 50 (population) US county

"Project Sunroof hasn't reached that address"

...well, alright then.

Was your house built in the last, what, 7-8 years?

(I'm curious if the data hasn't been refreshed since the site launched.)

My house was built 71 years ago.

My hunch is that they require LiDAR data for the estimate and LiDAR data is currently collected one municipality at a time and stored in many different places. It's a data fragmentation problem.

In comparison, ground elevation data (omitting buildings and trees) can be collected using radar from orbit. That's why we have consistent and widely available elevation data like SRTM for the entire planet.

There's pretty good high-res LiDAR data available for my location (84741), but no availability of Project Sunroof. So either they're not aggregating all available public data (certainly possible), or something else is holding them back.

FWIW this project is used by companies involved in solar installations all over the US and is instrumental in furthering the adoption of residential solar. it's kind of a big deal. that said, it's more of a jumping off point for better estimates, and many other companies are around that do this now.

Google's Project Sunroof - https://news.ycombinator.com/item?id=25548270 - Dec 2020 (2 comments)

Project Sunroof - https://news.ycombinator.com/item?id=11941520 - June 2016 (4 comments)

Google Project Sunroof - https://news.ycombinator.com/item?id=10879312 - Jan 2016 (1 comment)

Project Sunroof - https://news.ycombinator.com/item?id=10073724 - Aug 2015 (223 comments)

They say they account for trees when calculating usable roof area but I can't find if they account for building shading in the total hours of sunlight.

I've been working on a tangential project [1] for helping people find sunny areas anywhere in the world based on topography and building data and must say that Google trove of LiDAR data is amazing. Does anyone know if it's accessible to developers?

[1] https://shademap.app/@47.60502,-122.33437,15z,1675804770099t...

Putting solar panels on roofs may have made sense when PV was expensive and the component cost was a big multiple.

But now? Far better to just stick a bunch of them in the desert where they can be inexpensively installed and maintained -- pass the savings along to consumers. This saves a lot of pain when people "invest" a bunch of money in labor intensive installations, only to figure out the difference between wholesale and consumer rates when they cannot sell power back to the grid at $.20/kWh.

At a macro scale, there is no reason to incentivize residential solar, other than "feels."

Payback period in most parts of the US is ~10 years for residential rooftop solar. The equipment will last at least 25-30 years, most likely longer at a lower capacity rating (panels degrade ~0.5-1%/year). It also prevents your utility from holding you hostage economically, because they'll just keep raising rates at the max rate the PUC allows and to max their equity return permitted.

I'm unsure why you think savings would be passed to ratepayers versus kept as profit by investor owned utilities? Might happen with coops and non profit utilities. Utility scale solar absolutely helps drive out fossil generation at the macro, but residential rooftop solar is a hedge against utility monopolistic practices.

https://www.utilitydive.com/news/electricity-prices-inflatio...

https://www.utilitydive.com/news/retail-electricity-prices-c...

https://www.wptv.com/money/consumer/florida-residents-strugg...

This tool, while cool, is an application of large-scale image processing and won't stand scrutiny over the estimates for individual buildings. Just as an example, it features Oklahoma City, but their model does not seem to include the mean time between being destroyed by hail or tornados. This would be a somewhat relevant point for OKC.

Rooftop racking strengthens roofs (racking is mounted to roof trusses with asphalt roofing, or friction mounts on seams are used with standing seam metal roofs), and panels are rated to withstand hail (1 inch hail at 50 mph speeds) and hurricane force winds. Homeowner's insurance will cover replacement of damaged panels and systems in the event of a loss. A few minutes with a search engine of your choice will confirm these facts.

A little searching would also show that the installed capacity in Oklahoma is just 1% of that in Arizona and I think there's a reason for that. Even of your insurance covers it, it's not like the insurer is operating a steady-state loss. You're paying for it, and that factors into an economic estimate.

If your panels are destroyed, you're going to get the replacement cost back from home insurance along with your roof repair claim. Unless your panels are destroyed in year one, your deductible will more than have been paid for by the panels' production.

No reason to prevent people from putting solar on their roofs if they wish.

I'm talking about tax rebates and credits given to people who put solar on their roofs. I'm sure these are good investments: they're subsidized!

I had the local solar reseller estimate for installing on my flat roof house. I would like to think I am keen on going green, but after three visits, some hard sell and spreadsheets showing when and how I would break even, i really felt like I was just engaging in a long-term finance deal. I would pay the same as I currently do, but to the solar finance lender, not to my local electricity company. And some of the cost benefit was speculated based on my local electricity company buying from me. I also had concerns about roof maintenance when it was inaccessible due to the panels. I also worried that to sell my house I would have to find a buyer willing to take-on the debt, the reseller unrealistically insisted the opposite, that this would add value. I didn't sign up, having seen companies fold, be acquired, change terms and conditions and essentially screw consumers, it just wasn't appealing.

This is exactly my experience as well. I got a lot of hard sales tactics that all sounded like "Pay us instead of your power company." I see no incentive to do that honestly, it just leaves a bunch of pain and complicated math for me to take care of and they eat up all the savings. From what I can tell, just like recycling, it is almost useless unless we all group together and do it at once. In this case, it is better for my power company to innovate that each individual take on the problem themselves.

> it is almost useless unless we all group together and do it at once.

There are coops that perform group buys if you don't feel sophisticated enough to compare bids and select an installer (which is a reasonable position to take): https://www.solarunitedneighbors.org/

https://reddit.com/r/solar is also a resource for having bids reviewed, asking for system and install advice, etc.

(i collectively own ~100kw of residential solar and provide guidance gratis to others constructing or purchasing their own systems)

That the equipment has a useable date of 25-30 years is why most people won't bother with it. Most people don't want to be getting near retirement age and replacing obsolete equipment with a 15 year break even replacement cost on their homes.

Easy solution to that, just mine crypto with all your excess solar power.

A few reasons PV on roofs might still make sense:

* Cheap land may be far away from consumers, requiring expensive and inefficient transmission lines.

* As PV costs come down, land cost become a higher fraction. Roofs do not require a land purchase.

* Some locations are politically unwilling or unable to do large scale solar. rooftop solar provides a route for personal adoption.

Finally remember the two options are not mutually exclusive!

Roofs have the distinct advantage of not requiring new land acquisitions and being close to consumption (cutting transmission costs).

That doesn't mean that problems that can be solved at the macro scale should only be solved at the macro scale. If, for whatever reason, the best solutions aren't being implemented that doesn't mean no solution should be implemented.

Building solar farms in the desert requires expensive grid upgrades.

Putting solar panels on your roof generates the electricity near where you consume it, which is a big advantage.

Rooftop solar can have payback periods as short as 5-10 years depending on the location. It makes a lot of sense for individuals to embrace that if they want, while utilities can also do their thing in the desert.

I used to agree with you, because if you look at the system as a whole, centralizing generation is more efficient. But I am currently paying 2/3 of my per kWh rate just for distribution. PGE (the northern california utility) is some combination of negligent, corrupt, and standard fare profit seeking, that I simply don't trust them to pass on the efficiency savings to me.

Until someone figures out superconduction at room temperature, there will always be transmission losses, which you don't get when you generate and store the power at the point of consumption.

Rooftop space is wasted space, it should be used for something. I'm not against solar in the desert, but at the same time deserts are actually beautiful and sensitive ecosystems. If there's space on roofs that can reduce the impact on wilderness, then we should use it.

Branding that as "feels" is dismissive to rational debate as in that line of reasoning there is no reason not to sell yosemite or joshua tree national park to developers other than "feels"

LOL/cry that it doesn't let you enter a current electric bill above $500, but does include Puerto Rico and Hawaii adresses.

Arcadia’s Arc platform also does this, except it uses actual energy and tariff data, and is used by every major solar installer in America to produce savings estimates accurate enough to write binding contracts against. Check it out: https://www.arcadia.com/arc

I hope someone has already written the eulogy for this service.

Their numbers also appear to be off. I compared my property to a house that has over 4x the amount of roof space for solar cells and their calculations only showed a 67% increase in the money recouped after 20 years.

Amusingly they don't appear to be using Google Maps for the data, or if they are, it's an old dataset. They still have the old image of my house before I remodeled, but GMaps has the new roof.

Use https://solcast.com/ has a free version no credit card, also had it hooked up to homeassisant via their API.

Weird. My address shows a pin on the map, with the house missing. Both neighbors are there, and gives valid estimates for them.

This is way more optimistic about my house compared to other websites that provide other information.

To be safe is it better to use a neighbor's address?

That would be completely pointless, since you will get results for your neighbor.

What if it’s in an area with very similar buildings

This is a local-only service for just a single nation.

Not useful for off-grid and no battery options.

A Google thing started then abandoned? Never heard of such a thing!

I wait for a google toilet since years!

So that's Googles best bet how to not being killed by Microsofts AI adventure

What? How do you even make that connection?

Reply on news.ycombinator.com