It’s cool they hired an actual person but if I was so happy with their work to write-up a whole article about it, I’d be crediting them in a much more prominent way!
Wayback Machine confirms it was there a few days ago.
People are just never happy and always going to have something to say. Even if you don't go the AI route and hire an illustrator, even if you credit work, someone has to be critical.
The obvious place to do it is in the writing. I had been expecting that during the entire paragraph, being art and liking what I was seeing I just wanted to see the artist portfolio.
In this day and era such a thing will feel like yet another ad. We react on a very defensive way against ads.
The way it is actually done, it generates genuine interest, and you have the link to the artist at exactly the point where your interest is at the highest.
As the spouse of an illustrator, I applaud the author for hiring an actual artist. We've seethed at the way artists work was scraped online for the training of these models, recreating their very style for the creation of these infringement engines.
The finished product benefited greatly. There's nothing quite like working with a genuine expert who knows the rules and tradeoffs on things like competition, color, etc. Author did good on providing the illustrator with a detailed brief. When the illustrator takes those guidelines and applies their style and training to the product, magic happens.
I'm a 3d modeler and a software engineer. AI didn't steal my work. Anybody could have looked at my repositories and art and learned from it. Just because something came along that is a really good learner doesn't make that suddenly be theft.
The case is slightly stronger if you want to talk about training that happened on our private code repositories, but publicly visible art and repositories are fair game for learners to learn.
Most artists I know never had a problem with other artists learning from them. But they do have a problem with corporations training on their data without any sort of consent and then selling a product that takes away their jobs, they never agreed to it and they never would have shared it to them.
Then they presumably are fine with open weights models made by hobbyists by scraping publicly available art on boorus and such, yes? because they are neither made by corporations nor are they selling any product.
(to be clear, I don't think this is an accurate description of how artists feel about e.g. publicly released diffusion models. The core is not that it's made by "corporations" or the for-profit aspect or even the copyright infringement part, it's simply the understandable and not-at-all-unique-to-artists resentment at having your job or passion automated or displaced)
True. they were fine providing advances to their field for others to learn and build from that shared legacy. They didn't agree for their work to be the base of auto-slopping at industrial scale. Quite easy to understand, except for AI-bros.
How is this morally different from e.g. weavers and loom operators being automated or ceramics or anything else? Or programmers?
We've been automating people "out of jobs" since at least the start of the industrial revolution. It's generally considered to be a goos thing.
We had a sort of weird special case for copyright preventing certain types of machine automation and I've heard plenty of complaints about how that worked, but it more or less did work for a while.
I feel like "don't agree" is a weird framing, because it assumes that all technological innovations must have the explicit and affirmative consent of all parties they could directly or indirectly affect. Not to mention that "artists" are not a monolith any more than "AI bros" are.
I could sort of get behind something like this in theory (weighing each new technology based on whether it furthers the values of the community, and rejecting it if it doesn't, kinda like what the Amish do), but I'm not sure it's being applied neutrally or consistently here, nor do I think it works in practice especially for something as decentralized as a diffusion model, which are surprisingly simple to build and train and can run on a home GPU.
I think the reason artists wouldn't like hobbyists doing it either is just that they've been soured on it by corporations. Artists didn't reject AI in its infancy, or things like deepdream. But AI as it stands very clearly is exploitative, and what's more, it's flooded spaces formerly for human art with low quality generic garbage. Having those experiences bleeds over into your feelings about things that don't match the exact definition of the reason you're upset to begin with.
>Then they presumably are fine with open weights models made by hobbyists by scraping publicly available art on boorus and such, yes? because they are neither made by corporations nor are they selling any product.
Why would anyone be fine with people explicitly not actually making art.
Scrapping and generating isn't art and isn't comparable to people tracing/copying/learning from existing art.
Multi-billion-dollar copy-machines are the same as a person just dropping by and having a look? If you can't tell the difference, what are you? Be you human or robot?
...yet. AI has evolved greatly from basic image generation. It's already pretty capable at using tools to great SVGs and 3D scenes that are more reliable than plain images. Sure, experts won't lose their jobs, but how will juniors become new experts?
To be fair, if an illustrator uses these models to write code, it is exactly the same thing. We built everything for decades, got scraped, and now we don't write code ourselves. Everyone uses it to write code, and nobody gets upset. If people use it for art, they're angry. We've both been used.
I think there's something more interesting going on. We (coders) use LLMs to write code and still get paid, because what we bring to the table is context across a multitude of coding tasks.
There are art projects like this, but many traditional art projects are quick one-shot deals. Make a logo, design a business card, make a tshirt design. Bite sized, do the work, get paid, move on.
LLMs are great at short context tasks. A lot of art is short context. Building the software that runs a bank is not.
If you want to work as a professional artist in the age of LLMs, move into long context work. If it isn't already, it will be impossible to make a living off of fiverr.
The really sad thing is that the natural instinct for many will be to screenshot the page and upload it with "please create me a home assistant dashboard in this style".
yes, it's amazingly/depressingly good at things that would take a $1000+ software, and years of education just a year ago. Now I'm doing this with leftover claude/codex weekly quotas( i do have both due to some testings at work).
I designed my basement renovation using a combination of image modelsand a 3D scanned Houzz Pro FloorPlan, and I gotta say it REALLY helped the contractor deliver on the vision that I had.
Decorating when you can easily swap out pieces and parts in images with realistic dimensions makes rapid iteration on physical things much faster and it's easy to figure out what "speaks" to you. Great use of these tools.
I’m genuinely curious how you use image models for this type of design. I need a way for my wife and I to visualize a remodel project and don’t know where to start or which tools are best for the job.
dump the plan, make sure it has height etc, after dump pictures if needed,
refs of colors. atropic has some nice recipes somewhere online, entire prompts you can feed into cc. i used their workflows feature, otherwise it's painfully slow.
do it in two steps, first generate just the named 3d model, walls, stairs etc inspect for correctness, after proceed to the fun part, materials, colors, etc.
I did it in extra one step, not sure if that was needed, i build a local harness first for extra small room 6 ft by 6ft, and after that expanded into real size. It installed lots of tools.
It’s not different . One case you chose tiles , you take a picture of your bathroom , and give it a rough sized of tile + room and it will
Give you visual check . Second option if you want it also to calculate materials . This is a 3d path . First one can be done by just chat , another one is coding agent .
I tried a couple of the web-based design-a-floorplan sites and all of them were kinda like using crayons to draw a circuit diagram. Ended up doing a pile of sketches on paper, then when I had what I liked finalised it in Visio and sent it through to an architect who turned it into proper plans which, since they were architect-drawn plans I had to check for f_ckery and then the builder checked it for stuff I'd missed but ended up with a rebuild that was exactly what we wanted.
Well with one bit of f_ckery that both of us missed until after the concrete had been poured because it was too small to see easily on the plan, and a second also too-small bit that I caught once they'd put the framing up and asked them to take it out again.
It's not sad, it's literally what those tools are for.
Yes, this is the "you wouldn't download a car" moment. Yes, I would.
(And no, I'm not doing it, because I have my own preferences, which involves hitting a very small target in design space, that is applying "dark cockpit" principles and a mix of industrial HIGs and nuclear reactor control room interface standards to LCARS.)
> It's not sad, it's literally what those tools are for.
If someone spreads poison on your street to kill the strays and mutts (yes, that’s a thing that happens) and it kills your dog, would you like someone to tell you “that’s not sad, that’s literally what poison is for”?
Using a tool for its intended use has zero relation to it being sad.
Yes, I do see that problem. I just wish we'd call it what it is instead of drawing imprecise and confusing analogies that are often thinly veiled appeals to emotion.
That would be the first step to addressing the actual issues: Who will advance the field, assuming that generative AI will always be limited to copying instead of innovating? What will be the place in society for the people that lose their job due to these changes? etc.
> I just wish we'd call it what it is instead of drawing imprecise and confusing analogies that are often thinly veiled appeals to emotion.
And I wish people on HN would steel man arguments and not assume intentions in others, but here we are. There was no “appeal to emotion”, veiled or otherwise. The example used harming dogs because hopefully most¹ of us can agree that’s repulsive and bad.
Everything else you’re saying has nothing to do with the point. AI isn’t the point (hint: that’s why I didn’t mention it), the point is spelled out in plain English at the end of the comment:
> Using a tool for its intended use has zero relation to it being sad.
¹ I promise you that qualifier is necessary. I’ve seen some vile opinions on HN.
My comment was specifically regarding your example of harm against animals, yes.
Why do you need to invoke an image of something widely considered repulsive to strengthen your point against something much more nuanced/controversial?
It was not to “strengthen the point”, it was to have a clear example where most of us can be on the same page of what is wrong. This is HN, anything other than an obvious example attracts an infinite swarm of nitpickers arguing that that whatever bad thing “is very good, actually”.
He didn't say it's illegal (even if it's debatable and kinda loophole-ish, but lets pretend for now that society has fully embraced a world-encompassing copyright exception if you're named Altman or Amodei).
He said it's sad, and no matter how you want to spin it, it still is incredibly sad and depressing that we're hailing the age of "semantic photocopiers" as some sort of intellectual achievement.
If you consider it incredibly sad and depressing for the majority of people to have cheap access to things that would have cost a small fortune due to being inevitably hand-crafted only a few years ago, I hope you're renouncing clothing manufactured using automated looms, printed books etc. as well?
Absolutely, I think fast fashion brands like Shein and TEMU-style shops are incredibly sad, toxic for the economy and the environment, and I've never bought anything from them.
Are you arguing that the only type of clothing an automated loom can help produce is fast fashion? Is every piece of clothing you own hand-woven, crocheted etc.?
Man I hate this argument. You're comparing cognition to reproduction. You're giving examples of duplication but comparing them to unique artifacts. It's not the same and you must understand this.
On top of this is the gross ethical concerns where all AI companies were allowed to steal all of this information/wisdom/art to train these models. At least with the power looms, a few skilled craftsmen were hired to help design them. This is blatant intellectual theft we're choosing to ignore because we're addicted to cheap products.
We choose temporary sensory pleasure over a healthy society.
I don't think the closest T-Shirt comparison is quite exactly either of those. E.g. you can get any image you like printed on a T-Shirt without hiring someone might be closer, though still not really the same. Thinking about that a bit more, everything seems like unique cognition until the technology to treat it as simple duplication exists. "Duplicate this project with this layout and style" has crossed that line now, just as taking a photo with a filter did long ago which itself long ago replaced a painter and so on. They technologies often produce something of lower quality than what the people they replaced would, few would say a photo filter is as good as a custom painting, but they tend to do the job well enough it replaces most of those kinds of workers.
I think the concern should be more on "you can't just extrapolate the same process forever and expect the same kind of outcome". Just because it was good in small amounts before doesn't necessarily mean it's inherently good to have an infinite capability to do that.
Regardless of all of that, I think it's important if we do have the AIs we're making we treat it as a common good in terms of wealth (regardless if people were paid) rather than a way to funnel a small amount of money to the companies who initially drive it as we are now. I.e. don't even mind if we "steal" (copy) everyone's available outputs as long as we all decide it should really benefit everyone rather than a handful.
> He said it's sad, and no matter how you want to spin it, it still is incredibly sad and depressing that we're hailing the age of "semantic photocopiers" as some sort of intellectual achievement.
We've been talking about being sad about a lot of different things in this thread.
But here's another thing to be sad about (somewhat related to what I was originally sad about): building "semantic photocopier" is probably quite an intellectual achievement, but so was the atom bomb. intellectual achievements, even impressive ones, can't be assumed to be good.
To hail a "semantic photocopier" requires a pretty myopic and parochial view. Looked at holistically, this stuff amounts to a project to rub out the human and replace it with cheap, inhuman slop. One aspect if that is when I see a picture like this, it's ruined because when I look at it I'm unsure if I'm communing with a human or getting fooled by a simulacrum. Another aspect is how the slop undermines the viability of creative and intellectually engaging professions.
We really ought to look at the kind of world we want to live in (which would probably have maximum satisfying creative work, and a minimum of drudgery), preserve the good parts of what we have, and focus on intellectual achievements to subtract the bad parts. But instead we're just mindlessly building whatever we can manage, damn the consequences or end result (which seems to be: more drudgery, less creative expression).
I want mine in watercolor. I’ve been wanting to take a watercolor class, so I figure this can be my mini-thesis project. I’ll still pass it to an AI to refine and animate.
Sad, but at the same time it's worth asking how many people would actually pay an artist to do this, versus those who would never do this unless it was for AI.
Agree, and same here. But upon some reflection I've concluded that I think I and you are the problem. We're essentially assisting AI in taking something away from us by immediately associating a style with AI.
The guy should cherish the style he likes and be proud – just like I refuse to give up my en and em dashes.
Fwiw, I'm currently having an house built, and I was asked to choose the colour of the garage door.
I tried giving AI all the plans, descriptions, 2D artist renders, and pictures of the house under construction so that it could make a render with a bunch of different options to help me choose.
The textures were surprisingly good, but the overall plan was a total disaster, even if I tried to guide it. It's two semi detached houses, but there was just one really long house. There was either one too many floor or one missing. The windows were totally random. And the orientation/positioning of the buildings on the plot had nothing to do with the plan. It even hallucinated stairs to reach the property when the whole thing is flat.
All of this to say I don't think AI is there yet to build this kind of illustrations based on just a floor plan and some pictures.
That would probably be better suited with a vision model of some sort or some other model specifically trained on the task at hand. It's not a problem with AI in general but rather using a general AI for a special use case.
You ought to give it a blender install and ask it to model the house for you & show you color options for the garage door, I bet it would be much closer. I had Astra w/ Codex model rigged miter saws, table saws and router tables for my carpentry simulator [0], it did a surprisingly decent job.
It's definitely a better approach. I was just checking what the models were capable of with a simple one off approach. And honestly, it mostly did what I required, which was the texture work. The rest didn't really matter in my case.
I'm sure with a bit more efforts/tokens, you can do something amazing, with a virtual visit and so on.
Gotta admit, pretty cool! I'm just like.. My interest in smart home stuff is near zero; am I alone?
I think it's just an extension of my not caring to "tinker" with my professional workspace. I have all stock OS backgrounds, always preferred a GUI editor/IDE. Barely any customizations that aren't tied to significant velocity improvements.. IDK. I'm not a "space" tinkerer; physical or digital.
I just like being able to control my lights and thermostat from my phone, I’ve never bothered with higher level automation outside of having a lamp turn on at dusk automatically lol
I have a time of use electric plan and adjust load (grow lights, dehumidifier, AC, battery backups) at certain times to save money. Smart home can be pure tinkering but it can control some pretty practical things.
You do you though, mechanical outlet timers work perfectly well.
I have kids, and I have to leave the house quickly, so being able to set the house for "away" mode or make sure you shut off all the lights when you are all in bed is really nice.
Adding off-peak controls for electrical devices, water heaters, thermostats, EV charging, etc just makes sense so you don't have to think too hard about it.
I also really enjoy setting lighting for different moods - specifically when watching movies, and that's easy to do from the couch :)
I basically don’t care and until recently I had no reason to. My house has no AC and I heat with wood.
But now I need a way to monitor the status of a system I’m using to prevent a water line from freezing, and suddenly all this home automation stuff seems very useful.
I also have limited interest, but alas, I have home assistant and I'm not sure I could fulfill my needs without it.
1 - My child is a runner. Zigbee door position sensors talk to home assistant and play sounds when a door is unlocked or opened.
2 - I'm super private. Giving people a key to my house is not something I'm comfortable with. I use Unifi Access + ESPHome + Zwave deadbolt to get the level of control I'm comfortable with. The doorbell also plays a sound in the house (in addition to the app on my phone)
3 - I have an under counter water cooler that runs way too hot. Zigbee temp sensor + Zigbee powerbar + simple exhaust fan keeps it cool.
4 - I dislike having 15 tv remotes and I have a more traditional AV setup (not sonos). I use events on the TV to control the audio (power on, volume, infer input selection)
AI has made all this a lot easier. I don't reall put a ton of effort in anymore, whereas before I'd have to fumble with configs.
Hah second the AI part. Started a home assistant setup years ago to setup water leak sensors with alerts to my phone, but didn't feel comfortable enough having home assistant accessible from the web, so besides writing brittle yaml files now also had to deal with open VPN etc etc and after major update my config files were incompatible and with kids just didn't get back to it.
However this week spent an hour with claude to migrate to tailscale, latest home assistant update, sensor setup and filtering for my rtl433 antenna, connected hue lights while we're at it, migrated to latest pihole version and put everything in git in one nice compose file with a step by step instruction on how everything was setup and which account sign ups were needed. I still checked every step and manually approved stuff, but the remote session feature made this easy to finish.
> I use Unifi Access + ESPHome + Zwave deadbolt to get the level of control I'm comfortable with.
That makes me uncomfortable, electronics controlling physical access sounds like a nightmare to me. I'm glad we have the options though, so different people can customize to the level that makes them feel safe.
What would really make you feel uncomfortable is there is near 100% probability that there are electronics controlling your physical access to buildings you use on a daily basis. Maybe not your home, but other acess.
And the most important thing(s) that any experienced HA automation geek knows is:
1) Always have a manual way to control a device if HA goes down.
I just feel incredibly uncomfortable with something as complex as modern software controlling something as basic as physical security. Now modern locks aren't exactly create, but adding on a layer of software just provides you with an additional layer that can fail.
For a home, I'd argue that you should fail open, for safety. So now I'm sleeping or out of the country and the software fails and the doors unlock.
Then there's the quality of the software. Home Assistance is probably fairly good, but the software in the devices probably isn't. All the demonstrations I've seen of home automation, regardless of supposed quality and interoperability has been, in my mind, yanking as all hell. If an IoT sensor reports nothing, is that because there's nothing to report, or has it lost connection, or is the configuration broken? I'd be constantly nervous about the later two. A practical example is tire pressure sensors, I despise them. They are more often than not misreporting, so you learn to ignore them.
And as you say: "any experienced HA automation geek knows", yeah, but I don't want to be a home automation geek, I just want the stuff to work. I don't want to be bothered by setup and configuration, I don't want to think about it, ever.
> Always have a manual way to control a device if HA goes down.
It is for this reason that my lights (Hue) and switches (Inovelli) are bound in ZigBee groups. HA adds some niceties, but the lights still work if it’s down. Accent lights I have bound to the Hue Bridge, because their app adds features HA doesn’t have, and they aren’t required for emergency egress.
The only gap I have, I think, is that because I’ve set all bulbs to respond to power restoration after a cut in last known state, if everything somehow lost its Zigbee binding, putting switches into dumb mode locally still wouldn’t help. However, since that’s an improbable scenario, I accept the remote possibility in exchange for not being awakened from lights coming on in the middle of the night. Plus I keep flashlights / headlamps throughout the house.
re #1 - I know families in similar positions, and a secondary lock placed high up on the door has been a lifesaver (literally), and gives much-needed peace of mind.
I'm the complete opposite of what you said in the second paragraph, I am very comfortable and appreciative of my tailored dotfiles, but I will say I don't stare at my light switch and think about tinkering with it usually. I do have one or two stupidly placed switches, like when the door to one room is closed I can't turn one of the ceiling fan lights on/off without opening the door (because the switch is outside the room), and also I have a bunch of switched outlets (awful to hit by accident if a computer is plugged into one) from boomers with a lamp fetish living here previously, but otherwise I'm happy with my light switches. Also I sit in the dark most of the time so even if I could turn every light in an area on/off at once, I'm not sure I'd do it often.
> My interest in smart home stuff is near zero; am I alone?
No, we were renovating the electrics and my boss got all excited and asked about all sorts of automation. We did none of it. The closet thing I have to smart home things is PoE in more network drops than I really need.
Frankly I don't want to deal with it, I have zero desire to debug my home. It even pisses me off that my car charger needs to be online, because I don't understand why it needs networking. If it's to prevent people from misusing it, just add the option of a padlock.
I have zero trust in IoT devices, both from a security perspective and from a functionality point of view.
It was fun to play with for a while. But after some integration broke for the third time I didn't bother to set things up again, and I really don't miss it that much.
I was not interested about smart home. At all. Then my father asked me to help him install an automatic switch for external lights in the yard. I never did that, I used a Shelly relay and Home Assistant, it was simpler (yes, sounds strange, but it was simpler). Then I found that I can automate some stuff in the house that was manual until then, so I did some.
4 years later I have more than 50 relays and sensors in 2 houses, controlling water pumps, irrigation, AC, thermostats, monitoring energy consumption and allowing energy optimizations and saving manual work. I have a smart water valve on my desk waiting to be installed and I need to add a few more thermostats (energy is so expensive these days), otherwise I am almost done.
I have nothing fancy like light themes and moods. I have solid solar-AC-heating optimizations with the temperature sensors and window open sensors; I have a decent irrigation setup and also external lights control. That's it. I am not a fan of tinkering with microcontrollers, I did not even modded or used custom firmware to any of it, I am just pragmatic in using the tech where I need and for what brings me benefits.
I think the audience for home automation is U-shaped. There are folks who are neck deep in it and care deeply, and others who couldn't care less, but not many in the middle. The middle 90% is decently served by Apple, Google and Amazon, too, so there's no need for any real passion or exploration by most anyway.
We're building a custom home and it's only been through this process that I've taken an interest in smart home / home automation. Why? Mostly because
1) building speakers into your walls & ceilings means you need to control their input somehow, and your gear is going to be in a wiring closet
2) if you don't want Alexas or Nest devices sitting around on horizontal surfaces, your voice control options are limited
3) you don't want your security camera footage stored in the cloud
4) you do want a single pane of glass for visibility into things like your HVAC, your irrigation system, your security system, your garage door status, and your AV gear.
Home Assistant is tremendously helpful as a FOSS alternative to Control4, Crestron, Savant, Lutron, and Josh.ai, but you can't entirely escape the "big tech" part of the ecosystem because each of the primary systems (say, your AV streaming or your power shades) will still require use of the OEM's product. Many of which have integrations with or APIs exposed to Home Assistant, but not all.
There are some quality of life improvements available through automation, but most of it is incremental and only enthusiasts would really care.
> So I thought: why not build a dashboard so cool that everyone would want to use it?
I think my grandma would think such a thing impossible.... Especially when it's 2am and you just need to get to the right app to turn the bathroom lights on...
beautiful illustration - serious question why do ppl still use dashboards? I've been using HA for almost 10 years now and the best part is that everything is automated and I don't have to touch anything.
In general using a large setback will save more energy by not constantly cooling the home than it costs to return the house to a desired setpoint later.
This is a common advice I'm not sure I've ever seen properly explained and proven. It's mostly "we asked the experts!" kind of 3min tidbits with a guy in a corporate polo shirt talking assertively to the camera.
Most people actually measuring it come up with more savings when cutting the AC during the longer spans when nobody's there (usually during the hottest part of the day).
This would only be true for very short time periods (minutes) that cause the compressor to power up more times than it would have otherwise as compressor startup incurs transiently high power consumption, but a thermostat will usually do this many times a day, so the period of time that this has any practical implication is very short (<1 hour). At longer time periods, it is always more energy efficient to disable cooling and restart it only when temperature control is again desired.
Even among people who service AC units I've encountered the misconception that it is more energy efficient to keep a space at the desired temperature than it is to let it heat up and then cool it back down again, but thermodynamics say it is not so. It could conceivably be more cost efficient if your energy costs are variable.
I remember the same misconception existing about fluorescent lamps in the past. "They consume more when turning on, so the gains vs incandescent are negated". Which was verifiably false to me at the time by checking the label on a lamp.
Air temperature is only one contributing factor for comfort in a home. Surface temperatures take longer to achieve and the larger the gradient the longer it can take.
There are multiple factors to take into account, including the highest temperature the home gets to, that is dependant on location and weather patterns.
If you keep the temperature in the home to 28-30 degrees then later cooling to lower temperatures before bed time you can reduce costs and energy expenditures than if you let it rise to 34 degrees or more and only lowering it when you get home. Or, what you save in costs you expend in otherwise like poor sleep.
Lowering it advance, on a schedule makes more sense if you want surface temperatures to also be reduced.
The blanket disabling of air conditioning when not home is not the only solution, and often is sub-optimal.
Pretty sure that is completely false, though it is popular to say. It 100% fails even a basic sniff test. The rate of energy transfer from outside to inside will be greatest when the inside is coolest, so the air conditioner will have to work harder overall to keep it that way.
It can be true, but usually not for AC units, as these are quite effective at only heating/cooling the ambient air and not the walls, ceiling etc.
Regular heating can waste a lot of energy to heat the entire structure of the house, especially if you have thick walls, so it does more make sense to try and keep the temperature stable.
The bigger HVAC win, in my opinion, if you have time-of-use billing, is to crank the temp when power is cheap (usually overnight), and, assuming your house is relatively well insulated, letting it ride through the rest of the day. This does lead to some funny things like the fact that I use heavier blankets in the summer than in the winter because I chill the house so far overnight, but for anyone whose climate is relatively moderate and with a relatively modern house, this can save quite a bit on power costs.
Admittedly, this will probably not work that well in a Minnesota winter, or an Arizona summer, but I think a surprising number of people could benefit.
I think it's worth asking, especially with agentic AI tasks that can cost upwards of $20, whether a human could've done a better job on at least one aspect of your project. Fiverr exists.
IDK how much the illustrator was paid (they're Jakarta based and seem like a freelancer), but I think the result was worth it.
The arts have always survived because of patrons, regardless of how favourable or unfavourable the economy was for them. I'm not saying this is the only solution to protecting it from AI, but it's one worth thinking about if you have the resources to do so.
That is lovely. Certainly puts my barebones 2D SVG floor plan to shame. (On the other hand, there's 20+ sensors on there which might be tricky to fit onto something like that image.)
I can't help but notice that the displayed CO2 levels in your example images are quite high. If they're screenshots of the system in action I would change the air in your home more often. If they're just made up numbers, then never mind :)
It's a wide spectrum. It absolutely can mean "if Samsung has a bad day my house becomes uninhabitable."
But speaking as somebody also using Home Assistant, I'd bet the author has everything set up without internet access, fully locally controllable, and entirely operable the old fashioned way if Home Assistant goes down.
That's about perfect for my office, keeps my hands limber and avoids carpal tunnel, but not warm enough to sweat. I do like it more like 19C at night for sleeping, though.
I like how this combines two kinds of creativity. One is the pure artistic aesthetic creativity of the illustrator, and the other is the engineering creativity of even thinking to hire an illustrator and building the software for it.
People think of creativity in one shape, but it comes in all shapes and sizes.
The AI-slop writing while talking about hiring a human to do the illustration...
"A flickering fireplace, a softly lit garden, and a smart home the whole family finally enjoys."
"not only in my house, but in a hand-drawn picture of it"
Really reaching witch hunt level "AI" spotting. Saw an ad on instagram the other day where all the comments were "we don't want your AI slop" until the actor in the ad showed up and proved that he was a real human who shot a real video.
>> Try starting the sprinkler with the button on the unit itself, then running out of the garden before you get soaked.
Go to the store. Buy a valve and some pipe. Mount the valve near the door like any other lazy person. Or, if the outside is too scary, buy an electric valve and wire it to an indoor light switch, maybe even a smart switch so you need not rise from the couch.
I've commissioned art before via the Krita web forum's job section. Lots of bright young artists, many of them students, that do great work for affordable prices to build their portfolios.
You don't really need to be in the 1% to be a patron if you need some art.
I'm not sure what that is supposed to convey. This person was able to scrape together a couple bucks to pay someone to draw a picture of their house in a specific way. You've never paid for any art or entertainment? Netflix?
*$900, and no, the average person can’t afford $900 for bespoke art (or more likely, has no interest in paying $900 for one). Heck, I don’t see how you can look at the explosion of AI art online and not come to the conclusion the average person would much rather have the same for near free than paying hundreds of dollars. To me, that’s a pretty obvious revealed preference.
The overwhelming majority of people of any income level will never pay to commission any kind of art or even be familiar with the possibility to consider as a practical option.
It's just not a typical thing to do pre or post AI. "Revealed preference" doesn't seem like a useful lens without some basic degree of awareness that a choice even exists.
And on the flip side, I have a friend who commissioned some kind of avatar for social media from an artist friend of friend for the first time in their life. They loved the result a lot and since then have commissioned a bunch of little random things once they realized it was a thing anyone could easily do with a few bucks. For stuff like their WoW character/guild, D&D, social media, etc.
They make less than six figures, each request was <$100 IIRC. Budgeted basically like any other gaming/hobby purchase for personal entertainment.
But i mean, i dont think he thought of this as an "art-object" more like a ... house-control-kiosk-whatever. And people spend tons of money on those things. Eink-screens for thousands. Control4 for whatever amount. Homey-hub for..300usd? Hue-lights for hundreds etc.
So, sure 900usd or a part of that system is a lot of money and I wouldn't do it. But I've spent 900usd on techthings with a very diminishing amount of return :)
I never said it was lack of budget, it's lack of interest. I'd honestly rather have a $30 print of a Picasso, a patent, a blueprint, or a movie poster than pay someone for custom art, nor do I think one needs to buy or commission something for it to have a personal connection (for example the ratty BTTF poster I had in college, or the Chicago Lyric Opera advertisement I saved out of a nursing home dumpster).
I've commissioned tons of art, and one of the great joys is the surprise at how they interpret your idea, and how it's always different than what you had in your mind. That joy is unique to commissioned art, I think.
Pro tip you don’t even need an illustrator. While I do encourage hiring humans to do this work you can also use PolyCam with a phone that supports lidar, export it to Claude with Blender and voila.
This is typically the kind of hacking project that is preposterously non-economically-viable for 99% of homelabbers, let alone regular folks. Now, for the first time ever, AI is making it actually palatable for everyone who wants it done to have it done... and certainly much cheaper too at an equivalent quality level.
For (1) the vast majority of people don't want to spend what's probably on the roundabout of a thousand dollars for their homelab's homepage. You can easily see this in almost every homelabbing discussion on reddit or even here, where everyone focuses on cheap/hackish solutions to avoid overspending... except in hardware of course =) It's kind of become a meme at this point.
For (2), certainly, especially so considering that the author gave the illlustrator blueprints, photos, instructions, etc. for the commissioned illustrator to convert into dynamically-compatible svgs.
There's a lot of top-down implementing there with very low bottom-up artistry. This doesn't at all take from the final quality -which is pretty good- but it makes it easily replicable using current-gen AI.
Commissioning art like this is at most a few hundred dollars. As was posted elsewhere, there are lots of young artists that need to pad out their portfolios, as well as artists in places where the cost of living is significantly cheaper. This isn't a multi-thousand-dollar thing that only the 1% could afford.
It's _better_ that the illustrations were made by an illustrator, but it still would have been cool and front-page-worthy if that hadn't been the case.
Not really. AI slop is an instant turn-off for me. It's like hearing someone talk about a dream they had; it is egregiously boring for everyone except the person relating it.
Some quick googling suggests that in the US, only 60% of people have iphones, and only 50% of iphones in use are pro or max. I would guess that HN skews towards more recent and higher end phones, but also skews towards android. Maybe saying most people don't have lidar is inaccurate, but I think it's still accurate to say the majority of people don't have lidar.
I used AI to get it draw it. What worked is giving floor plan I already had and putting some labels on it - brown dining table here, fridge here, etc.
What didn't work was giving it bunch of photos and asking it to just figure it out. For things it did figure out - it was nice, but it made way too make mistakes.
The entire point of it was to put light switches and heaters on the map so I don't have to fiddle with a number of toggles in a table/list.
p.s. I wish I had huge eink display for home assistant dashboards. Some concepts I made and seen online are glorious.
This is the illustrator: https://linktr.ee/owenyeconiel
It’s cool they hired an actual person but if I was so happy with their work to write-up a whole article about it, I’d be crediting them in a much more prominent way!
They credited the artist at the bottom of the article. Not sure if it was there before you commented or not
Wayback Machine confirms it was there a few days ago.
People are just never happy and always going to have something to say. Even if you don't go the AI route and hire an illustrator, even if you credit work, someone has to be critical.
The obvious place to do it is in the writing. I had been expecting that during the entire paragraph, being art and liking what I was seeing I just wanted to see the artist portfolio.
Scroll down two inches.
What would you want?
In this day and era such a thing will feel like yet another ad. We react on a very defensive way against ads.
The way it is actually done, it generates genuine interest, and you have the link to the artist at exactly the point where your interest is at the highest.
I think part of the problem is that he included the illustrator so prominently in the title. He could have just said:
I used an illustration to…
As the spouse of an illustrator, I applaud the author for hiring an actual artist. We've seethed at the way artists work was scraped online for the training of these models, recreating their very style for the creation of these infringement engines.
The finished product benefited greatly. There's nothing quite like working with a genuine expert who knows the rules and tradeoffs on things like competition, color, etc. Author did good on providing the illustrator with a detailed brief. When the illustrator takes those guidelines and applies their style and training to the product, magic happens.
I'm a 3d modeler and a software engineer. AI didn't steal my work. Anybody could have looked at my repositories and art and learned from it. Just because something came along that is a really good learner doesn't make that suddenly be theft.
The case is slightly stronger if you want to talk about training that happened on our private code repositories, but publicly visible art and repositories are fair game for learners to learn.
Most artists I know never had a problem with other artists learning from them. But they do have a problem with corporations training on their data without any sort of consent and then selling a product that takes away their jobs, they never agreed to it and they never would have shared it to them.
Then they presumably are fine with open weights models made by hobbyists by scraping publicly available art on boorus and such, yes? because they are neither made by corporations nor are they selling any product.
(to be clear, I don't think this is an accurate description of how artists feel about e.g. publicly released diffusion models. The core is not that it's made by "corporations" or the for-profit aspect or even the copyright infringement part, it's simply the understandable and not-at-all-unique-to-artists resentment at having your job or passion automated or displaced)
True. they were fine providing advances to their field for others to learn and build from that shared legacy. They didn't agree for their work to be the base of auto-slopping at industrial scale. Quite easy to understand, except for AI-bros.
How is this morally different from e.g. weavers and loom operators being automated or ceramics or anything else? Or programmers?
We've been automating people "out of jobs" since at least the start of the industrial revolution. It's generally considered to be a goos thing.
We had a sort of weird special case for copyright preventing certain types of machine automation and I've heard plenty of complaints about how that worked, but it more or less did work for a while.
>We've been automating people "out of jobs" since at least the start of the industrial revolution. It's generally considered to be a goos thing.
By who? The corporate bag holders?
I feel like "don't agree" is a weird framing, because it assumes that all technological innovations must have the explicit and affirmative consent of all parties they could directly or indirectly affect. Not to mention that "artists" are not a monolith any more than "AI bros" are.
I could sort of get behind something like this in theory (weighing each new technology based on whether it furthers the values of the community, and rejecting it if it doesn't, kinda like what the Amish do), but I'm not sure it's being applied neutrally or consistently here, nor do I think it works in practice especially for something as decentralized as a diffusion model, which are surprisingly simple to build and train and can run on a home GPU.
I think the reason artists wouldn't like hobbyists doing it either is just that they've been soured on it by corporations. Artists didn't reject AI in its infancy, or things like deepdream. But AI as it stands very clearly is exploitative, and what's more, it's flooded spaces formerly for human art with low quality generic garbage. Having those experiences bleeds over into your feelings about things that don't match the exact definition of the reason you're upset to begin with.
>Then they presumably are fine with open weights models made by hobbyists by scraping publicly available art on boorus and such, yes? because they are neither made by corporations nor are they selling any product.
Why would anyone be fine with people explicitly not actually making art.
Scrapping and generating isn't art and isn't comparable to people tracing/copying/learning from existing art.
Multi-billion-dollar copy-machines are the same as a person just dropping by and having a look? If you can't tell the difference, what are you? Be you human or robot?
Robot we be...just two robots out roboting it up.
> AI didn't steal my work
...yet. AI has evolved greatly from basic image generation. It's already pretty capable at using tools to great SVGs and 3D scenes that are more reliable than plain images. Sure, experts won't lose their jobs, but how will juniors become new experts?
It's generous of you to license your work for anyone to use. Not everyone did. AI, without a doubt, stole from people's work.
To be fair, if an illustrator uses these models to write code, it is exactly the same thing. We built everything for decades, got scraped, and now we don't write code ourselves. Everyone uses it to write code, and nobody gets upset. If people use it for art, they're angry. We've both been used.
> and nobody gets upset
[citation needed]
I think there's something more interesting going on. We (coders) use LLMs to write code and still get paid, because what we bring to the table is context across a multitude of coding tasks.
There are art projects like this, but many traditional art projects are quick one-shot deals. Make a logo, design a business card, make a tshirt design. Bite sized, do the work, get paid, move on.
LLMs are great at short context tasks. A lot of art is short context. Building the software that runs a bank is not.
If you want to work as a professional artist in the age of LLMs, move into long context work. If it isn't already, it will be impossible to make a living off of fiverr.
>Everyone uses it to write code, and nobody gets upset.
A lot of us are very upset outside of the techbro sloperator pool of Ycombinator types.
1. Hiring a human illustrator is the right thing to do, and I applaud this guy.
2. AI kinda ruined this style for me. The first thing I did when I saw it was start searching for AI incongruities which is sad.
The really sad thing is that the natural instinct for many will be to screenshot the page and upload it with "please create me a home assistant dashboard in this style".
I spent last 2 months in construction stores. A good chunk of people drop a size in claude/chatgpt, and put tiles in their existing picture etc.
You can feel differently but it saves you so much money, time, and bring you so much pleasure. Not the other way around.
just claude/chatgpt, no specialized app?
yes, it's amazingly/depressingly good at things that would take a $1000+ software, and years of education just a year ago. Now I'm doing this with leftover claude/codex weekly quotas( i do have both due to some testings at work).
I designed my basement renovation using a combination of image modelsand a 3D scanned Houzz Pro FloorPlan, and I gotta say it REALLY helped the contractor deliver on the vision that I had.
Decorating when you can easily swap out pieces and parts in images with realistic dimensions makes rapid iteration on physical things much faster and it's easy to figure out what "speaks" to you. Great use of these tools.
I’m genuinely curious how you use image models for this type of design. I need a way for my wife and I to visualize a remodel project and don’t know where to start or which tools are best for the job.
dump the plan, make sure it has height etc, after dump pictures if needed,
refs of colors. atropic has some nice recipes somewhere online, entire prompts you can feed into cc. i used their workflows feature, otherwise it's painfully slow.
do it in two steps, first generate just the named 3d model, walls, stairs etc inspect for correctness, after proceed to the fun part, materials, colors, etc.
I did it in extra one step, not sure if that was needed, i build a local harness first for extra small room 6 ft by 6ft, and after that expanded into real size. It installed lots of tools.
Just upload the pictures to a chatGPT chat and tell it what you want, My wife uses it for this all the time.
The difference in the replies to this comment are striking.
One is a relatively complicated workflow producing 3D assets, the other is "just give pictures to chatgpt and tell it what you want".
And neither shows the outcome as to illustrate/support their point, which is phenomenally ironic considering the context..
It’s not different . One case you chose tiles , you take a picture of your bathroom , and give it a rough sized of tile + room and it will Give you visual check . Second option if you want it also to calculate materials . This is a 3d path . First one can be done by just chat , another one is coding agent .
I tried a couple of the web-based design-a-floorplan sites and all of them were kinda like using crayons to draw a circuit diagram. Ended up doing a pile of sketches on paper, then when I had what I liked finalised it in Visio and sent it through to an architect who turned it into proper plans which, since they were architect-drawn plans I had to check for f_ckery and then the builder checked it for stuff I'd missed but ended up with a rebuild that was exactly what we wanted.
Well with one bit of f_ckery that both of us missed until after the concrete had been poured because it was too small to see easily on the plan, and a second also too-small bit that I caught once they'd put the framing up and asked them to take it out again.
My wife designed our pool and our patio this exact way. She spent hours on it and it worked perfectly.
I can look out my window and see the exact scene she made in chatGPT.
Really impressive how useful it has become on some very general things
That's a great idea. I may just try it.
It's not sad, it's literally what those tools are for.
Yes, this is the "you wouldn't download a car" moment. Yes, I would.
(And no, I'm not doing it, because I have my own preferences, which involves hitting a very small target in design space, that is applying "dark cockpit" principles and a mix of industrial HIGs and nuclear reactor control room interface standards to LCARS.)
> It's not sad, it's literally what those tools are for.
If someone spreads poison on your street to kill the strays and mutts (yes, that’s a thing that happens) and it kills your dog, would you like someone to tell you “that’s not sad, that’s literally what poison is for”?
Using a tool for its intended use has zero relation to it being sad.
What analogy are you trying to draw here? Creating a copy/derivative work of something notably doesn't destroy the original.
If it destroys the market for illustrators then it effectively kills any future innovation in the field.
Yes, I do see that problem. I just wish we'd call it what it is instead of drawing imprecise and confusing analogies that are often thinly veiled appeals to emotion.
That would be the first step to addressing the actual issues: Who will advance the field, assuming that generative AI will always be limited to copying instead of innovating? What will be the place in society for the people that lose their job due to these changes? etc.
> I just wish we'd call it what it is instead of drawing imprecise and confusing analogies that are often thinly veiled appeals to emotion.
And I wish people on HN would steel man arguments and not assume intentions in others, but here we are. There was no “appeal to emotion”, veiled or otherwise. The example used harming dogs because hopefully most¹ of us can agree that’s repulsive and bad.
Everything else you’re saying has nothing to do with the point. AI isn’t the point (hint: that’s why I didn’t mention it), the point is spelled out in plain English at the end of the comment:
> Using a tool for its intended use has zero relation to it being sad.
¹ I promise you that qualifier is necessary. I’ve seen some vile opinions on HN.
My comment was specifically regarding your example of harm against animals, yes.
Why do you need to invoke an image of something widely considered repulsive to strengthen your point against something much more nuanced/controversial?
It was not to “strengthen the point”, it was to have a clear example where most of us can be on the same page of what is wrong. This is HN, anything other than an obvious example attracts an infinite swarm of nitpickers arguing that that whatever bad thing “is very good, actually”.
> are often thinly veiled appeals to emotion.
What's your problem with emotion?
Nothing, emotions are just what they are. Rousing them in other people to sway them in your favor is pretty manipulative though, I'd say.
> Rousing them in other people to sway them in your favor is pretty manipulative though, I'd say.
Yes, I agree. That was not what I was doing.
He didn't say it's illegal (even if it's debatable and kinda loophole-ish, but lets pretend for now that society has fully embraced a world-encompassing copyright exception if you're named Altman or Amodei).
He said it's sad, and no matter how you want to spin it, it still is incredibly sad and depressing that we're hailing the age of "semantic photocopiers" as some sort of intellectual achievement.
If you consider it incredibly sad and depressing for the majority of people to have cheap access to things that would have cost a small fortune due to being inevitably hand-crafted only a few years ago, I hope you're renouncing clothing manufactured using automated looms, printed books etc. as well?
Absolutely, I think fast fashion brands like Shein and TEMU-style shops are incredibly sad, toxic for the economy and the environment, and I've never bought anything from them.
You conveniently didn't address the tougher problem GP mentioned: books.
Are you arguing that the only type of clothing an automated loom can help produce is fast fashion? Is every piece of clothing you own hand-woven, crocheted etc.?
GP: "Are you avoiding clothing on automated looms, like every piece in every major store you have visited in your lifetime?"
easyThrowaway: "I'll pretend your question is about fast fashion and TEMU!"
Man I hate this argument. You're comparing cognition to reproduction. You're giving examples of duplication but comparing them to unique artifacts. It's not the same and you must understand this.
On top of this is the gross ethical concerns where all AI companies were allowed to steal all of this information/wisdom/art to train these models. At least with the power looms, a few skilled craftsmen were hired to help design them. This is blatant intellectual theft we're choosing to ignore because we're addicted to cheap products.
We choose temporary sensory pleasure over a healthy society.
I don't think the closest T-Shirt comparison is quite exactly either of those. E.g. you can get any image you like printed on a T-Shirt without hiring someone might be closer, though still not really the same. Thinking about that a bit more, everything seems like unique cognition until the technology to treat it as simple duplication exists. "Duplicate this project with this layout and style" has crossed that line now, just as taking a photo with a filter did long ago which itself long ago replaced a painter and so on. They technologies often produce something of lower quality than what the people they replaced would, few would say a photo filter is as good as a custom painting, but they tend to do the job well enough it replaces most of those kinds of workers.
I think the concern should be more on "you can't just extrapolate the same process forever and expect the same kind of outcome". Just because it was good in small amounts before doesn't necessarily mean it's inherently good to have an infinite capability to do that.
Regardless of all of that, I think it's important if we do have the AIs we're making we treat it as a common good in terms of wealth (regardless if people were paid) rather than a way to funnel a small amount of money to the companies who initially drive it as we are now. I.e. don't even mind if we "steal" (copy) everyone's available outputs as long as we all decide it should really benefit everyone rather than a handful.
Literal photocopiers are a pretty major intellectual achievement.
> He said it's sad, and no matter how you want to spin it, it still is incredibly sad and depressing that we're hailing the age of "semantic photocopiers" as some sort of intellectual achievement.
We've been talking about being sad about a lot of different things in this thread.
But here's another thing to be sad about (somewhat related to what I was originally sad about): building "semantic photocopier" is probably quite an intellectual achievement, but so was the atom bomb. intellectual achievements, even impressive ones, can't be assumed to be good.
To hail a "semantic photocopier" requires a pretty myopic and parochial view. Looked at holistically, this stuff amounts to a project to rub out the human and replace it with cheap, inhuman slop. One aspect if that is when I see a picture like this, it's ruined because when I look at it I'm unsure if I'm communing with a human or getting fooled by a simulacrum. Another aspect is how the slop undermines the viability of creative and intellectually engaging professions.
We really ought to look at the kind of world we want to live in (which would probably have maximum satisfying creative work, and a minimum of drudgery), preserve the good parts of what we have, and focus on intellectual achievements to subtract the bad parts. But instead we're just mindlessly building whatever we can manage, damn the consequences or end result (which seems to be: more drudgery, less creative expression).
I want mine in watercolor. I’ve been wanting to take a watercolor class, so I figure this can be my mini-thesis project. I’ll still pass it to an AI to refine and animate.
Sad, but at the same time it's worth asking how many people would actually pay an artist to do this, versus those who would never do this unless it was for AI.
Raises hand
This is cool and I'd like to play around with it, but I'm not going to hire an illustrator for it, at least not to start with.
Isn't this how we invented fire and started civilization?
"One dude did a thing and the rest copied it coz it seemed cool. Everybody wants to be cool..."
If "everyone wants to be cool", copying fire seems like a mistake...
Agree, and same here. But upon some reflection I've concluded that I think I and you are the problem. We're essentially assisting AI in taking something away from us by immediately associating a style with AI.
The guy should cherish the style he likes and be proud – just like I refuse to give up my en and em dashes.
> I've concluded that I think I and you are the problem
Oh please. If someone shits in your sandwich, you're not the problem for disliking the flavor of salami for the next 2 years.
Fwiw, I'm currently having an house built, and I was asked to choose the colour of the garage door.
I tried giving AI all the plans, descriptions, 2D artist renders, and pictures of the house under construction so that it could make a render with a bunch of different options to help me choose.
The textures were surprisingly good, but the overall plan was a total disaster, even if I tried to guide it. It's two semi detached houses, but there was just one really long house. There was either one too many floor or one missing. The windows were totally random. And the orientation/positioning of the buildings on the plot had nothing to do with the plan. It even hallucinated stairs to reach the property when the whole thing is flat.
All of this to say I don't think AI is there yet to build this kind of illustrations based on just a floor plan and some pictures.
That would probably be better suited with a vision model of some sort or some other model specifically trained on the task at hand. It's not a problem with AI in general but rather using a general AI for a special use case.
You ought to give it a blender install and ask it to model the house for you & show you color options for the garage door, I bet it would be much closer. I had Astra w/ Codex model rigged miter saws, table saws and router tables for my carpentry simulator [0], it did a surprisingly decent job.
[0] https://sawdust.diy
It's definitely a better approach. I was just checking what the models were capable of with a simple one off approach. And honestly, it mostly did what I required, which was the texture work. The rest didn't really matter in my case.
I'm sure with a bit more efforts/tokens, you can do something amazing, with a virtual visit and so on.
Groverhaus for the modern age?
Gotta admit, pretty cool! I'm just like.. My interest in smart home stuff is near zero; am I alone?
I think it's just an extension of my not caring to "tinker" with my professional workspace. I have all stock OS backgrounds, always preferred a GUI editor/IDE. Barely any customizations that aren't tied to significant velocity improvements.. IDK. I'm not a "space" tinkerer; physical or digital.
I just like being able to control my lights and thermostat from my phone, I’ve never bothered with higher level automation outside of having a lamp turn on at dusk automatically lol
I have a time of use electric plan and adjust load (grow lights, dehumidifier, AC, battery backups) at certain times to save money. Smart home can be pure tinkering but it can control some pretty practical things.
You do you though, mechanical outlet timers work perfectly well.
I have kids, and I have to leave the house quickly, so being able to set the house for "away" mode or make sure you shut off all the lights when you are all in bed is really nice.
Adding off-peak controls for electrical devices, water heaters, thermostats, EV charging, etc just makes sense so you don't have to think too hard about it.
I also really enjoy setting lighting for different moods - specifically when watching movies, and that's easy to do from the couch :)
>Barely any customizations that aren't tied to significant velocity improvements
I never was interested in the tinkering side. The functional side has been very handy.
For example, smart thermostats reduced the heating bill 40% while leaving the place warmer where we actually are.
Smart dimmers on lamps let you dim them in the evening, or turn all the lights off with one push.
Your use cases may vary and the answer may be none do it but a good level of smart home things are quite practical.
I basically don’t care and until recently I had no reason to. My house has no AC and I heat with wood.
But now I need a way to monitor the status of a system I’m using to prevent a water line from freezing, and suddenly all this home automation stuff seems very useful.
I also have limited interest, but alas, I have home assistant and I'm not sure I could fulfill my needs without it.
1 - My child is a runner. Zigbee door position sensors talk to home assistant and play sounds when a door is unlocked or opened.
2 - I'm super private. Giving people a key to my house is not something I'm comfortable with. I use Unifi Access + ESPHome + Zwave deadbolt to get the level of control I'm comfortable with. The doorbell also plays a sound in the house (in addition to the app on my phone)
3 - I have an under counter water cooler that runs way too hot. Zigbee temp sensor + Zigbee powerbar + simple exhaust fan keeps it cool.
4 - I dislike having 15 tv remotes and I have a more traditional AV setup (not sonos). I use events on the TV to control the audio (power on, volume, infer input selection)
AI has made all this a lot easier. I don't reall put a ton of effort in anymore, whereas before I'd have to fumble with configs.
Hah second the AI part. Started a home assistant setup years ago to setup water leak sensors with alerts to my phone, but didn't feel comfortable enough having home assistant accessible from the web, so besides writing brittle yaml files now also had to deal with open VPN etc etc and after major update my config files were incompatible and with kids just didn't get back to it.
However this week spent an hour with claude to migrate to tailscale, latest home assistant update, sensor setup and filtering for my rtl433 antenna, connected hue lights while we're at it, migrated to latest pihole version and put everything in git in one nice compose file with a step by step instruction on how everything was setup and which account sign ups were needed. I still checked every step and manually approved stuff, but the remote session feature made this easy to finish.
> I use Unifi Access + ESPHome + Zwave deadbolt to get the level of control I'm comfortable with.
That makes me uncomfortable, electronics controlling physical access sounds like a nightmare to me. I'm glad we have the options though, so different people can customize to the level that makes them feel safe.
What would really make you feel uncomfortable is there is near 100% probability that there are electronics controlling your physical access to buildings you use on a daily basis. Maybe not your home, but other acess.
And the most important thing(s) that any experienced HA automation geek knows is:
1) Always have a manual way to control a device if HA goes down.
2) Local only, no cloud.
I just feel incredibly uncomfortable with something as complex as modern software controlling something as basic as physical security. Now modern locks aren't exactly create, but adding on a layer of software just provides you with an additional layer that can fail.
For a home, I'd argue that you should fail open, for safety. So now I'm sleeping or out of the country and the software fails and the doors unlock.
Then there's the quality of the software. Home Assistance is probably fairly good, but the software in the devices probably isn't. All the demonstrations I've seen of home automation, regardless of supposed quality and interoperability has been, in my mind, yanking as all hell. If an IoT sensor reports nothing, is that because there's nothing to report, or has it lost connection, or is the configuration broken? I'd be constantly nervous about the later two. A practical example is tire pressure sensors, I despise them. They are more often than not misreporting, so you learn to ignore them.
And as you say: "any experienced HA automation geek knows", yeah, but I don't want to be a home automation geek, I just want the stuff to work. I don't want to be bothered by setup and configuration, I don't want to think about it, ever.
> Always have a manual way to control a device if HA goes down.
It is for this reason that my lights (Hue) and switches (Inovelli) are bound in ZigBee groups. HA adds some niceties, but the lights still work if it’s down. Accent lights I have bound to the Hue Bridge, because their app adds features HA doesn’t have, and they aren’t required for emergency egress.
The only gap I have, I think, is that because I’ve set all bulbs to respond to power restoration after a cut in last known state, if everything somehow lost its Zigbee binding, putting switches into dumb mode locally still wouldn’t help. However, since that’s an improbable scenario, I accept the remote possibility in exchange for not being awakened from lights coming on in the middle of the night. Plus I keep flashlights / headlamps throughout the house.
re #1 - I know families in similar positions, and a secondary lock placed high up on the door has been a lifesaver (literally), and gives much-needed peace of mind.
Example: https://www.amazon.com/dp/B07DR9CMGZ
I'm the complete opposite of what you said in the second paragraph, I am very comfortable and appreciative of my tailored dotfiles, but I will say I don't stare at my light switch and think about tinkering with it usually. I do have one or two stupidly placed switches, like when the door to one room is closed I can't turn one of the ceiling fan lights on/off without opening the door (because the switch is outside the room), and also I have a bunch of switched outlets (awful to hit by accident if a computer is plugged into one) from boomers with a lamp fetish living here previously, but otherwise I'm happy with my light switches. Also I sit in the dark most of the time so even if I could turn every light in an area on/off at once, I'm not sure I'd do it often.
Smart home stuff I mostly agree. But there's still plenty of semi dumb smart home stuff you can do that's nice.
Like plugging in an old AC into a smart plug to manually turn it on remotely (make sure it had enough amps)
I will say though monitoring my air quality with a proper "smart" device (dedicated to monitoring) has been very eye opening
> My interest in smart home stuff is near zero; am I alone?
No, we were renovating the electrics and my boss got all excited and asked about all sorts of automation. We did none of it. The closet thing I have to smart home things is PoE in more network drops than I really need.
Frankly I don't want to deal with it, I have zero desire to debug my home. It even pisses me off that my car charger needs to be online, because I don't understand why it needs networking. If it's to prevent people from misusing it, just add the option of a padlock.
I have zero trust in IoT devices, both from a security perspective and from a functionality point of view.
> My interest in smart home stuff is near zero; am I alone?
You are not. I have no interest in having my appliances spy on me, or hit me with sudden unexpected costs or subscriptions, or being hacked.
https://reddit.com/r/theinternetofshit/
It was fun to play with for a while. But after some integration broke for the third time I didn't bother to set things up again, and I really don't miss it that much.
It's just a hobby like any other, some people will like it, some people won't care for it. There's nothing wrong with that.
I was not interested about smart home. At all. Then my father asked me to help him install an automatic switch for external lights in the yard. I never did that, I used a Shelly relay and Home Assistant, it was simpler (yes, sounds strange, but it was simpler). Then I found that I can automate some stuff in the house that was manual until then, so I did some.
4 years later I have more than 50 relays and sensors in 2 houses, controlling water pumps, irrigation, AC, thermostats, monitoring energy consumption and allowing energy optimizations and saving manual work. I have a smart water valve on my desk waiting to be installed and I need to add a few more thermostats (energy is so expensive these days), otherwise I am almost done.
I have nothing fancy like light themes and moods. I have solid solar-AC-heating optimizations with the temperature sensors and window open sensors; I have a decent irrigation setup and also external lights control. That's it. I am not a fan of tinkering with microcontrollers, I did not even modded or used custom firmware to any of it, I am just pragmatic in using the tech where I need and for what brings me benefits.
I think the audience for home automation is U-shaped. There are folks who are neck deep in it and care deeply, and others who couldn't care less, but not many in the middle. The middle 90% is decently served by Apple, Google and Amazon, too, so there's no need for any real passion or exploration by most anyway.
We're building a custom home and it's only been through this process that I've taken an interest in smart home / home automation. Why? Mostly because
1) building speakers into your walls & ceilings means you need to control their input somehow, and your gear is going to be in a wiring closet
2) if you don't want Alexas or Nest devices sitting around on horizontal surfaces, your voice control options are limited
3) you don't want your security camera footage stored in the cloud
4) you do want a single pane of glass for visibility into things like your HVAC, your irrigation system, your security system, your garage door status, and your AV gear.
Home Assistant is tremendously helpful as a FOSS alternative to Control4, Crestron, Savant, Lutron, and Josh.ai, but you can't entirely escape the "big tech" part of the ecosystem because each of the primary systems (say, your AV streaming or your power shades) will still require use of the OEM's product. Many of which have integrations with or APIs exposed to Home Assistant, but not all.
There are some quality of life improvements available through automation, but most of it is incremental and only enthusiasts would really care.
> So I thought: why not build a dashboard so cool that everyone would want to use it?
I think my grandma would think such a thing impossible.... Especially when it's 2am and you just need to get to the right app to turn the bathroom lights on...
beautiful illustration - serious question why do ppl still use dashboards? I've been using HA for almost 10 years now and the best part is that everything is automated and I don't have to touch anything.
How do you automate clothing too? My whole family always glance at the temperature and weather on the dashboard before we go out.
Do you automate music selection too? My music taste is very different depending on mood, not to mention my kids.
Or cameras?
I also automate lights but it's pretty nice to be able to control them manually too. Or just pause automations during winter/summer.
Granted I don't use a layout like OP but still.
The illustrator did beautiful work, it's amazing how much different your program became from the original dashboard to one simple home gui.
"When everyone leaves, all the air conditioners switch off."
That's often substantially more inefficient than keeping it at the temperature desired, or adjusting the desired temperature to mitigate it somewhat.
I’m curious why you think that.
In general using a large setback will save more energy by not constantly cooling the home than it costs to return the house to a desired setpoint later.
depends on the duration and temperature differential
This is a common advice I'm not sure I've ever seen properly explained and proven. It's mostly "we asked the experts!" kind of 3min tidbits with a guy in a corporate polo shirt talking assertively to the camera.
Most people actually measuring it come up with more savings when cutting the AC during the longer spans when nobody's there (usually during the hottest part of the day).
This would only be true for very short time periods (minutes) that cause the compressor to power up more times than it would have otherwise as compressor startup incurs transiently high power consumption, but a thermostat will usually do this many times a day, so the period of time that this has any practical implication is very short (<1 hour). At longer time periods, it is always more energy efficient to disable cooling and restart it only when temperature control is again desired.
Even among people who service AC units I've encountered the misconception that it is more energy efficient to keep a space at the desired temperature than it is to let it heat up and then cool it back down again, but thermodynamics say it is not so. It could conceivably be more cost efficient if your energy costs are variable.
I remember the same misconception existing about fluorescent lamps in the past. "They consume more when turning on, so the gains vs incandescent are negated". Which was verifiably false to me at the time by checking the label on a lamp.
Except it's not a misconception.
Air temperature is only one contributing factor for comfort in a home. Surface temperatures take longer to achieve and the larger the gradient the longer it can take.
There are multiple factors to take into account, including the highest temperature the home gets to, that is dependant on location and weather patterns.
If you keep the temperature in the home to 28-30 degrees then later cooling to lower temperatures before bed time you can reduce costs and energy expenditures than if you let it rise to 34 degrees or more and only lowering it when you get home. Or, what you save in costs you expend in otherwise like poor sleep.
Lowering it advance, on a schedule makes more sense if you want surface temperatures to also be reduced.
The blanket disabling of air conditioning when not home is not the only solution, and often is sub-optimal.
No thanks. My measuring equipment says otherwise.
Pretty sure that is completely false, though it is popular to say. It 100% fails even a basic sniff test. The rate of energy transfer from outside to inside will be greatest when the inside is coolest, so the air conditioner will have to work harder overall to keep it that way.
It can be true, but usually not for AC units, as these are quite effective at only heating/cooling the ambient air and not the walls, ceiling etc.
Regular heating can waste a lot of energy to heat the entire structure of the house, especially if you have thick walls, so it does more make sense to try and keep the temperature stable.
The bigger HVAC win, in my opinion, if you have time-of-use billing, is to crank the temp when power is cheap (usually overnight), and, assuming your house is relatively well insulated, letting it ride through the rest of the day. This does lead to some funny things like the fact that I use heavier blankets in the summer than in the winter because I chill the house so far overnight, but for anyone whose climate is relatively moderate and with a relatively modern house, this can save quite a bit on power costs.
Admittedly, this will probably not work that well in a Minnesota winter, or an Arizona summer, but I think a surprising number of people could benefit.
I think it's worth asking, especially with agentic AI tasks that can cost upwards of $20, whether a human could've done a better job on at least one aspect of your project. Fiverr exists.
IDK how much the illustrator was paid (they're Jakarta based and seem like a freelancer), but I think the result was worth it.
The arts have always survived because of patrons, regardless of how favourable or unfavourable the economy was for them. I'm not saying this is the only solution to protecting it from AI, but it's one worth thinking about if you have the resources to do so.
> Fiverr exists
I wonder how many contractors on Fiverr are actually just meat proxies at this point.
Great job and a cracking write up. The notes on making sure you get layers etc right so you can animate effectively is great advice.
Thank you.
I did almost the same but it was completely AI , and started from renovation. it build me material list looked for discounts, saved me ~$2k.
Tricky question how do you know that person did a real drawing, not generated it in codex/claude code?
It doesn't really matter if they're happy with the result does it? Many of us here generate stuff in Claude Code for our employers.
There's still expertise and experience used in prompting, reviewing, refining it - a professional using AI doesn't automatically make it a scam.
I love that you hired an illustrator to make this!
That is lovely. Certainly puts my barebones 2D SVG floor plan to shame. (On the other hand, there's 20+ sensors on there which might be tricky to fit onto something like that image.)
I can't help but notice that the displayed CO2 levels in your example images are quite high. If they're screenshots of the system in action I would change the air in your home more often. If they're just made up numbers, then never mind :)
Interesting
I know of an artist that drew peoples’ houses for a living and had been doing it for decades. People were very proud to hang his work in their actual homes. https://www.aaa.si.edu/collections/surveys/chicago/glessner-...
The author did an excellent job of giving the artist a very detailed brief. Kudos. Put good in, get good out.
Very nice illustration, but when I hear the words "smart home" I reach for my, um, analog kitchen timer.
https://biggaybunny.tumblr.com/post/166787080920/tech-enthus...
I'm inclined to agree.
The "mid-wit" home automation philosophy, I think, strikes a nice middle ground.
https://news.ycombinator.com/item?id=37859437
It's a wide spectrum. It absolutely can mean "if Samsung has a bad day my house becomes uninhabitable."
But speaking as somebody also using Home Assistant, I'd bet the author has everything set up without internet access, fully locally controllable, and entirely operable the old fashioned way if Home Assistant goes down.
where are you based, how much you paid for it ?
They posted a link here which has illustrator’s commission site.
This looks terrific and feels like a project where a lot of love and effort went, so good to read about it in the age of AI
did you try openrealtor.co for a quick floorplan generation? basically a video to floorplan in like 5 minutes
This is incredibly charming. I wish my house evolved at a slow enough pace to be worth doing this to
I want to do this too, great idea!
"I hired an artist but only credit them in a tiny footnote."
this is so cool! i love it.
Perhaps “cool” is the wrong word. Did you notice the temperature recorded inside the house? 24degC (26 in the office). A little too warm for my taste.
Unless it was taken in the summer, in which case it's really cold.
Am I wooshing a joke? That's 75F which is hardly cold in any season.
For summer that's right about the upper end of tolerable range if I don't want to hear the A/C starting/stopping just for a few degrees of cooling.
That's about perfect for my office, keeps my hands limber and avoids carpal tunnel, but not warm enough to sweat. I do like it more like 19C at night for sleeping, though.
I like how this combines two kinds of creativity. One is the pure artistic aesthetic creativity of the illustrator, and the other is the engineering creativity of even thinking to hire an illustrator and building the software for it.
People think of creativity in one shape, but it comes in all shapes and sizes.
The AI-slop writing while talking about hiring a human to do the illustration...
"A flickering fireplace, a softly lit garden, and a smart home the whole family finally enjoys." "not only in my house, but in a hand-drawn picture of it"
Pangram says 90% Human. And this didn't come of as particularly AI to my eyes.
Really reaching witch hunt level "AI" spotting. Saw an ad on instagram the other day where all the comments were "we don't want your AI slop" until the actor in the ad showed up and proved that he was a real human who shot a real video.
Have you tried using AI for the same?
>> Try starting the sprinkler with the button on the unit itself, then running out of the garden before you get soaked.
Go to the store. Buy a valve and some pipe. Mount the valve near the door like any other lazy person. Or, if the outside is too scary, buy an electric valve and wire it to an indoor light switch, maybe even a smart switch so you need not rise from the couch.
This is absolutely incredible and I can't put into words how wonderful this is.
This is amazing!
Really well done, nice work!
Beautiful, and such a cool idea, would love to do this with my house.
This is pretty awesome man. I had zero insight into home automation and this really made me want to learn more. Well done
appreciate you sharing the artist name!
How very inspiring and full of character in the age of worthless slop
Cool, but if you can afford to hire an illustrator to make a custom dashboard for your lights and thermostat, you're clearly in the 1%.
I've commissioned art before via the Krita web forum's job section. Lots of bright young artists, many of them students, that do great work for affordable prices to build their portfolios.
You don't really need to be in the 1% to be a patron if you need some art.
Disposable income and being in the 1% are entirely different things.
I'm not sure what that is supposed to convey. This person was able to scrape together a couple bucks to pay someone to draw a picture of their house in a specific way. You've never paid for any art or entertainment? Netflix?
Calling $900 "a couple bucks" shows how wide a gap there is between perception and reality though.
How much do you they think they paid? By home improvement standards, might not be out of line at all.
I agree it just feels weird to spend on “software” for home improvement, but sometimes it may make a bigger difference than the hardware!
You say it like it’s a bad thing?
And?
Its $100 at most. Thats something majority of people can afford.
*$900, and no, the average person can’t afford $900 for bespoke art (or more likely, has no interest in paying $900 for one). Heck, I don’t see how you can look at the explosion of AI art online and not come to the conclusion the average person would much rather have the same for near free than paying hundreds of dollars. To me, that’s a pretty obvious revealed preference.
The overwhelming majority of people of any income level will never pay to commission any kind of art or even be familiar with the possibility to consider as a practical option.
It's just not a typical thing to do pre or post AI. "Revealed preference" doesn't seem like a useful lens without some basic degree of awareness that a choice even exists.
And on the flip side, I have a friend who commissioned some kind of avatar for social media from an artist friend of friend for the first time in their life. They loved the result a lot and since then have commissioned a bunch of little random things once they realized it was a thing anyone could easily do with a few bucks. For stuff like their WoW character/guild, D&D, social media, etc.
They make less than six figures, each request was <$100 IIRC. Budgeted basically like any other gaming/hobby purchase for personal entertainment.
Idk why people work so cheaply. Less than 100. They're probably working for like 10/hr
But i mean, i dont think he thought of this as an "art-object" more like a ... house-control-kiosk-whatever. And people spend tons of money on those things. Eink-screens for thousands. Control4 for whatever amount. Homey-hub for..300usd? Hue-lights for hundreds etc.
So, sure 900usd or a part of that system is a lot of money and I wouldn't do it. But I've spent 900usd on techthings with a very diminishing amount of return :)
Ever buy any pictures or posters for your wall? Same thing.
I don’t think I’ve ever spent more than…$30 a piece for posters vs $900 for this commission.
Perhaps at some point your budget will allow you to buy or commission something more personal that brings you pleasure to look at every day.
I never said it was lack of budget, it's lack of interest. I'd honestly rather have a $30 print of a Picasso, a patent, a blueprint, or a movie poster than pay someone for custom art, nor do I think one needs to buy or commission something for it to have a personal connection (for example the ratty BTTF poster I had in college, or the Chicago Lyric Opera advertisement I saved out of a nursing home dumpster).
Okay, good for you.
It's ok to have simple tastes! You do you!
I've commissioned tons of art, and one of the great joys is the surprise at how they interpret your idea, and how it's always different than what you had in your mind. That joy is unique to commissioned art, I think.
Wall art in most of America is wooden words from Hobby Lobby that say "Gather" or "Togetherness" and are mass produced by a factory in China.
$900 for digital art isn't exactly cheap, but you sure as hell don't need to be pull a $750K salary to afford it.
https://www.reddit.com/r/homeassistant/comments/1wyyrd6/comm...
Pro tip you don’t even need an illustrator. While I do encourage hiring humans to do this work you can also use PolyCam with a phone that supports lidar, export it to Claude with Blender and voila.
Pro tip the whole point is that it was made by an illustrator, and that's also the reason it's on the frontpage.
Maybe the reason is simply that it's a cool idea?
This is typically the kind of hacking project that is preposterously non-economically-viable for 99% of homelabbers, let alone regular folks. Now, for the first time ever, AI is making it actually palatable for everyone who wants it done to have it done... and certainly much cheaper too at an equivalent quality level.
I'm sorry, I have a hard time understanding you. Correct me if I'm wrong, are you saying that
1. paying an illustrator is non-economically-viable for 99% of homelabbers? and
2. AI would have done it at an equivalent quality level?
Yes to both.
For (1) the vast majority of people don't want to spend what's probably on the roundabout of a thousand dollars for their homelab's homepage. You can easily see this in almost every homelabbing discussion on reddit or even here, where everyone focuses on cheap/hackish solutions to avoid overspending... except in hardware of course =) It's kind of become a meme at this point.
For (2), certainly, especially so considering that the author gave the illlustrator blueprints, photos, instructions, etc. for the commissioned illustrator to convert into dynamically-compatible svgs.
There's a lot of top-down implementing there with very low bottom-up artistry. This doesn't at all take from the final quality -which is pretty good- but it makes it easily replicable using current-gen AI.
Commissioning art like this is at most a few hundred dollars. As was posted elsewhere, there are lots of young artists that need to pad out their portfolios, as well as artists in places where the cost of living is significantly cheaper. This isn't a multi-thousand-dollar thing that only the 1% could afford.
It's _better_ that the illustrations were made by an illustrator, but it still would have been cool and front-page-worthy if that hadn't been the case.
Not really. AI slop is an instant turn-off for me. It's like hearing someone talk about a dream they had; it is egregiously boring for everyone except the person relating it.
Most people don't have lidar phones, and commissioning this art is about the same price as buying a lidar phone.
> Most people don't have lidar phones
the most popular phones in the US (iPhone pro max and pro) have had lidar cameras for the past 6 models...
Most people don't have lidar phones. The statement is still very true, even if you said US only. What are you trying to say?
Some quick googling suggests that in the US, only 60% of people have iphones, and only 50% of iphones in use are pro or max. I would guess that HN skews towards more recent and higher end phones, but also skews towards android. Maybe saying most people don't have lidar is inaccurate, but I think it's still accurate to say the majority of people don't have lidar.
Regular photogammetry is probably sufficient for this sort of project.
Damn that is a good tip. Thanks, dude.
Ugh
"Pro tip, you can generate soulless slop to avoid supporting your fellow humans" is not a very compelling case.
This is beautiful. Reminds me floor796.
I used AI to get it draw it. What worked is giving floor plan I already had and putting some labels on it - brown dining table here, fridge here, etc.
What didn't work was giving it bunch of photos and asking it to just figure it out. For things it did figure out - it was nice, but it made way too make mistakes.
The entire point of it was to put light switches and heaters on the map so I don't have to fiddle with a number of toggles in a table/list.
p.s. I wish I had huge eink display for home assistant dashboards. Some concepts I made and seen online are glorious.