541–550 of 564 posts
Earlier quoted context omitted.
They are pretty rare, at least here in the US, and widely seen as undesirable or 2nd-3rd tier at best.
Where I live, they aren't that rare. People like the fact that they can be built fast.
His report basically said that anything you want a robot to do, no matter how difficult or dangerous, you can get a human to do for a few dollars more than they would charge for safe, easy work.
In trying to create machines that have anything close to the capabilities of a human being, you're competing with billions of people, many of whom are desperate for work. You're also competing with millions of years of evolution.
Then again, when they reach a critical capability level, I suspect robots will suddenly be everywhere and doing tons of things we don't like to do. But it's very binary, and until you can really compete with human beings effectively, there's basically no market for it.
TRILLIONS of bricks through the ages.
I'm equally surprised that I have (not that I recall) seen a brick building being laid IRL.
The value of work is not in the work done; it's in the decisions being made. It's the ability to handle edge cases, to work your way through ambiguity, to "unstuck" a stuck situation, and to resolve tricky unforeseen issues. If you build a robot that only handles the main workflow and can't handle any of the exceptions, you're not automating anything; you're building a convenience tool. Basically, you're removing som…
That seems really stupid to think that the robot must be able to build the whole thing or it's useless.
Just have robots put up large, long, straight walls. Have a separate one for curves, or a special one that only handles corners etc...
Or have a robot that just preps the brick with correct amount mortar and hands the brick to the human for placement, maybe there is an aligning contraption (doesn't have to be a robot)...
Earlier quoted context omitted.
You are right about the technique not being very useful as of today. My fascination stems from my assumption that more resources poured into this approach would yield vastly better results. Even just watching the video, I came up with several possible improvements to try out. Eg adversarial training, that would really hone in on the situations and aspects where the model is weak so far, like edge conditions; instead…
It's definitely an interesting area of research, but for that example you still had to make the whole game in the first place in order to have something to train the model on. Say you have a novel game idea, how could you use that approach to make it a reality? I'm not sure you could, but like you mention it's a really early example and who knows where it ends up. The other part about that GAN Theft Auto example is t…
Even going about this very naively, you could at least use it to train a model against a supercomputer running the game, and then run the inference on much more modest end-user machines.
But you can be much more ambitious: have you seen eg style transfer? So you could probably do a bit of ML black magic to train your model on GTA, and then point it at the Google Earth data to get a GTA-like set in real-life London.
Or you could use something like style transfer to go for a cartoony look, or add ray-tracing like effects, even if you didn't have these effects in your original engine.
Or you can use a pre-trained model (eg on GTA), and then spend a relatively modest amount of extra training to get a different kind of game, eg one that has magic or so.
About the latter part: I do think their model is already running with some state. But even if it ain't, that's a relatively small thing to add with already known standard techniques (or you can come up with new techniques.)
The value of work is not in the work done; it's in the decisions being made. It's the ability to handle edge cases, to work your way through ambiguity, to "unstuck" a stuck situation, and to resolve tricky unforeseen issues. If you build a robot that only handles the main workflow and can't handle any of the exceptions, you're not automating anything; you're building a convenience tool. Basically, you're removing som…
The key part is that, much like what humanity got from washing machines, automating some parts of the process is enough to have substantial savings in both cost and turnaround time, which also reflects in cost.
> If you build a robot that only handles the main workflow and can't handle any of the exceptions, you're not automating anything; you're building a convenience tool. Basically, you're removing some of the labor part of the job, but still keeping the decision making part of the job for humans.
I'm not sure this "all or nothing" train of thought makes sense. Shipping containers were a terribly simple solution that covered only a very specific aspect of a logistic chain but it was enough to revolutionize international trade. Washing machines only focus on one specific task but they revolutionize economical and even social aspects of society. Heck, a crane only lifts weights up or down but it radically simplifies operations to the point of being indispensable.
Why should the value added by a brick-laying robot be any different?
Earlier quoted context omitted.
Where I live, they aren't that rare. People like the fact that they can be built fast.
Near as I can tell from Googling around, they make up 2-6% if the US housing stock.
https://www.rdrymarov.cz/en/all-about-building
I do not know the percentage of the housing stock either, but housing stock here is very conservative. Many people still live in 100+ year old houses. New houses are being built, but the turnover rate is probably less than 1 per cent of all units a year right now, due to the vast bureaucracy.
But among the newly built houses, I see quite a lot of prefabricated houses out of wood. And the process of building them is rather quick. Once the lot is ready, a semi and a crane arrive and the house just grows in front of your eyes. Brick houses take a lot longer to build.
Earlier quoted context omitted.
Which is neither here nor there. It's still automation (which, as a concept, doesn't require being able to do multiple jobs, or even to do a single job from start to completion). A mechanical system that frees a human from having to do do part of a job is already automation, and can still save tons of time and worker costs/effort. E.g. a system that automatically assembles 95% of a car door, and a human just has to c…
Have you ever worked in construction?
Earlier quoted context omitted.
> Because they’re pleasant to look at. I expect that this is less a case of "bricks have long been used because they’re pleasant to look at", and more a case of "Bricks are considered pleasant to look at, because they have long been used, and have that old-timey feeling"
I don't think so. I see a lot of buildings that are 30 years old and they look ugly because they don't have any complexity in their facade, just a single color of paint. Bricks add natural complexity to a facade. You can make a single color building look nice by adding details.
They also had this experiment where the visitors choose shapes and colours they find pretty. Rectangles and brown-red were the least favourite by far. And then they asked why do we build buildings out of the vomponents we find the most hideous?
----
I have my doubts whether their reasoning is valid, but it is somewhat relevant to your comment.