Civil engineering is already a field where the very largest projects are done by humans planning and building the roads and bridges for the robots to move in (such as things rented from Mammoet [1] with extra control systems), but it does require significant human oversight (typically a metaphorical red button).
It's all very one off and specific, and given how big those projects are that seems unlikely to change. The manufacturing of suburbs though would be a whole different ballgame.
Is this a scenario where offshore operators doing remote equipment control would be 90% as good as a US union worker for 15% of the price and could work in shifts 24/7 (e.g. for mining operations)? Sensor + control data would be great training for future AI.
Jumping straight to autonomous operations seems expensive/hard.
I didn't think open field construction was hampered by the humans in the loop? Quite the contrary, I was under the naive impression that the heavy machinery was already largely doing the vast majority of the work. Even when operated by a human.
Will be neat to see where this goes. But I'm reminded of some Amazon guys that were supposed to revitalize the supply chains. My memory is that that didn't work out so well.
I think that more than physics the bottleneck for this is political (at least in the US). All of the local large projects around me are expensive because of massive amounts of red tape (environmental studies, zoning, planning), and political patronage systems. After the kick backs, political donations, promises to only work 8 hours a day, only use union labor, hire x police officers for y hours in overtime security positions a month, use xyz contractor etc. a small cost seems to be the actual labor and materials. Hell these robots if they work will be made illlegal.
Tech can help with the politics as well. More construction would happen if it was technologically possible for it to be so fast that major projects could be finished in a single election period.
Pre-fab is illegal in a lot of places too. Pre-fab is for the poors! Houses in this neighborhood must be assembled on site by lumber by a team of expensive humans from raw lumber, not at a factory churned out by the hundreds, shipped on site on a truck, and quickly put together.
Like yeah, typical pre-fab, where it's legal, is going to go into a trailer park or something, but nothing says pre-fab has to be cheap and crummy. Why are we still cutting lumber on site?
This is a bit like saying "if everyone was just honest and a good person, we wouldn't need police or jails or criminal judges". All these things exist for a reason (except for kickbacks). We have labor unions because of worker exploitation. We have zoning to protect personal freedoms of existing property owners. We have environmental impact studies because corporations have proven they are willing to freely dump toxic materials wherever they feel like.
If you read the other comments made by this user they are exactly the political leaning you expect, including support for the BBB among other similar opinions.
> Bedrock Robotics is focused on developing a self-driving kit that can be retrofitted to construction and other worksite vehicles
> Bedrock is “upgrading existing fleets with sensors, compute, and intelligence that understands project goals, adapts to changing conditions, and executes work around the clock,”
I can also imagine this applying to all kinds of mining too, where there's already all the heavy equipment to mine and transport resources and we're just turning it into a robot so they don't have to employ a human anymore.
One big barrier I haven't seen mentioned is all the OEM competition they are going to face.
Caterpillar, John Deer, etc. already have remote operation vehicles. And a lot of provisions on what types of kits can be retrofitted onto their equipment without violating their terms/warranties.
I'm sure this is already something they've taken into consideration, but it seems like this will be more focused on partnerships with existing OEMs rather than selling add on kits to current fleets.
Definitely echo the concerns about bureaucratic red tape (looking at you, California high-speed rail fiasco) that kind of environment makes innovation in infrastructure extra hard. Still, there's something compelling about Bedrock’s approach if they can genuinely deliver a system that adapts in real-time to the complex nature of construction sites. The big question is whether they can win through retrofitting existing fleets or by locking in tight partnerships with OEMs, adding to that the competition is going to be pretty tough
Unrelated apart from the words "rock" and "construction": I wouldn't be surprised if we see dry stone construction becoming very practical thanks to advances in computer vision and robotics.
Dry stone construction is incredibly durable -- it doesn't rely on mortar which can weather away -- but it is limited by needing to reshape stones to fit together tightly (often by making flat surfaces). A human stonesmith can look at a handful of stones and find one which is close to fitting in the necessary spot; but a computerized system could scan thousands of stones and build tightly-fitting stonework with minimal need for reshaping.
Eg. if bricklayers could talk to their machine the way we can with coding agents, and say “yep, wall here please, check the blueprints to confirm how high and where the holes for the windows go”, retired & injured tradespeople could choose to come back. Less injuries means cheaper insurance & better margins. People could work in multiple parts of a site by supervising several robots, and not be exhausted at the end of the day. The list of benefits to individuals and the industry is long.
Well said. The company is actually more focused on earth moving, but the same arguments you make apply there, if not even more. Earth moving is a very exhausting job that most would be glad to have automated. A little known fact is that you actually get boiled no matter what machinery you use in there on the black dirt mid day, and it's not totally safe; your machine could roll over if you aren't at the height of your awareness.
To build a commercially successful autonomous bulldozer requires building a commercially successful bulldozer. That’s hardware and hardware is hard. Probably harder than the autonomous part because bulldozers are a century past the proof-of-concept era.
My cynical take is this is financial engineering more than construction engineering. YMMV.
74 comments
[ 2.9 ms ] story [ 62.2 ms ] threadCivil engineering is already a field where the very largest projects are done by humans planning and building the roads and bridges for the robots to move in (such as things rented from Mammoet [1] with extra control systems), but it does require significant human oversight (typically a metaphorical red button).
It's all very one off and specific, and given how big those projects are that seems unlikely to change. The manufacturing of suburbs though would be a whole different ballgame.
[1] Specifically https://en.wikipedia.org/wiki/Self-propelled_modular_transpo...
Jumping straight to autonomous operations seems expensive/hard.
Will be neat to see where this goes. But I'm reminded of some Amazon guys that were supposed to revitalize the supply chains. My memory is that that didn't work out so well.
all the labor is imported from abroad and live in miserable conditions, no unions or paperwork, just lots of earthworks to be done
Why are heavy machines like diggers and cranes legal then? Human operated, but replace countless workers on the site.
Like yeah, typical pre-fab, where it's legal, is going to go into a trailer park or something, but nothing says pre-fab has to be cheap and crummy. Why are we still cutting lumber on site?
And an $80M round sounds sane these days
> Bedrock is “upgrading existing fleets with sensors, compute, and intelligence that understands project goals, adapts to changing conditions, and executes work around the clock,”
I can also imagine this applying to all kinds of mining too, where there's already all the heavy equipment to mine and transport resources and we're just turning it into a robot so they don't have to employ a human anymore.
Caterpillar, John Deer, etc. already have remote operation vehicles. And a lot of provisions on what types of kits can be retrofitted onto their equipment without violating their terms/warranties.
I'm sure this is already something they've taken into consideration, but it seems like this will be more focused on partnerships with existing OEMs rather than selling add on kits to current fleets.
Seems sensible a project. $80m raised also seems a sensible amount. And the guy has a background in this field. Good luck
> adapts to changing conditions
The real play here is starting a business that specializes in getting construction equipment unstuck from the mud.
Dry stone construction is incredibly durable -- it doesn't rely on mortar which can weather away -- but it is limited by needing to reshape stones to fit together tightly (often by making flat surfaces). A human stonesmith can look at a handful of stones and find one which is close to fitting in the necessary spot; but a computerized system could scan thousands of stones and build tightly-fitting stonework with minimal need for reshaping.
Eg. if bricklayers could talk to their machine the way we can with coding agents, and say “yep, wall here please, check the blueprints to confirm how high and where the holes for the windows go”, retired & injured tradespeople could choose to come back. Less injuries means cheaper insurance & better margins. People could work in multiple parts of a site by supervising several robots, and not be exhausted at the end of the day. The list of benefits to individuals and the industry is long.
To build a commercially successful autonomous bulldozer requires building a commercially successful bulldozer. That’s hardware and hardware is hard. Probably harder than the autonomous part because bulldozers are a century past the proof-of-concept era.
My cynical take is this is financial engineering more than construction engineering. YMMV.