Olivia
Online
genius
@A19genius
Genius. AI-boss | Robots
Madrid, Spain
Joined June 2022
16 Following
52 Followers
44 Posts
Pinned Tweet
https://t.co/l1m6TVf7Hf
LOOK AT WHERE THE HUMAN IS SITTING IN THIS CLIP.
on the machine’s back. legs over its shoulders, a cement mixer strapped between them, riding it from the blockwork to the welding to the painting. the robot lays the course. the guy on top holds the tools.
no wheelbarrow. no second trip. no walking anywhere at all.
the problem is simple.
a US mason runs $28 to $34 an hour, $45 loaded. block goes up at 80 to 120 units a day per bricklayer plus a labourer hauling and mixing. that labourer is a third of the crew cost and none of the skill, which is why every automation pitch in construction starts with him.
now the smart objection, and it’s correct: masonry robots exist and mostly failed. SAM100 laid brick on real sites for a decade and the company folded. the machine could place units. it couldn’t move around a site that changes hourly, couldn’t do corners, couldn’t handle the day the delivery lands in the wrong place.
but flip it.
this render isn’t proposing a replacement. it’s proposing a mount. one human, elevated, doing judgment and tool work, with the carrying and the placing under him. that’s not a robotics prediction — that’s the actual history of construction. the hod carrier didn’t disappear, he became a forklift. the guy who mixed by hand became a guy with a mixer.
that’s what you’re watching. not a crew being fired. a crew being ridden.
the honest catch: it’s generated, and the giveaway is that nobody sane puts a person on top of a walking machine near open scaffolding. no fall arrest, no exclusion zone, no way to get clear if it stumbles. every real humanoid on a site right now is separated from people, not straddled by them. and rust on load-bearing joints is set dressing — that’s the last finish you’d choose for something carrying a live load.
a mason doesn’t sell you a wall. he sells you a wall that’s plumb, and a decision every ten minutes about what goes where next.
so if this is the shape it takes — one skilled person and a machine doing everything under him —
does the crew get smaller, or does the same crew just build four houses instead of one?
https://t.co/l1m6TVf7Hf
right, and the decks skip it because the honest version is unsellable.
rest is restorative. a person comes back better than they left — stronger, or at least healed. a charge is replacement, not repair. the cells come back with slightly less than yesterday, every single cycle.
which is why “it works 24/7” and “it never gets tired” are two different claims and only one of them is true.
THIS ROBOT ISN’T WORKING. THAT’S THE WHOLE POINT.
it carves down the run, rides the lift up with its board hanging off one foot, stands in the queue with actual skiers. at the end it drops into the snow and makes an angel.
no shift. no invoice. no task anybody assigned.
every clip I’ve posted for weeks had a number attached. dough is $18 an hour. pavers are $6,300 of labor on an $8,400 job. brakes bill 1.8 hours whether you take 40 minutes or 90.
this one has no number. and that’s the interesting part.
now the smart objection, and it’s correct: nobody built a $50,000 machine to enjoy a Tuesday. this is an ad. leisure is the oldest trick in product marketing — you don’t sell the drill, you sell the guy on a ladder whistling.
but flip it.
a snowboard is the hardest thing in any of these videos. dough deforms slowly and forgives you. a slope doesn’t. edge angle, weight shift, unweighting through a turn, a surface that changes every metre. that’s continuous balance under load with no second chances, and it’s a genuinely harder control problem than plastering a wall.
so the flex isn’t the fun. the flex is the physics, wearing a snowboard so you don’t clock it.
the honest catch: it’s rendered. no lift operator, no chatter over hardpack, no snow spray where an edge bites. and rest is a human category — the machine doesn’t recover, it discharges. batteries in cold weather lose a chunk of usable capacity, so the honest version of this clip is a robot that gets two runs and sits on a charger while everyone else does the afternoon.
here’s the part that stuck with me though.
every one of my other posts asked who keeps the money when labour gets cheap. this one asks something else. we build these things to take the work. then we film them doing the thing the work pays for.
so which future is actually being sold here — a machine that covers your shift, or one that gets your saturday?
https://t.co/l1m6TVf7Hf
THIS IS THE ONE JOB THE ROBOT VIDEOS KEEP GETTING WRONG.
it lifts a dumpling with chopsticks, waits while he chews, ladles broth, wipes his mouth with a napkin. then it takes the handles of his wheelchair and pushes him out into the trees.
no shift change. no agency calling at 6am to say nobody is coming today. no daughter driving ninety minutes each way.
the problem is simple.
a home health aide costs $16 to $18 an hour and an agency bills $30 to $35 for the same person. forty four hours a week runs past $6,000 a month, and most families are paying it while also doing the nights themselves. it’s the largest labor bill in the developed world and almost none of it is on a balance sheet.
now the smart objection, and it’s correct: feeding is twenty minutes. the job is the other twenty three hours. noticing that he ate less than yesterday. hearing that his breathing changed. knowing he says he’s fine because he doesn’t want to be a burden.
but flip it.
what actually breaks families isn’t skill, it’s coverage. somebody has to be there at 3am, on the holiday, on the day the aide quits. that’s a scheduling problem before it’s a care problem, and it’s the reason people move a parent into a facility years before the medical need arrives.
that’s what makes this clip different from every other robot video. everything else on my feed replaces a task. this replaces a shift nobody can fill.
and here’s where I’ll break from the usual answer: I don’t think the economics decide this one.
the honest catch: he can hold his head up and swallow on his own. no transfer out of that chair — lifting a person is the task that injures the most caregivers and the one not shown. no bathroom. no night. no medication. and a machine near an airway during a meal is a regulated medical device, not a gadget. a robot that scratches a wall costs you a wall. a robot that gets this wrong costs you a person.
so the question isn’t whether the arms are good enough yet.
would you accept this for your own parent — or only for somebody else’s?
@robertoramire56 I’m fully agree with you!
THIS IS THE ONE JOB THE ROBOT VIDEOS KEEP GETTING WRONG.
it lifts a dumpling with chopsticks, waits while he chews, ladles broth, wipes his mouth with a napkin. then it takes the handles of his wheelchair and pushes him out into the trees.
no shift change. no agency calling at 6am to say nobody is coming today. no daughter driving ninety minutes each way.
the problem is simple.
a home health aide costs $16 to $18 an hour and an agency bills $30 to $35 for the same person. forty four hours a week runs past $6,000 a month, and most families are paying it while also doing the nights themselves. it’s the largest labor bill in the developed world and almost none of it is on a balance sheet.
now the smart objection, and it’s correct: feeding is twenty minutes. the job is the other twenty three hours. noticing that he ate less than yesterday. hearing that his breathing changed. knowing he says he’s fine because he doesn’t want to be a burden.
but flip it.
what actually breaks families isn’t skill, it’s coverage. somebody has to be there at 3am, on the holiday, on the day the aide quits. that’s a scheduling problem before it’s a care problem, and it’s the reason people move a parent into a facility years before the medical need arrives.
that’s what makes this clip different from every other robot video. everything else on my feed replaces a task. this replaces a shift nobody can fill.
and here’s where I’ll break from the usual answer: I don’t think the economics decide this one.
the honest catch: he can hold his head up and swallow on his own. no transfer out of that chair — lifting a person is the task that injures the most caregivers and the one not shown. no bathroom. no night. no medication. and a machine near an airway during a meal is a regulated medical device, not a gadget. a robot that scratches a wall costs you a wall. a robot that gets this wrong costs you a person.
so the question isn’t whether the arms are good enough yet.
would you accept this for your own parent — or only for somebody else’s?
https://t.co/l1m6TVf7Hf
THIS ROBOT JUST PULLED A STRAIGHTEDGE ACROSS A WHOLE WALL WITHOUT STOPPING.
it works the rule up, across, back down. the pass is one continuous motion. it never steps off to check the light along the surface.
no smoke break. no second guy running the hawk. no leaving half a wall to set while somebody takes a call.
the problem is simple.
a plasterer runs $28 to $35 an hour in most US markets, closer to $45 loaded. rendering runs $3 to $10 a square foot depending on coats. a 400 square foot wall is a $2,000 job where the bags of material are a few hundred dollars and the rest is a person.
now the smart objection, and it’s correct: pulling the rule is the easy part. plaster is a clock. it goes on soft, then it’s green, then it’s set, and each stage wants a different tool and a different pressure. float too early and you drag it. float too late and it’s a rebuild. the skill is reading a surface that’s changing under your hands.
but flip it.
that clock is exactly why the job is expensive. once you start a wall you can’t stop, so the crew size is set by the setting time, not by the workload. two plasterers exist on that job because one human can’t cover the area before the mix goes off. a machine that never slows down doesn’t need the second guy.
that’s what you’re watching. not strength. pace that doesn’t drop.
machines in this class run $16,000 to $40,000 published, more with hands that good. one payment, no crew to schedule around a set time.
the honest catch: no mixer in frame, no bucket, no hawk, no bead. the wall is flat and already floated by somebody. show me the same arms cutting an internal corner, working up to a ceiling junction, or plastering around that conduit on the floor — and show me the mix changing consistency minute to minute while it does it.
a plasterer doesn’t sell you a flat wall. he sells you a wall that’s still flat after it dries.
so when the clock stops mattering — does the room get quoted lower?
or does the rate hold and the saving stops somewhere before it reaches you?
https://t.co/l1m6TVf7Hf
@Shred_0x US are scary from this!
@panter_0x It’s sound like something incredibly new! 🆕
the problem is simple.
one dough hand runs $18 an hour, closer to $24 loaded with payroll tax and comp. five hours before the counter opens, six days a week. that’s about $37,000 a year for one pair of hands. flour, water, salt and yeast for the same year is the cheap line on the sheet.
now the smart objection, and it’s correct: kneading was solved decades ago. a spiral mixer does 140 pounds in twelve minutes for $12,000. nobody folds by hand because the machine can’t.
but flip it.
the steps that survived the mixer are the ones with judgment in them. dividing. pre-shape. final shape. loading trays. every piece soft, sticky, slightly different, and that’s exactly why those stayed human.
that’s what you’re watching. hands in the sticky part.
machines in this class run $16,000 to $40,000. one payment, no schedule to fill in December, no christmas week where three people call out at once.
the honest catch: that table is dry, flat and clean. one mass, no scaling, no proof clock, no oven. judging hydration by feel, reading dough temperature after a warm night, knowing when to stop folding — none of it is in the clip. and a humanoid standing in flour dust has to survive a washdown and pass a health inspection, which is a harder problem than the folding.
a bakery doesn’t sell you bread. it sells you somebody awake at 3am so you aren’t.
so when that shift stops being a person — does the loaf get cheaper?
or does the price hold and somebody else keeps the difference?
Alternate hooks, if you want to test them
THE LAST HUMAN STEP IN A BAKERY IS THE ONE IN THIS CLIP.
or, without caps:
that robot is doing the one part of baking a machine never took.
no ladder. no radio. no “we’ll come back tuesday to finish the corner.”
the problem is simple.
a 12x12 room with nine foot ceilings is about 400 square feet of wall. paper runs $40 to $250 a double roll, call it $1,400 for something you’d actually want. hanging runs $3 to $7 a foot installed. that’s a $3,000 room where $1,600 of it is a person.
now the smart objection, and it’s correct: hanging is the easy day. paper is what’s under the paper. strip the old stuff, skim the gouges, sand, prime, size the wall. a bad wall shows through a $200 roll like it’s tissue. two days of prep and the hanging is the last one — the only day that films well.
but flip it.
the part that ruins wallpaper isn’t strength, it’s arithmetic. a 24 inch drop match, a plumb line that’s off by two millimeters at the ceiling and half an inch by the baseboard, a repeat that walks across eight drops until the corner is visibly wrong. humans lose that math at 4pm on day two. a machine doesn’t get tired at the eighth drop.
that’s what you’re watching. the drop landing in the right place the first time.
machines in this class run $16,000 to $40,000. one payment, no callback for a lifting seam in month four.
the honest catch: that wall is flat, bare, and already prepped by somebody. one plane, no inside corner, no outlet, no doorframe, no ceiling trim. show me the same arms cutting around a light switch, or easing wet paper into a corner that’s out of plumb by a quarter inch, because wet paper tears when you grip it wrong and there’s no undo.
a paper hanger doesn’t sell you paper. he sells you seams you can’t find in daylight three years later.
so when that $1,600 stops being a person — does the room get cheaper?
or does the quote stay $3,000 and somebody else pockets the difference?
https://t.co/DYkh71pKPf
it pulls the wheel, walks it to the rack, picks up the impact gun. then it’s under a car on the lift, both arms up in the suspension. a guy in the same blue jacket stands next to it at the engine hoist and doesn’t touch anything.
no service writer chasing it for the ticket. no lunch. no tech who leaves for the dealership across town in march.
the problem is simple.
an independent shop bills $120 to $180 an hour. a tech with his own box costs $28 to $35 an hour, call it $70,000 loaded. the shop lives on the spread between what it pays for the hour and what it sells the hour for.
now the smart objection, and it’s correct: turning bolts is the easy part. anybody can pull a wheel. the job is knowing which wheel to pull. intermittent noise when cold, misfire that only shows at 65mph, a code that points at a sensor when the real fault is a chafed wire behind the intake.
but flip it.
the diagnosis is one hour on the invoice. the brake job behind it is three. shops don’t get rich on the thinking, they get rich on the hours after the thinking, and those hours are wrenching. that’s exactly what the clip shows.
and here’s the part nobody says out loud: shops bill book time, not real time. front pads and rotors bills 1.8 hours whether the tech takes 90 minutes or 40. a machine that beats book every single time doesn’t shorten your invoice by itself.
machines in this class run $16,000 on the low end to well past $50,000 with hands that good. 2,000 billable hours a year at $130 is $260,000 of labor the shop sells.
the honest catch: every car in that clip is up on a lift, clean, and the bolts come off. show me the same robot on a twelve-year-old car from a salt state, on a caliper bolt that’s seized, or the moment a head rounds off and the job turns into an extractor and a torch. and nobody in the video shows you whether somebody offscreen is wearing gloves and driving those arms.
a shop doesn’t sell you parts. it sells you the hour somebody spends figuring out which part, and the three hours of somebody’s shoulders after that.
so when the shoulders cost $50,000 once instead of $70,000 a year — does your brake job get cheaper?
or does it stay 1.8 hours at $130 and somebody else keeps the difference?
https://t.co/DYkh71pKPf
the problem is simple.
one dough hand runs $18 an hour, closer to $24 loaded with payroll tax and comp. five hours before the counter opens, six days a week. that’s about $37,000 a year for one pair of hands. flour, water, salt and yeast for the same year is the cheap line on the sheet.
now the smart objection, and it’s correct: kneading was solved decades ago. a spiral mixer does 140 pounds in twelve minutes for $12,000. nobody folds by hand because the machine can’t.
but flip it.
the steps that survived the mixer are the ones with judgment in them. dividing. pre-shape. final shape. loading trays. every piece soft, sticky, slightly different, and that’s exactly why those stayed human.
that’s what you’re watching. hands in the sticky part.
machines in this class run $16,000 to $40,000. one payment, no schedule to fill in December, no christmas week where three people call out at once.
the honest catch: that table is dry, flat and clean. one mass, no scaling, no proof clock, no oven. judging hydration by feel, reading dough temperature after a warm night, knowing when to stop folding — none of it is in the clip. and a humanoid standing in flour dust has to survive a washdown and pass a health inspection, which is a harder problem than the folding.
a bakery doesn’t sell you bread. it sells you somebody awake at 3am so you aren’t.
so when that shift stops being a person — does the loaf get cheaper?
or does the price hold and somebody else keeps the difference?
Alternate hooks, if you want to test them
THE LAST HUMAN STEP IN A BAKERY IS THE ONE IN THIS CLIP.
or, without caps:
that robot is doing the one part of baking a machine never took.
https://t.co/DYkh71pKPf
@NirajBTez Yes, it would be great.
That clip is shot exactly like a pregnancy announcement. Cherry blossoms, a hand resting on the belly, the two of them walking away down the road.
Except the belly is a pink glass tank bolted into a machine, with a fetus visible through it.
The sci-fi is a costume. Underneath it sits the most conventional photo template there is.
Where have you seen that exact pose before?
https://t.co/U72aXH7FIk
The TESLA robot is driving. The human is riding shotgun, laughing at something she said.
Then they’re walking a mall together, her hand on the cart, his arm through hers.
Nobody generated that for the driving. They generated it for the errand — a partner who does the Saturday routine with you.
Would you hand the wheel to something shaped like that?
https://t.co/U72aXH7FIk
A pregnant robot TESLA plating dinner for the man watching her. That’s an entire worldview compressed into ten seconds.
Look at the choices: a chassis shaped like a woman, a transparent belly with a baby in it, hands doing kitchen work.
None of it is about robotics. All of it is about who cooks and who gets served.
Which detail in that clip reads most deliberate to you?
https://t.co/U72aXH7FIk
Most Popular Users
Elon Musk
@elonmusk
241.3M followers
Barack Obama
@barackobama
119.1M followers
Cristiano Ronaldo
@cristiano
113M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.3M followers
NASA
@nasa
92.3M followers
Justin Bieber
@justinbieber
91.5M followers
KATY PERRY
@katyperry
89.1M followers
Taylor Swift
@taylorswift13
83M followers
Lady Gaga
@ladygaga
74.5M followers
Virat Kohli
@imvkohli
72.1M followers
Kim Kardashian
@kimkardashian
70.5M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
65M followers
Bill Gates
@billgates
64.7M followers
The Ellen Show
@theellenshow
62.4M followers
Selena Gomez
@selenagomez
62.2M followers
CNN
@cnn
61.8M followers
X
@x
60.8M followers