Olivia
Online
Aurelian
@aurel1ans
A relic of a forgotten age
Joined April 2013
84 Following
94 Followers
16 Posts
@explosss1ve BORK could be a huge memecoin
at 98, philip carret was asked to explain how he survived 75 years on wall street, he answered with one word
April 28, 1995. Louis Rukeyser was sitting across from a man who had begun investing in 1919, founded the Pioneer Fund in 1928 and watched the market collapse one year later.
By the time Carret appeared on Wall Street Week, he had lived through the Great Depression, World War II, Korea, Vietnam, inflation and recessions. Warren Buffett said he had “the best truly long-term investment record of anyone I know.”
Rukeyser asked him to compress three quarters of a century in markets into the single most important lesson.
“Patience.”
It sounds almost too clean, but Carret’s version of patience required preparation. He wanted healthy balance sheets and reasonable earnings, with as little debt as possible. After a margin call in 1924, he swore off borrowed money and later said that decision was one of the main reasons he made it through the 1930s.
He also admitted that the market still surprised him. He couldn’t reliably call its next move, so he owned companies that didn’t require him to.
That is the part modern investors keep trying to engineer away.
Everyone wants a better forecast. Carret spent 75 years building something more useful: the ability to be wrong about timing without being forced out of the game.
Patience works much better when your portfolio can survive it.
https://t.co/Mqhcmc3NQD
A ROBOT JUST BEAT THE HUMAN HALF-MARATHON WORLD RECORD. IT ALSO CRASHED INTO A RAILING NEAR THE FINISH LINE.
april 19 2026, beijing. a humanoid named lightning, built by honor, ran 21.1km in 50:26.
the human world record is jacob kiplimo's, set in lisbon a month earlier. lightning beat it by nearly seven minutes.
near the finish, it hit a railing. a handler caught it, stood it back up, and it finished anyway.
that detail is the one nobody puts in the headline, and it's the one that actually tells you something. this isn't a general-purpose robot that happens to run fast. it's a machine built for exactly one task, with 95cm legs and a liquid cooling system designed in-house, that still needed a human standing trackside in case it fell over.
a year earlier, the same race humiliated this category completely. 21 robots started. 6 finished. the winner took 2 hours 40 minutes, and every team on the course had trainers running alongside with duct tape and battery packs.
this year: 112 teams, 300+ robots, 40% of them navigating without remote control for the first time.
the honest read isn't "robots are faster than humans now." it's that a narrow, specialized machine with no other job than this one can already beat elite humans at their own specialty, the same way a bicycle always could. general-purpose is still the hard problem. this was the easy one, solved in a single year, with a person still standing by the rail.
https://t.co/QEbyuIBse3
MIT ACCIDENTALLY TURNED A 2008 FINANCE COURSE INTO A RECORDING OF A FINANCIAL SYSTEM BREAKING IN REAL TIME
Andrew Lo was teaching Finance Theory I at MIT Sloan in the fall of 2008. Lehman Brothers filed for bankruptcy on September 15, markets were convulsing, and concepts his students were supposed to learn from textbooks were suddenly unfolding outside the classroom.
That makes the course almost impossible to recreate today. Lo teaches present value, bonds, leverage, portfolio theory, risk and efficient markets while the biggest financial crisis in generations keeps entering the conversation.
And some of the lessons hit very differently when you know what was happening outside.
Leverage makes a good trade look brilliant on the way up and makes being slightly wrong lethal on the way down. Diversification works until positions that looked unrelated suddenly start moving together. An asset can have a perfectly reasonable theoretical value and still become a disaster when nobody wants to buy it from you.
The nastiest lesson is liquidity.
You can be right about what something is worth and still get destroyed if you need cash before the market agrees with you.
That is why watching this in 2026 feels different from watching another finance course recorded after the fact. Nobody in that room had the luxury of knowing exactly how the crisis would end. They were learning the machinery while parts of that machinery were failing in public.
MIT put the entire course online for free.
There are thousands of videos explaining what went wrong in 2008.
Very few let you sit inside a finance classroom while it was going wrong.
https://t.co/3hWKMIBJcw
WE BUILT HUMANOID ROBOTS AND SOMEHOW SENT THEM STRAIGHT BACK TO THE POTATO FIELD.
No perfect lab floor. No staged warehouse demo.
This thing is standing in actual dirt, swinging a hoe around and digging potatoes out of the ground like somebody’s grandfather on a Saturday morning.
The weirdest part is how normal it already looks.
We spent decades imagining humanoid robots building spaceships.
They might end up harvesting dinner first.
A SINGLE INJECTION MOLD CAN COST $500,000 AND OUTLAST THE FACTORY IT SHIPS TO.
128 cavities. 128 parts per cycle. tolerances held to ±0.005 mm across a million shots, through repeated heat, pressure and mechanical movement, without losing alignment.
the steel is hardened H13 or S136. cooling circuits run water through channels inside it, temperature controlled to within 1-2°C, cycle after cycle.
to make one of these you need CNC machining, EDM, grinding, polishing, heat treatment, metrology, mold flow simulation and hot-runner engineering, all coordinated, all precise.
china does roughly 68% of global plastic injection mold supply.
that's not a trade statistic. that's who decides how cheaply and quickly everything from car parts to medical devices gets made at scale, for every factory that uses their tooling.
these aren't the machines. they're what makes the machines possible.
any country serious about manufacturing has to answer one question first: can you make the mold?
MECHANICS JUST GOT A ROBOTIC COWORKER.
This humanoid robot is working inside an auto shop, moving around a vehicle and handling mechanical tasks like it actually belongs there.
No custom factory floor. No perfectly controlled demo room. Just a robot standing where a mechanic normally would.
That matters more than the flashy movements.
Cars are already full of repetitive jobs that require the same motions hundreds of times. A machine that can see the workspace, pick up tools, move around the vehicle, and repeat those tasks for hours could change the economics of an entire repair shop.
The mechanic of the future might not disappear.
He might just supervise 10 robots instead of turning the same wrench 200 times a day.
The weird part is how normal this is starting to look.
A $50,000 ROBOT COOKS DINNER AND THEN SITS WITH YOUR GRANDMOTHER
that's IRON running a full domestic shift. stove, pans, tea service, then pulling up a chair beside an elderly relative and staying there.
care work was supposed to be the last human job. too much patience, too much presence, too much of the thing you can't automate, which is showing up.
watch it plate a meal and pour tea and that line gets thin.
fifty thousand is roughly a year of part-time home care in most of the US. one year against a machine you own.
what a caregiver actually sells is years of trust assembled one visit at a time. she knows your mother stopped eating properly in march. she knows which stories get told twice and to let them.
this thing arrives with a spec sheet.
the interesting question isn't whether the arms work. it's who decides that presence and attendance are the same purchase.
https://t.co/QEbyuIBse3
THEY GAVE THE ROBOT A BADGE NUMBER.
030809, stencilled on a hi-vis vest, over a black humanoid with a police light bar on its back. shenzhen, walking a patrol route beside a swat unit.
the costume is the tell. you don't put a vest and a service number on a tool. you put them on a colleague.
and the walking genuinely is solved. flat pavement, managed crowd, squad pace, and the gait holds without a wobble. that problem is closed.
what nobody films is the robot deciding anything. every second of this is a machine going where it was told, next to people who are doing the actual policing.
the crowd with their phones up is the deployment. deterrence is an optics product, and optics ship years ahead of autonomy.
the first genuinely interesting clip will be the one where it has to make a call.
THEY GAVE THE ROBOT A BADGE NUMBER.
030809, stencilled on a hi-vis vest, over a black humanoid with a police light bar on its back. shenzhen, walking a patrol route beside a swat unit.
the costume is the tell. you don't put a vest and a service number on a tool. you put them on a colleague.
and the walking genuinely is solved. flat pavement, managed crowd, squad pace, and the gait holds without a wobble. that problem is closed.
what nobody films is the robot deciding anything. every second of this is a machine going where it was told, next to people who are doing the actual policing.
the crowd with their phones up is the deployment. deterrence is an optics product, and optics ship years ahead of autonomy.
the first genuinely interesting clip will be the one where it has to make a call.
A $50,000 HUMANOID ROBOT WALKED INTO A BATHROOM WITH A MOP. THE HARD PART ISN'T THE MOPPING
this is the task everyone assumed humanoids would take first, and it's the one they're worst at.
a wheeled floor scrubber has a fixed base, a known footprint, and one motion. it drives a floor plan, scrubs, docks, recharges. thousands of them are running tonight and nobody films them.
a humanoid has to do all of that while staying upright, then reach around a fixture, then adjust grip pressure on a mop handle it can't feel properly.
that's the actual gap. not intelligence. contact.
pushing a mop means constant force against an uneven surface, changing every second as the head catches on grout, a drain, a corner. a human wrist handles that without thinking. a robot arm has to sense it, decide, and correct, in real time, on every stroke.
which is why the mop is the interesting part of this clip and the walking isn't.
the price stops people, and it's the wrong thing to look at. what matters is how long it ran before someone stopped filming.
restrooms are the last thing automation gets, not the first. this is a robot attempting the hardest version of a job the boring machines already won.
EVERY SWING OF THAT HAMMER IS DOING MORE DAMAGE TO THE ROBOT THAN TO THE WALL
almost every humanoid on the market runs harmonic drives in its joints. high reduction, zero backlash, small enough to fit inside a shoulder. that's what makes the movement look precise.
it's also what makes impact catastrophic.
a strain wave gearbox works by flexing a thin metal cup. brilliant under smooth continuous load, terrible under shock. a hammer strike sends a spike back through the tool, through the wrist, into that flexing component in a fraction of a millisecond.
the joint that makes a robot look elegant is the joint that hates being hit.
the swing itself is the other problem.
a human doesn't push a sledgehammer, they whip it. energy loads from the legs through the hips into the torso and reaches the hands last, with tendons storing and releasing it like springs along the way.
rigid actuators can't do that. they apply torque directly, so matching the same impact takes far more motor power than the muscle version.
and every strike shoves the robot backwards, because the wall pushes back with exactly equal force. a person absorbs it through braced legs. a biped absorbs it through a balance controller nobody designed for a shockwave.
that's why every demolition machine that actually works sits on tracks and weighs half a tonne.
the recoil goes into the ground instead of into the gearbox.
https://t.co/QEbyuIBse3
NEARLY 500 MILLION FARMS ON EARTH ARE SMALLER THAN TWO HECTARES. NOT ONE OF THEM WAS EVER GOING TO GET A COMBINE
that's 84 to 85 percent of every farm in the world, and it's the market agricultural machinery skipped entirely.
the numbers are lopsided in a way most people have backwards. farms over 1,000 hectares are 0.1 percent of all holdings and operate more than half the world's farmland. the largest one percent work over 70 percent of it.
the 500 million small ones? about 9 to 12 percent of the land.
so the machines got built for the land, not for the farms. a combine is engineered for the field that has the acreage to justify it, and roughly six out of seven farms on the planet never did.
in eastern and southeast asia the average holding is 0.8 hectares. under two acres. no full-size harvester is ever driving into that.
and here is why it matters rather than just being a statistic. in low and lower-middle income countries, smallholders under two hectares produce nearly 60 percent of the food, according to FAO's 2025 assessment. globally they account for roughly 35 percent of food production.
for a thousand years the only tool small enough for a small field was a human body.
a bipedal robot is slow. it will never outpace a combine on open plains, and it doesn't have to. it fits somewhere a combine physically cannot go, which is most of the farms in existence.
whether this specific clip is a working deployment or a demo is a separate question, and worth asking every time.
but the gap it's aimed at is real, and it's the biggest one in agriculture that nobody has ever built for.
LOOK AT WHAT THIS ROBOT IS ACTUALLY DOING. IT'S PLASTERING A WALL, AND THAT'S HARDER THAN ANYTHING IN A BACKFLIP VIDEO
plastering is the exact task humanoids are worst at, which is why almost nobody films it.
the wall isn't flat. it's cracked, uneven, with old render still clinging in patches. the robot has to press hard enough to bond the material and soft enough not to gouge it, adjusting on every stroke based on what it feels through the tool.
that's contact-rich work. force control against a surface it can't fully see, changing constantly.
a backflip is a fixed trajectory. floor is flat, nothing pushes back unpredictably, and the whole motion is planned before it starts.
this isn't. every square metre of wall is a different problem.
which is why the interesting question isn't whether it can do one stroke. it's how long it ran before someone stopped recording.
the clips that matter in this field are the boring long ones nobody shares.
LOOK AT WHAT THIS ROBOT IS ACTUALLY DOING. IT'S PLASTERING A WALL, AND THAT'S HARDER THAN ANYTHING IN A BACKFLIP VIDEO
plastering is the exact task humanoids are worst at, which is why almost nobody films it.
the wall isn't flat. it's cracked, uneven, with old render still clinging in patches. the robot has to press hard enough to bond the material and soft enough not to gouge it, adjusting on every stroke based on what it feels through the tool.
that's contact-rich work. force control against a surface it can't fully see, changing constantly.
a backflip is a fixed trajectory. floor is flat, nothing pushes back unpredictably, and the whole motion is planned before it starts.
this isn't. every square metre of wall is a different problem.
which is why the interesting question isn't whether it can do one stroke. it's how long it ran before someone stopped recording.
the clips that matter in this field are the boring long ones nobody shares.
A $50,000 HUMANOID ROBOT WALKED INTO A BATHROOM WITH A MOP. THE HARD PART ISN'T THE MOPPING
this is the task everyone assumed humanoids would take first, and it's the one they're worst at.
a wheeled floor scrubber has a fixed base, a known footprint, and one motion. it drives a floor plan, scrubs, docks, recharges. thousands of them are running tonight and nobody films them.
a humanoid has to do all of that while staying upright, then reach around a fixture, then adjust grip pressure on a mop handle it can't feel properly.
that's the actual gap. not intelligence. contact.
pushing a mop means constant force against an uneven surface, changing every second as the head catches on grout, a drain, a corner. a human wrist handles that without thinking. a robot arm has to sense it, decide, and correct, in real time, on every stroke.
which is why the mop is the interesting part of this clip and the walking isn't.
the price stops people, and it's the wrong thing to look at. what matters is how long it ran before someone stopped filming.
restrooms are the last thing automation gets, not the first. this is a robot attempting the hardest version of a job the boring machines already won.
A $50,000 HUMANOID ROBOT WALKED INTO A BATHROOM WITH A MOP. THE HARD PART ISN'T THE MOPPING
this is the task everyone assumed humanoids would take first, and it's the one they're worst at.
a wheeled floor scrubber has a fixed base, a known footprint, and one motion. it drives a floor plan, scrubs, docks, recharges. thousands of them are running tonight and nobody films them.
a humanoid has to do all of that while staying upright, then reach around a fixture, then adjust grip pressure on a mop handle it can't feel properly.
that's the actual gap. not intelligence. contact.
pushing a mop means constant force against an uneven surface, changing every second as the head catches on grout, a drain, a corner. a human wrist handles that without thinking. a robot arm has to sense it, decide, and correct, in real time, on every stroke.
which is why the mop is the interesting part of this clip and the walking isn't.
the price stops people, and it's the wrong thing to look at. what matters is how long it ran before someone stopped filming.
restrooms are the last thing automation gets, not the first. this is a robot attempting the hardest version of a job the boring machines already won.
WATCH HOW SMOOTHLY THIS ROBOT MOVES. THAT SMOOTHNESS IS THE WHOLE ENGINEERING PROBLEM
what makes movement look robotic isn't speed or shape. it's abrupt changes in acceleration.
start a joint too suddenly, stop it too sharply, and the eye reads it as machine instantly. humans almost never do that. we ease into and out of every motion without thinking.
so making a robot look fluid means planning trajectories that minimise those jolts across every joint at once, while the balance controller is separately predicting where the centre of mass will be a fraction of a second ahead.
both of those run continuously, and they fight each other. smooth motion wants gentle corrections. staying upright sometimes needs a sharp one.
the dance is choreographed. the balance underneath it isn't.
that's the part actually worth watching.
https://t.co/QEbyuIBse3
A $50,000 HUMANOID ROBOT WALKED INTO A BATHROOM WITH A MOP. THE HARD PART ISN'T THE MOPPING
this is the task everyone assumed humanoids would take first, and it's the one they're worst at.
a wheeled floor scrubber has a fixed base, a known footprint, and one motion. it drives a floor plan, scrubs, docks, recharges. thousands of them are running tonight and nobody films them.
a humanoid has to do all of that while staying upright, then reach around a fixture, then adjust grip pressure on a mop handle it can't feel properly.
that's the actual gap. not intelligence. contact.
pushing a mop means constant force against an uneven surface, changing every second as the head catches on grout, a drain, a corner. a human wrist handles that without thinking. a robot arm has to sense it, decide, and correct, in real time, on every stroke.
which is why the mop is the interesting part of this clip and the walking isn't.
the price stops people, and it's the wrong thing to look at. what matters is how long it ran before someone stopped filming.
restrooms are the last thing automation gets, not the first. this is a robot attempting the hardest version of a job the boring machines already won.
https://t.co/QEbyuIBse3
Last Seen Users on Sotwe
Most Popular Users
Elon Musk 
@elonmusk
241.6M followers
Barack Obama
@barackobama
119M followers
Cristiano Ronaldo
@cristiano
114.2M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.2M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.4M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
89.9M followers
Taylor Swift
@taylorswift13
83.8M followers
Lady Gaga
@ladygaga
75.3M followers
Virat Kohli
@imvkohli
73.2M followers
Kim Kardashian
@kimkardashian
70.8M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.2M followers
Bill Gates
@billgates
65M followers
Selena Gomez
@selenagomez
63M followers
The Ellen Show
@theellenshow
62.3M followers
CNN
@cnn
61.8M followers
X
@x
60.7M followers