Forget lidar.
One single camera.
Runs in real time & is open source:
A streaming 3D model that reconstructs scenes live, at ~20 FPS, over long sequences.
End-to-end.
Optimization tricks, cleanup steps?
Nope.
And it beats both streaming and even some offline methods.
Perception is becoming software-first.
Closer to machines that see and understand the world as it unfolds.
Thanks for sharing, @YinghaoXu1
📍Models: https://t.co/gnSDy919eX
Project page: https://t.co/zgpkgBvcik
Code: https://t.co/Js0MzHE387
Paper: https://t.co/FrzMojMMZC
——
Weekly robotics and AI insights.
Subscribe free: https://t.co/9Nm01QUcw3
@IlirAliu_ One of the tedious projects I’ve done was on the depth analysis using data of a #LiDAR, and yes you are right as AI has done exceptionally well in this domain. #tesla as well (if I’m not mistaken), is a living proof of it.
A reporter tries to tell Tucker Carlson that Elon Musk is “undermining” Britain’s democracy.
He bursts out laughing, and it’s not about Elon at all. It’s about Britain.
REPORTER: “Can we take a moment to talk about Elon Musk? Because there are many in Britain who are concerned that he is trying to undermine Britain’s democracy.”
TUCKER: “Britain’s what?”
REPORTER: “Democracy.”
[Tucker looks directly at reporter, sees she is serious, then laughs]
TUCKER: “You guys don’t have a democracy! What are you talking about?!”
REPORTER: “We do, we do. We do, we do. We do have a democracy…”
TUCKER: “George Galloway would be the Prime Minister if you had a democracy.”
The engineers behind Claude Code just published a 5-PAGE PDF on loop engineering for LLM code.
They run an LLM through 5 loops: compile errors, static analysis, tests, until the code gets way better.
Every step is written out, so you can drop it into your own workflow immediately.
Read it, then check the article below. Worth every second.
This is the EXACT architecture OpenAI uses to build AI agents.
They just dropped a 34-page guide, I compressed it into one page.
6 stages, one loop, everything you actually need.
Read it, then go to the step by step guide on building LOOPS for your agents below.
Mark Zuckerberg said something so quietly devastating that even he does not seem to understand what he gave away.
Zuckerberg: “If the intelligence of a 10,000-person company is not greater than the intelligence of a single person, then what are we doing here.”
He asked it as a rhetorical question.
It stopped being rhetorical the moment he finished the sentence.
A company was never a mind.
It was a translation layer, built so one person’s vision could survive contact with a thousand strangers who would never fully understand it.
Every meeting, every manager, every layer between an idea and the person executing it was the cost of that translation.
We just called that cost the company, and mistook it for the value.
Meta proved it this year.
Thousands of roles cut.
Thousands more reassigned into the machine that no longer needs a translator.
Zuckerberg asked what ten thousand people are for, if they are not smarter than one.
There is a harder question underneath it.
A company was never about being smarter than anyone.
It was about reaching further than any one person’s hands could go alone.
AI does not make you smarter than ten thousand people.
It removes the only reason you ever needed ten thousand people.
That does not measure what you are worth.
It never did.
It only ever measured how far your own mind could reach before it needed other people to carry it further.
Reach used to cost a payroll.
Now it costs your attention.
The gate was never about intelligence.
It was about who got to multiply themselves.
For a hundred years, that gate opened for almost no one.
Zuckerberg: “Instead of having relatively few people be able to harness the power of a 10,000-person organization… I think in the future almost everyone is going to have that.”
He asked what ten thousand people are for, if they are not smarter than one.
You were never the ten thousand.
You were always the one.
Tim Cook, who told The Wall Street Journal that the jump in costs was unlike anything he had seen “in any area in over 40 years.”
Biggest price jump in anything I’ve ever seen too. https://t.co/aypJGgssnN
Analyse psychologique terrifiante de Fiodor Dostoïevski, tirée de Les Frères Karamazov :
« Ne te mens pas à toi-même. Celui qui se ment à lui-même et finit par croire à ses propres mensonges devient incapable de reconnaître la vérité. Il ne la voit plus ni en lui-même, ni autour de lui, et finit par perdre le respect de sa propre personne comme celui des autres.
Et lorsqu’il ne respecte plus personne, il cesse d’aimer. Alors, privé d’amour et cherchant seulement à se distraire, il s’abandonne aux passions et se laisse entraîner vers les plaisirs les plus vils, jusqu’à sombrer dans une forme de bestialité.
Et tout cela n’arrive que parce qu’il ment sans cesse : aux autres comme à lui-même. »
Deleting a file looks weightless.
No smoke. No sound.
No visible wound in the universe.
But physics says something stranger: when information is truly erased, the world must pay a small thermodynamic price.
This is Landauer’s principle.
In 1961, Rolf Landauer, working at IBM, asked a question that sounded almost too philosophical for engineering:
What is the physical cost of computation?
The answer was not hidden in the screen. It was hidden in the word erase.
A bit can be 0 or 1. If you reset it to 0, regardless of whether it began as 0 or 1, you have destroyed information about its past state. Two possible histories have been compressed into one final result.
That is not just a logical operation.
It is a physical act.
Landauer’s insight was that logical irreversibility becomes physical irreversibility. If a computation throws away information, the missing distinction does not simply vanish into nothing. The entropy has to go somewhere.
Usually, it goes into the environment as heat.
The famous bound is:
E ≥ kBT ln 2
Here, kB is Boltzmann’s constant, T is temperature, and ln 2 appears because one bit has two possible states.
At room temperature, this energy is incredibly tiny, around 3 × 10⁻²¹ joules per erased bit. Your laptop wastes vastly more energy than this for practical reasons: resistance, imperfect circuits, speed, error control, architecture.
But the principle is not about today’s inefficient machines.
It is about the basement floor of physics.
Landauer’s principle tells us that information is not an abstract ghost. A bit must live somewhere: in a voltage, a magnetic domain, a trapped particle, a quantum state, ink on paper, or neurons in a brain.
To change information, you must change a physical system.
To erase information, you must erase a physical distinction.
This also helped clarify Maxwell’s demon, the imaginary creature that seemed able to beat the second law of thermodynamics by sorting fast and slow molecules.
The demon’s trick depends on information. It measures molecules, remembers results, and uses that knowledge to extract work. But to keep operating, the demon must reset its memory.
That erasure has a cost.
The demon does not defeat thermodynamics.
It becomes part of thermodynamics.
This is why Landauer’s principle is so beautiful. It does not make physics more mystical. It makes information less mystical.
The lesson is not that thoughts create reality or that the universe is magically made of data.
The lesson is more disciplined:
Knowledge needs a body.
Memory needs a medium.
Forgetting leaves a trace.
In ordinary life, deleting something feels like making it disappear.
In physics, disappearance is never so innocent.
Even a lost bit has to go somewhere.
🇨🇳 | Dos niños de China han establecido un récord mundial en la resolución de un cubo de Rubik en pareja, completando el rompecabezas en solo 8.57 segundos.
This is why smart people rarely build businesses
Jensen Huang stood in front of a room of Stanford graduates and told them he hopes they suffer.
He wasn't being cruel. He was being precise.
His argument: people with very high expectations have very low resilience. And resilience, not intelligence, is what decides who actually makes it. A Stanford grad has spent their whole life as the smartest person in the room. They've rarely been tested by real failure. So when something finally breaks, they break with it.
Then he said the line every founder should sit with: "Greatness is not intelligence. Greatness comes from character. And character isn't formed out of smart people, it's formed out of people who suffered."
He would know. At nine, Huang was scrubbing toilets at a Kentucky boarding school his family hadn't realized was a reform school. As a teenager he bussed tables at Denny's. In 1993 he started NVIDIA in a Denny's booth, and nearly lost it more than once in the years that followed. The character was built decades before the valuation showed up.
This is why he uses the words "pain and suffering" inside NVIDIA with what he calls great glee. He isn't trying to shield his best people from the hard part. He's trying to give it to them on purpose.
Talent gets you into the room. The people who stay are the ones who were broken once and learned they could rebuild.
It is humbling to consider that if we harness just 1 millionth of the Sun’s power for AI, that will be much more than a million times the intelligence of all of humanity
This is actually insane. 🔥
NVIDIA built a single AI that replaces decades of animation technology.
15,000 frames per second. 2 milliseconds of delay. For robots and games.
It's called MotionBricks. One neural network trained on 350,000 motion clips. You tell it: go here, pick up this sword, do it in zombie style.
It generates everything else foot placement, balance, transitions, follow-through in real time.
No animation graphs. No hand-crafted transitions. No per-task training. One model does it all.
In the demo, a character navigates a space, picks up a sword, vaults over a bench, sits down, and switches between zombie, injured, and skipping styles. Every single frame generated live by the AI.
And here's the part that matters most.
This is now the core motion engine inside NVIDIA's GR00T the stack powering humanoid robots used in research globally.
Same AI that animates game characters is now teaching robots how to move their entire body.
vc: @NVIDIARobotics