@HarryStebbings Do not be fooled by this, Wix and every company in the space is going to zero in ~5 years.
This is not Salesforce, it's a glorified HTML wrapper leeching off low information customers. It's over and they all know it.
@MattLoszak The coolest part of this new tech-industrial revolution is non boring SCADA systems.
The first time I opened a valve from my laptop instead of a control room with no windows and windows 95' graphics I almost wept
The fact that we have multiple start-ups producing nuclear reactors is insane.
This whole industry went into hibernation for 40 years. Seeing it bloom again with such ferocity is inspiring.
Building a nuclear reactor and splitting atoms was the coolest thing I've been a part of.
If you’ve ever wondered what it’s like to turn on a reactor, here’s what stood out to me:
PART 1: THE APPROACH
First, you need a "spark" to start the "fire".
In this case, that is a neutron source (AmBe), which quietly outputs a constant rate of neutrons in the middle of the core.
Next, you add the fuel.
We'd calculated that we'd need 12.2 fuel assemblies to go critical. Assembling and then loading them all one-by-one was a slow, methodical process that took days.
Each fuel assembly has enough energy potential to power ~600 homes for 3 years with no refueling. And yet is light enough that a single human could lift it.
Once you install the neutron source and then a single fuel assembly, you're actually already splitting atoms:
The neutrons from the source trigger some uranium atoms in the fuel to split, which emits a few more neutrons, which go on to cause a few more splittings (fission).
But as long as we had less than 12 fuel assemblies inserted, we knew there was basically a ~0% chance of going critical.
Why? Two main reasons:
First, there wasn’t enough fuel loaded yet to have a self-sustaining chain reaction. You need at least a certain amount of fuel per unit volume to pull it off.
Second, because the control rods were still inserted. Control rods are made of material that absorbs neutrons, and at first, the number of neutrons being created is less than the number of neutrons being absorbed by the rods and the rest of reactor, or leaking.
Once we'd loaded 13 assemblies, we knew we had a chance to achieve criticality.
It was 10pm on Friday July 3rd, so timing was tight. If this approach didn’t lead to criticality, we would have had to wait until Monday July 6th, failing the President’s Executive Order.
We started to pull out the control rods, slowly, in increments.
We'd remove them by a few inches, then stop, and see what happened with the charts.
The graphs started to come to life.
The main chart that matters is the neutron counts per second. As the rods were removed, the neutron count increased, then would stabilize.
Why did it stabilize?
The neutron source emits a constant rate of neutrons. Each time we withdrew the control rods, fewer neutrons were absorbed, allowing every neutron from the source to trigger a larger cascade of fissions.
But as long as the number of neutrons being released was less than the number being absorbed (sub-critical), the graph would always plateau at some new, ever higher level of counts per second.
So we kept going, searching for signs of criticality.
The neutron counts per second reached 100, then 400, 800, 1200, 2000, and kept plateauing all the way up to 20,000 counts per second.
For a moment, I was afraid our calculations had been wrong, and we'd actually needed 14 fuel assemblies to go critical. This would have pushed our criticality date to AFTER July 4th…
But then...
The ability to kick off a factory while the product and production line is still undefined is a unique Tesla trait. This is an underappreciated superpower.
When we broke ground on the Berlin/Austin Cell factories the dry electrode process was still years from stability. We invested hundreds of millions building them anyways.
Hundreds of design changes, retrofits, tool swaps later... Somehow it works. It's so kind of magic I've only ever seen at Tesla.
@JeanClaudeFox2@actsmaniac My girlfriend paid like €3k. I think it depends on how subsidized the rent was.
As an expat paying 2X local rent it didn't impact me lol
@JohnSpence33693@zslayback Americans drive ~3 trillion miles on the highway system each year, that's 15,000 round trips to the sun and back.
Beyond that, it's estimated to be responsible for ~5% of GDP or about $1.5 trillion dollars a year in economic activity.
It's not some abandoned cold war project
@doyendon@aelluswamy@Tesla_AI I remember hand labeling FSD data ~8 years ago. They would cater really nice food at Deer Creek and Karpathy would hang out on weekends.
It seemed so ambitious back then, it's incredible to see how far it's developed.
@paleonormie Not to mention local assessment fees are killing the market for small multifamily projects
5-10 unit complexs were built everywhere in the 60-70's. Today the economics make it almost impossible.
Now you need giant units or a giant volume of units for the math to shake out
@IterIntellectus As a consumer, I'm a huge fan of the insane levels of competition in the AI space
As an executive/board member, I would be in a stress induced coma trying to juggle user retention and costs with an IPO imminent.
At a certain scale, labor unions become deeply malevolent and anti progress.
IG Metall is willing to kill Volkswagen rather than accept some plant closures the company needs to survive.
This is a cartel actively destroying German industry, there is no other interpretation.
@UrbanCourtyard My courtyard in Berlin had one shoddily made table built out of a tombstone and some scrap lumber.
3 old guys sat there all day every day chain-smoking and drinking Sterni
Less than 100 miles away, about 50 tons of steel and copper are spinning at exactly 3,600 RPM. A giant rotor slews magnetic fields through copper windings and, per Faraday and Maxwell, induces a synchronized 60 Hz sine wave across an entire grid.
That electromagnetic wave propagates through grain oriented silicon steel, laminated transformer cores, kilometers of aluminum conductor steel reinforced cable, SF₆ insulated switchgear, substations switching hundreds of kilovolts, distribution transformers, oxygen free copper wiring, silicon MOSFETs, ferrites, multilayer ceramic capacitors, and the USB C cable plugged into my phone.
There, power feeds a capacitive touchscreen. A transparent matrix of indium tin oxide is deposited on glass. The controller continuously scans the grid. My finger changes the mutual capacitance by a fraction of a picofarad, perturbing the local electric field just enough for dedicated analog front ends to detect before software reconstructs the touch.
Iron ore. Bauxite. Silica sand. Copper ore. Lithium brines. Rare earths.
Refined. Alloyed. Crystallized. Doped. Implanted. Deposited. Etched. Polished.
We purified sand into nearly perfect single crystal silicon, grew boules, sliced wafers, placed dopant atoms with nanometer precision, fabricated chips containing tens of billions of transistors, synchronized them with clocks billions of times per second, and engineered them to execute matrix multiplications at a scale no human could comprehend.
Somewhere along the way, matmuls started to look like us.
That is an astonishing thought.
The towers of abstraction are staggering. No one person built this. No one person fully understands it. Yet every morning millions of people wake up tired, answer emails, argue over specifications, chase margins, fix defects, make payroll, and try to cover their mortgages. Somehow the whole machine keeps humming.
A planetary Swiss watch assembled from physics, markets, bureaucracy, ambition, error, and human necessity.
And after all of that, I touch a piece of conductive glass and step into the Roman forum of modern thought, where millions of minds test ideas against one another in real time.
Clockwork humanity.
It staggers the mind.