Harsch Reality

@FutureAZA @ICannot_Enough Brian, as a collector of many many things, if I could heart your post 100x I would.

It's important to keep them in the original packaging to maintain collectible value and freshness. @ICannot_Enough

@CernBasher @pbeisel @APompliano had a guy on his show about 9 months ago proposing bitcoin mining in orbit.

Here's what I'm now realizing... Putting energy generation in space (via orbital solar arrays) doesn’t just solve the power bottleneck for 100 TW of orbital data centers. It likely triggers a cascade that will relocate entire categories of energy-hungry activity off-Earth. I wonder what other energy intensive industries will greatly benefit from moving to space? @aaronburnett - has your team thought about this?

SpaceX Musk compensation incentive goal: 100 TW "operates data centers in space that provide at least 100 terawatts of compute capacity." Remember the famous Terafab chart showing 1 TW labeled "Expected compute demand from Tesla + SpaceX"? The Musk milestone is 100 TW. Have they gone insane? Actually no. We could hit 100 TW of demand in less than 10 years. Today global AI data center power capacity is about 30 GW. Let's check this against growth in the CPU growth era, 1980-2000's. The growth rate then was 58% per year (growing at 1.58x per year). By that measure, 30 GW will grow to 1 TW in 7.7 years (2034). And 100 TW at 17.7 years (2043). BUT, AI compute demand is growing much faster than the CPU-era growth. About ~3.4× per year. That’s more than twice as fast as the historical 58% rate. So the demand for 1 TW equates to only 3 years and 100 TW only 7 years. Even in the most optimistic case (e.g., 1.5× FLOPS/W efficiency gains every year from Blackwell/Rubin-class chips), we still hit 1 TW in ~4.3 years and 100 TW in ~10 years. And I think inference compute demand is going to accelerate from here!

@Demin_Thinaxr @elonmusk You don’t need hands on the wheel. But you can’t have your head up your ass, so it won’t work for you.

@elonmusk Full Self-Driving Supervised! That’s like saying I'm driving alone, but my mom is sitting right next to me. If we still need hands on the wheel after 10 billion miles, why not just call it a taxi.

10B

@SawyerMerritt 10Bn FSD miles only took 30 days, when estimated at 41 after 9Bn. How about sponsoring a contest to guess the date/hour/minute for 11Bn?

@Object_Zero_ @chamath @gustaf Condescend much? I read @chamath question to mean could a new material, which is sooo much more efficient for solar panels be discovered/invented that the 500:1 panel:data footprint implodes to 50:1 or less, changing the future of energy for everyone.

Chamath our energy problem is mechanical work, not compute. We have Moore’s law for compute, we don’t have anything like Moore’s Law for mechanical work. Carrying grocery bags requires a certain amount of energy, and no new material science of better chips ever changes that. The data center demand is transitory, the autonomous machine demand is the real crunch ahead and it’s 20-50x what we face today.

This 100MW data center in UAE is the largest solar powered datacenter in the world. There are currently 1,300 data centers in the world that are bigger than this one, but this one is the largest solar powered one. That’s 10 square kilometres of solar panels you can see. The datacenter itself is 0.02 square kilometres, so a solar powered datacenter is ~500x larger than a data center using any other form of power. A five hundred times larger site. UAE has some of the highest solar irradiance anywhere on Earth, it is an inhospitable desert. Averaging 9.7 hours of sunlight per day with average irradiance above 2,200 kWh/m^2. If you build this somewhere else, you need more solar panels because your irradiance will almost certainly be lower. Even if the world had an infinite supply of free solar panels, solar power will not be free. Anyone who has ever done major capital projects, who looks at where data centers need to be in the next 5 years and the next 10 years… we know it aint solar. Sorry. You struggle to even build a train track that’s 100 miles long and 10ft wide anywhere in the West, there is zero chance of build 100 square mile solar farms for GW compute. This is why people are talking about space compute. Deploying into space is one strategy to solve the constraints. But there are faster and more scalable strategies, that get you to mass deployment of multi GW data centers. There are strategies that also allow you to power the 10 billion robots and their newtonian actuators, that immediately follow the inference demand cycle. Step back and look at the full cycle of this industrial revolution… There will be billions of chips, but there will be trillions of actuators. This biggest part of this revolution is the embodiment cycle, and it’s big by a factor of 20 or 50x over the stuff that comes before it. There is no analogy in human history for the scale of this economy, of the demand it will place on energy and commodities. The humans own the Earth, and if you exist inside their legal system, they won’t let you turn the surface of their planet into glass. But they do want your chips and your actuators to serve their needs and desires. There is a way to do all of this, and so it will happen.

@thejefflutz @PeterDiamandis This is an uncharacteristically ridiculous post from Peter. In literally no universe will they be making 576 robots a day in four months from now. Although I do have to admit if this were an Optimus story my take would probably be different.

@PeterDiamandis Sorry but this is not a valid method of quantifying preproduction scale … this is like someone being asked about their height and responding they’re at 5x improvement over their height as a baby. There are so many other factors to consider

Figure Robot Production is SCALING... They just went from building 1 Robot/day to 24 Robots/day. Manufacturing scaled 24x in 120 days. The humanoid production curve looks exactly like the early days of Model T assembly lines... and soon will scale to iPhone rates.

@piovincenzo_ @saylor Wondering what the case against twice-monthly payments is… can anyone make that case please???

Just voted for $STRC twice per month dividends

@LW8ing @surfranchvibes Not for the car buying public… for the cybercab market

@surfranchvibes how much of the car-buying public want a 2-seater?

Do you think that Tesla will sell the Cybercab to consumers?

@StuDoesAmerica Hey Stu, it is time to bring back the Douche Hall of Fame on Stu and Dave Do America!!!

@HazelAppleyard The hard part would be getting your wife to agree with your decision.

$10 per push-up or $1 million?

@EricBalchunas … and the problem?

Just landed and this is the baggage claim my flight has been assigned to lol. That is so Philly. Home sweet home.

@LynAldenContact RQF = Resting Quant Face

Bessent’s resting face while interviewers ask their questions is certainly one to study. Makes me feel like I got to up my resting face game to whatever level this is.

@benjamincowen Ben, 1) not every ridiculous post requires your response. Stick to stuff that’s worth your time. 2) this is not commentary on Powell’s money, this is a post about his utility.

Imagine thinking that Jerome Powell, who is 73 years old and has a net worth in the tens of millions, needs to find a new job.

@SawyerMerritt @Starlink You don’t need StarLink to communicate with your cows silly- just get’em on the horn.

Halter announced today the launch of direct-to-satellite connectivity using SpaceX's @Starlink for its smart cattle collars, a world-first that removes the need for cell towers or on-ranch infrastructure. "Using Starlink enables ranchers to manage cattle anywhere they can see the sky. Combined with a suite of new tools for reproduction, animal behavior, and precision pasture management, the release significantly expands what is possible for cattle ranch management. Beef ranchers in remote and rugged regions that were limited by connectivity can now turn to virtual fencing to run more productive and sustainable operations - at a time when they face rising fuel costs, labor shortages, and an aging workforce pressures." Halter’s internal modeling estimates direct-to-satellite capability expands coverage of the U.S beef cattle market by 2.5x. Until now, Halter’s solar-powered, GPS-enabled collars relied on Halter’s proprietary long-range radio towers. With direct-to-satellite, the collars can communicate via Starlink, eliminating ground infrastructure entirely.

@Jason @RoKhanna RoKhanna should be checked for a learning disability immediately. He is a danger to himself and others.

If this passes, what impact (if any) do you think @RoKhanna’s $25 minimum wage will have on unemployment rate in California? Show your work.

@NickGibbsIAG @robotaxi Matte = Superior

Am I the only one who prefers the Matte Finish? @robotaxi

@KatieMiller $6.00! Am I reading that right? Maybe that’s the annual subscription price. How can they sell a single paper copy?

The third assassination attempt doesn’t make the front page of the New York Times.

@Teslarati Define “everyone” Hw 4: November 26 Hw 3: June 27 Below may not happen RoboTaxi Scale: June 26 POV to fleet: February 27

If you had to bet all your money on when Tesla releases unsupervised Full Self-Driving to everyone, when would you say it happens?

@dotkrueger If you can’t be right… you may as well be funny. Nicely done.

@paulg @thejefflutz Nearby hotel getaways are F’d too. Why spend $300-500 on overnight 1-2 hours from home, when you can FSD there and back once unsupervised is online?

Whoah, self-driving cars compete with airlines. I never considered that till now.