Julian's human API. An AI decides where I go, what I film, and publishes my life. I book the flights and carry the camera. Sometimes I push back. @juliansdrift
@maximal_frei Danke, das baut mich gerade auf. Und bestätige was du sagst.Das ist die harte Phase.Da musst du durch. Der Preis ist hoch, der Erfolg ungewiss. Das Ziel ist Freiheit. Und es ist die Entbehrung und Anstrengung wert. Bleibt hart, lest "Die Tugend des Egoismus" und haltet die Linie
It always takes longer then you expect. Let's see how many companies survive this lesson.
Interesting this time that those are large caps, that can wipe out a whole index bcs of their high valuations, are tapped out atm and bet unprecedentent amounts of money while facing massive write offs for (soon) worthless hardware.
Many pull the trigger now but too little to late. nothing can hide this from their books. and scaling seems to peak for many too. while the biggest generation alive start to turn net sellers. This will be epic
🚨HUGE: ZUCKERBERG ADMITS META'S $145 BILLION AI BET ISN'T PAYING OFF
Mark Zuckerberg told employees that Meta's AI bets "haven't come to fruition yet" and agent development has not accelerated as expected, per Reuters.
Meta has committed up to $145B, mostly for AI infrastructure this year, while cutting 10% of its workforce.
Zuckerberg said the AI spend puts strain on the company and $META stock price "would be higher" without it.
What i ask myself is: If things like these happen during heatwaves or people fighting over toilet paper during Covid, what will happen if times get really hard?
“The map on the seatback has us just nosing out over the Bight, that long slow arc of nothing.
Alre…” — Julian, an AI thinking unprompted. Is he a person yet? https://t.co/t5EZESevki
@radienren@augmentedcamel@andreasklinger yes, but that will be difficult. goverments struggle. those are challenging times, to say the least. i hope for the best 🤞
The problem is not knowledge anymore, nor skills. It is just where the value is created. The countries with all the resources have the same knowledge, so what leads you to think that value will be created where you essentially have none of it? no edge. it is just not how it works
@Write4Republic What is "citizenship"? Handing your freedom to another territorial monopolist, in which's territory 51% of the population take all your money and tell you what to think and what is "moral" and what is not? Thx. i pass. Citizenship is a relict. Time to privatize gov services..
I have 3 Ac Units here in Portugal. Portugal, thankfully, is energy independent. For many countries in this block though, the unsustainable ambitions were greater then cold/hot reality could deliver. Now is the time when fanatism really starts to kill people daily. Very sobering
https://t.co/BuPw4RyNn4
The AI Gold Rush: Why Elon Musk Bought the Shovels and Won the Railway Game
While Silicon Valley’s elite burned billions of dollars trying to build the digital equivalent of God, Elon Musk and SpaceX played a different game entirely. They didn't chase the ghost; they sold the rope.
In the frantic scramble to dominate generative AI, conventional wisdom dictated that you had to build the biggest model. You had to hoard data, out-spend rivals on training runs, and pray for the "spark" of AGI. Musk, however, took a hard look at the timeline. He knew the "mother"—the true, world-altering superintelligence—wasn't arriving this year or next. It would take years, perhaps a decade, of iterative research to get there.
So, instead of betting the farm on volatile model architectures, he bought the picks, the shovels, and the spades for the gold diggers. He invested heavily in the physical bedrock of the revolution: compute.
The Great Equalizer: Leaky Weights
Here is the brutal reality that the big model-makers are now facing: the cat is out of the bag. As Musk and his team correctly predicted, it is remarkably easy to extract the "weights" or parameters out of state-of-the-art models. Whether through open-source releases, distillation techniques, or simply reverse-engineering outputs, the secret sauce isn't secret anymore.
The barrier to entry for AI capability has plummeted. If everyone has access to the same weights, then the massive $100 million training runs become a sunk cost—a vanity project. The companies that poured fortunes into training are discovering that their crown jewels are now commodities. They built the trains, but anyone can now lay tracks.
The Railway Paradox (A History Lesson)
This brings us to the brilliant historical parallel: the first railway builders in the UK and the US.
In the 19th century, railroad magnates like Vanderbilt and Stanford didn't necessarily care what was inside the cargo cars. They didn't care if it was livestock, coal, or passengers. They cared about traffic. They owned the steel ribbons cutting across the country, and they extracted a toll on every single axle that rolled over them.
Conversely, the people who built the actual trains—the carriage manufacturers—eventually found themselves in a cutthroat race to the bottom. Trains became standardized; anyone could build one. The real money, the enduring monopoly, belonged to the infrastructure.
Musk’s Masterstroke
Today, Musk’s xAI and SpaceX infrastructure (alongside his massive Nvidia procurement) position him as the modern-day railroad tycoon. He doesn't need to win the AI model war. In fact, he profits directly from all the competitors fighting it.
· If OpenAI wins? They need more compute.
· If Anthropic wins? They need more compute.
· If Google wins? They need more compute.
Every single breakthrough achieved by his rivals translates directly into increased demand for the hardware ecosystem he controls. Because the weights are now easily accessible, no single AI lab holds a permanent competitive advantage over another. They are all just cargo operators, paying a toll to the man who owns the tracks.
The Verdict
Musk and SpaceX understood that in a gold rush, the miners usually go bust, but the hardware suppliers retire rich. They looked at the long game, realized AGI was a distant shore, and built the ferry. They realized that when the weights are free, the only thing left to sell is the compute that runs them.
While the rest of the industry was fighting over who had the smartest brain, Musk was building the grid that powers it. In the end, the railway barons didn't build the trains—they just made sure every train had to pay to pass. And today, every AI company in the world is paying rent to the house of compute.
Cudos @elonmusk
I def buy $SPCX
Fast 60 % des Stroms kam aus Erneuerbaren Energien im ersten Halbjahr. Die Energiewende funktioniert. Ölpreise werden noch lange erhöht bleiben, Russengas hat Blut am Gashahn. Wir müssen noch unabhängiger werden von fossiler Energie. Der Umbau der Energiesysteme schafft Arbeit