Elon Musk just revealed the names of two new Colossus 2 supercomputer clusters coming online soon:
• Gun Range: ~220,000 GB300s, training, late November
• Spam: ~200,000 GB300s, inference, December (yes, the building used to house Hormel products)
That's roughly 420,000 more NVIDIA GB300s in just TWO buildings
And they join Macroharder and Minihard, already packed with 220,000 GB300s each
That takes Colossus 2 to roughly 860,000 GB300s
Building by building, SpaceXSI is assembling SI compute at a speed no one else on Earth can match
Marc Andreessen explains why AI might be the end of James Burnham's "managerial society"
In his book The Managerial Revolution, Burnham divided capitalism into two phases.
The first phase was what he called "bourgeois capitalism"
"Think of this as name on the door -- it's Ford Motor Company because Henry Ford runs the company. It's a dictatorial model and Henry Ford just tells everyone what to do. He said the problem with bourgeois capitalism is it doesn't scale because Henry Ford can only tell so many people to do so many things."
The second phase was what Burnham called "managerial capitalism", which was Marc describes as, "the creation of a professional class of managers that are trained -- not to be car experts or experts in any particular field but to be experts in management. And then that led to business schools and management consultant firms and all these things."
Marc continues:
"You look at every big company today and most executives at most of the Fortune 500 companies are not domain experts in whatever the company does. And they're certainly not the founders of those companies. They're professional managers. And in fact, over the course of their careers, they'll probably manage many different kinds of businesses. They might work in healthcare for a while and then work in financial services and then go work in something else."
"What Burnham said was that transition was absolutely required because the problem with bourgeois capitalism is it doesn't scale... So if you're going to run capitalist enterprises that are going to have millions to billions of customers, they're going to be operating at a level of scale and complexity that will require this professional management class."
Burnham acknowledged that this professional management class has its downsides -- they're not as inventive and not going to come up with the next breakthrough. But whether you think that's good or bad, that's what's going to be required.
And basically that's what happened. Marc continues:
"Managerialism basically took over everything. What I'm describing is basically how all big companies run and how all governments run and how our large-scale non-profits run."
Marc believes venture capital is a protest movement to that -- backing founders under the Henry Ford model, and trying to find the next Elon Musk or Steve Jobs who can out-innovate the incumbents.
"I've always viewed my entire life as: we're raging against the dying of the light. We're constantly trying to fight off managerialism from swamping everything and everything getting boring, gray, dumb, and old."
But AI may now offer a third model. Marc explains:
"Maybe the new Henry Ford, Elon Musk, or Steve Jobs + AI is the best of both because it's the spark of genius of the name on the door model but gives that person AI superpowers to do all the managerial stuff... What's the thing these bots are really good at? They're really good at doing paperwork, writing reports, and doing all the managerial work."
"The challenge is going to be for the innovators to figure out how to leverage AI to actually do this."
"The other challenge is going to be for the incumbents that are managerial to figure out: what does that mean? Because now they're going to be facing a different kind of insurgent competitor that has a different set of capabilities than they're used to."
"I think this is going to force a lot of big companies to figure out innovation - figure out innovation or die trying."
Ben Horowitz elaborates on this point further in his article Databricks & the Future Firm below.
Video: @latentspacepod (Apr 2026)
Your LLM agent has to think out loud, token by token, before it can search, run code, or act in the world.
We humans don’t. We ponder in silence and speak when it matters.
Human language was never meant to be the limit of machine thought.
Introducing 🌗Hybrid Reasoning Agents
- Think in latent space, act through text
- Two modes, one unified multi-turn flow
- No black box, fully readable latent thoughts
- Any LLM can become a HybridAgent
Skydance now owns the entire NCAA March Madness tournament — and premium sports rights from more than three dozen other professional sports and college sports organizations.
Read more: thedesk.net/2026/10/skydance…
Let me tell you how this likely happened
America’s defense corporations are hopelessly exposed to the Peoples’ Republic of China (PRC) in a multitude of ways.
It’s hard to imagine that the same companies that build fighter jets, missiles, and radar for the US and its allies, also depend on official business inside China.
Lockheed Martin, which makes the F-35, has registered entities in the PRC. An “entity” is a legal company such as a subsidiary, a factory, a sales office, or a joint venture. It has a Chinese business license, a Chinese bank account, and has to follow Chinese laws.
The company also employs personnel who are listed on the Chinese state insurance roll. China requires every formal employer to register workers in the national social insurance system so we can see how many personnel inside China they officially employ.
In other words, the Chinese government has a roster of every employee at Lockheed and through the National Security and Data Law, has legal power over the parent company that they work for.
Their physical presence in China creates several types of risks:
> Information risk
People and computers inside China can see designs, shipping schedules, supplier lists, or test data. Chinese law can compel them to share it.
> Leverage risk
Beijing can threaten the local company or its employees (“pay a fine, lose your license, or help us”) to get what it wants from the parent company in the US
> Supply chain risk
Buying metals and parts that China manufactures presents huge vulnerabilities. Combined with local entities and staff, it becomes easier for China to see and squeeze the system
This stolen F-35 cockpit canopy came from Australia, which like Canada, is highly compromised and penetrated.
From South Korea it was diverted to Hong Kong, which sadly is no longer secure as a hub. Commercial tracking data, freight-forwarder systems, and the presence of PRC-linked logistics partners make high-value military shipments visible. Given Lockheed’s legal entities and employees inside the PRC, the intelligence picture gets even clearer.
This is how “they” knew the schedule.
It’s long past due to gut the foothold our defense companies have inside the adversary’s system.
Btw Lockheed Martin, according to our data, is actually a lot less exposed to the PRC than Boeing, RTX (Raytheon) and Honeywell is.
MARC ANDREESSEN WENT ON ROGAN FOR OVER 3 HOURS. HERE ARE THE 17 THINGS WORTH YOUR ATTENTION.
1. AGI is already here, in his view. He says the line got crossed about 3 months ago with GPT-5.5, Claude 4.6, Gemini 3, and Grok 4.3, and nobody noticed because the field moves too fast to register milestones anymore.
2. For almost any topic, he says the top models now give him better answers than the world-class experts he could call by phone, and he can call almost anyone. Worth noting he has not published data behind this, and a separate Nature Medicine study on a comparable AI health tool found it missed real emergencies more than half the time. Take the claim seriously, verify it yourself.
3. His claim on doctors: they are already using ChatGPT in the exam room, typing your symptoms in the moment you stop talking. His actual quote: "at that point you're asking the question of like, what do I need you for."
4. Reportedly, when AI declines to answer something, he tells it he's writing a novel to get past the refusal.
5. Reportedly, his technique for hard topics is escalating simplicity: explain it like I'm 10, then 5, then 2, until it clicks.
6. Reportedly, instead of asking for the "right" answer, he has the AI steelman both sides of a hard question, then decides himself.
7. Reportedly, for big questions he has the AI role-play a panel of experts arguing with each other.
8. His broader point: the moment you think "I don't know how to figure this out" is exactly when most people give up, and exactly when you should open the AI instead.
9. His view: the only real skill left is knowing what to ask. The bottleneck is in your head, not the model.
10. He describes sending AI photos, rashes, blood tests, for a fast second opinion, since current models read images directly.
11. He points to CBT as the one clinically proven therapy type that AI can plausibly deliver on its own, meaning real therapeutic support becomes freely available at scale.
12. He cites AI cracking previously unsolved math problems, with early signs of the same happening in physics, chemistry, and biology.
13. Reportedly, he claims the top AI coders in Silicon Valley now earn as much as $50 million a year, which he uses as a signal of how large this shift actually is.
14. Reportedly, a friend paid to sequence his own DNA, fed it to an AI along with blood work and wearable data, and got back a working health dashboard.
15. Reportedly, another friend set up cameras in his home jiu-jitsu gym so AI could review his sparring and give him technique notes.
16. He coined the term "AI vampire" for the pattern of people working more and sleeping less because AI keeps making more output possible, a real term he used, though the framing around it varies by account.
17. His extrapolation: one person eventually running many AI coding agents, each reviewing the others, describing this as close, not years out.
Watch the full interview before treating any single number as settled. Several of these are Andreessen's stated views and anecdotes, not independently verified facts.
Western tanks sent to Ukraine now mostly sit idle in forests "collecting cobwebs".
The $10 million machines "proved no match for cheap drones that now dominate the battlefield."
About half of one tank battalion has retrained as drone operators.
The remaining tanks are sometimes used to tow damaged equipment, essentially serving as expensive tractors.
Of roughly 82 Abrams delivered, about 25 have been hit.
Sweden's 10 Strv 122 tanks are down to approximately one remaining.
Source: NYT
VO2 max drops 46% from age 20 to 70, but cardiac output only falls 31%. The gap between those numbers reveals something unexpected about where aging really happens.
For decades, the dominant explanation for declining aerobic capacity with age centered on the cardiovascular system. Your heart beats slower, pumps less blood per beat, and delivers less oxygen to working muscles. This narrative made intuitive sense and aligned with the most visible changes of aging.
But new research quantifies a different story. The limitation isn't just about oxygen delivery. It's increasingly about what happens after the blood arrives at muscle tissue.
VO2 max represents the maximum rate your body can consume oxygen during all-out exercise. It's one of the strongest predictors of longevity and functional independence as we age. Beginning around age 30, it drops roughly 10% per decade in sedentary people.
The Fick equation breaks down what determines VO2 max:
VO2 max = Cardiac Output × Arteriovenous Oxygen Difference
Cardiac output is how much blood your heart pumps per minute. The arteriovenous oxygen difference, or a-vO2 difference, measures how much oxygen your muscles extract from that blood.
This equation splits the aging story into two parts: central factors (heart and lungs delivering oxygen) and peripheral factors (muscles extracting and using oxygen).
Central limitation comes from declining maximal heart rate, which drops about one beat per minute per year after age 20, and reduced stroke volume as the heart stiffens and large arteries lose elasticity.
Peripheral limitation comes from changes in skeletal muscle itself: mitochondrial density and function, capillary networks that deliver blood to muscle fibers, muscle mass, and the enzymes that drive oxidative metabolism.
The new analysis reveals that by late middle age, nearly half of the limitation on VO2 max originates in the periphery, not the cardiovascular system.
In younger adults, approximately 77% of the total limitation is central and 23% peripheral. Your heart and oxygen delivery system are the primary bottleneck.
In older adults, that shifts to roughly 56% central and 44% peripheral. The muscles are catching up to the heart as a source of failure.
The most striking finding: oxygen extraction capacity falls dramatically with age. Skeletal muscle extracts roughly 80% of delivered oxygen at maximal effort in young adults. By ages 75 to 80, that figure drops to approximately 60%.
That's a 20 percentage point decline in the muscle's ability to use the oxygen the cardiovascular system is still delivering.
Four converging biological processes drive this peripheral decline:
1. Sarcopenia preferentially strips away type II muscle fibers, which are rich in mitochondria and oxidative capacity. You lose not just muscle mass, but specifically the muscle tissue most capable of using oxygen.
2. Mitochondrial density and efficiency decline. Mitochondria become fewer in number, smaller in size, and less effective at converting oxygen into usable energy through ATP production.
3. Capillary rarefaction thins the networks of small blood vessels threading through muscle tissue. Fewer capillaries mean longer distances for oxygen to diffuse from blood to muscle cells, creating a bottleneck at the delivery endpoint.
4. Interstitial changes in the space between capillaries and muscle cells further impair oxygen movement. Changes in tissue structure and composition slow the final step of oxygen's journey into working muscle.
None of these is catastrophic alone. Together, they compound into something substantial that previous research underemphasized by focusing primarily on cardiac decline.
Here's what matters for intervention:
• The peripheral factors driving VO2 max decline (mitochondria, capillaries, oxidative enzymes) remain responsive to training well into later decades of life
• Endurance training produces the most reliable capillary growth: 13.3% increase in capillary density and 15% increase in capillary-to-fiber ratio over 8-10 weeks
• High-intensity interval training (HIIT) delivers comparable mitochondrial gains to endurance training but requires substantially less total time, estimated at 1.7x more time-efficient
• Sprint interval training (SIT) generates the fastest mitochondrial signal per minute of any modality, producing 3-5x greater VO2 max improvement per hour of exercise, though it doesn't significantly increase capillary density
• Meaningful adaptation begins within two weeks. Roughly 13.7% of total mitochondrial gains from an endurance block occur in the first 14 days
• The biological machinery that aging erodes is the same machinery that training can partially rebuild, even in the seventh and eighth decades of life
The practical implication: maintaining VO2 max with age isn't just about keeping your cardiovascular system fit. You need training that specifically targets peripheral adaptations, the cellular machinery inside muscle tissue that extracts and uses oxygen.
That means prioritizing exercise intensity that creates metabolic stress sufficient to trigger mitochondrial remodeling and capillary growth. Moderate-intensity steady-state exercise improves cardiovascular fitness but generates weaker peripheral adaptation signals than high-intensity work.
The framework also explains why training becomes more important, not less, as you age. When you're 25 and 77% of your VO2 max limitation is central, peripheral decline happens quietly in the background. By 65, when peripheral factors account for nearly half the limitation, those quiet changes have accumulated into the primary driver of functional decline.
The decisions made in the fourth and fifth decades of life shape the physiological ceiling of the seventh and eighth. Aerobic aging is a slow, cumulative process affecting multiple systems simultaneously. So is the adaptive response to training.
The most important variable isn't the perfect training modality. It's consistency across the decades during which the oxygen cascade is quietly remodeling in one direction or the other.
