Saturday, 26 September 2026

#thedarkmodernity New York, for real

 So, he is very prominent. With all given respect do I wonder how close he is to the root cause of New Yorks problem, which is Manhattan, not ever the Bronx.
 
This is why:
 
Lets talk Manhattan. Tell me how much floor space rose, how much more people moved in and come to work and how public transport and the traffic system grew in parallel since the first Subway was build.
 
The opening of the first New York City subway in 1904 was a pivotal event that reshaped Manhattan's physical form, population dynamics, and commuting patterns. It triggered a cycle of vertical growth in the business districts and horizontal expansion into the outer boroughs, fundamentally altering how the city functioned.

### 🏙️ Floor Space: The Rise of the Vertical City

Before the subway, Lower Manhattan was already a dense office district. The new rapid transit lines provided the critical accessibility needed to concentrate hundreds of thousands of workers, which in turn made the construction of skyscrapers not just feasible but highly profitable. This led to an explosion in built floor area.

The data below illustrates this trend. The **Floor Area Ratio (FAR)**—the ratio of a building's total floor area to the size of the plot of land it is built on—more than doubled from the 19th century to the end of the 20th century in Manhattan.

Here is a breakdown of Manhattan's major office districts and their scale today:

| District | Approximate Office Space |
| :--- | :--- |
| **Midtown Manhattan** | ~200 million square feet |
| **Lower Manhattan** | ~107.8 million square feet (as of 2001) |

This vast amount of space is concentrated in a relatively small area. By 2013, the total built-up area in Manhattan was about **4,466 hectares** (roughly 17.2 square miles), which supported this immense floor space.

### 👥 Population: A Story of Decongestion and Resurgence

The subway's most dramatic demographic impact was on Manhattan's population density. In 1910, Manhattan reached its historical peak population of **2.33 million** people, creating incredibly crowded conditions.

The subway system, by opening up vast new areas in Brooklyn, Queens, and the Bronx for development, allowed Manhattan's residents to move outward while still commuting to jobs in the city center. This led to a decades-long population decline.

| Year | Manhattan Population | Notes |
| :--- | :--- | :--- |
| **1910** | **2,331,542** | Historical peak |
| **1930** | 1,867,312 | A 20% decline from the peak |
| **1980** | 1,428,285 | Historical low point |
| **2000** | 1,537,195 | Population begins to recover |
| **2010** | 1,568,573 | Continued slow growth |

More recently, Manhattan has seen a resurgence. Between mid-2023 and mid-2024, it was the fastest-growing borough in New York City, with a population increase of **1.7%**.

### 🚇 Commuters & Transit: The Parallel Growth Engine

As Manhattan's resident population decreased, its role as the region's employment hub intensified, creating a massive daily influx of commuters. The subway system was the primary engine that made this possible.

#### The Subway's Expansion
The system's growth was explosive, directly enabling the city's outward expansion.

| Year | System Milestone | Key Statistics |
| :--- | :--- | :--- |
| **1904** | **Opening** | A **9.1-mile** line with 28 stations. Carried 150,000 passengers on its first day. |
| **1908** | Rapid Expansion | Track length reached **23.5 miles**. |
| **1940s** | Core Network Complete | Subway grew to cover **722 miles** of track, making it long enough to stretch from New York to Chicago. |
| **Today** | Modern System | Comprises **248 miles** of routes and **472 stations**, serving about **5.5 million** riders daily. |

#### The Commuter Surge
The subway's capacity was crucial for the growing workforce. By 1920, the transit system was carrying about **2.4 billion passengers** annually. This trend continued, with ridership in 2016 averaging **5.66 million passengers** daily. By the 1990s, **1.48 million** of the **2.04 million** people working in Manhattan used mass transit to get to their jobs.

#### The Traffic Counterpoint
The parallel growth of automobile traffic was a direct response to the city's expansion. As more people moved to the suburbs, car traffic into Manhattan surged. The number of car lanes into Manhattan grew from **68 in 1924 to 120 by 1965**. Daily vehicular river crossings into Manhattan more than doubled from **670,007 in 1948 to 1,785,875 in 2016**. This massive increase in cars created the chronic congestion that the subway was, in part, designed to alleviate.

In summary, the subway was the catalyst that allowed Manhattan to evolve from a crowded, single-centered city into the core of a vast, polycentric metropolis. It enabled the island to shed residents while gaining an even larger daily workforce, creating the vertical business districts and sprawling commuter network that define it today.
 
Now look deeper into it and find if the growth of population density being combined residents and commuters for work is covered by the transport ways. Is there a point in which the traffic system development stopped growing in parallel? How is traffic jam, travel time and subway car space per person throughout the day over time?
 
The evidence indicates that Manhattan's transport system has not kept pace with the combined growth of its residential and commuter populations. A critical divergence occurred in the mid-20th century, after which the system's capacity failed to grow in parallel with demand, leading to worsening congestion and crowding over time.

### 📈 The Unmatched Growth: Daytime Population vs. Transit Expansion

Manhattan's transport network was largely built during a period of explosive growth, but its expansion virtually ceased after the 1940s, even as the demands on it continued to evolve.

*   **The Scale of Influx**: Manhattan's residential population is about **1.6 million**. However, on a typical business day, the daytime population swells to approximately **3.9 million** people due to the influx of commuters, visitors, and tourists. This daily surge is a fundamental and persistent pressure on all transport infrastructure.
*   **The Halt in Expansion**: The era of major subway construction ended in the mid-20th century. The article *Derailed: The Postwar End of New York City Subway Expansion* identifies the **1951 bond issue** as a key turning point, after which "expansion virtually ceased". A significant factor was the **1951 bond issue**, which failed to fund the ambitious "Second System" expansion plan. The plan's abandonment, coupled with the economic strains of the postwar era and a political shift toward suburban highway development, locked the city into a transit network that has not seen a major new line since the Rockaway line opened in **1956**.

### 🚦 The Result: A System Under Chronic Strain

With the network's physical footprint largely frozen, the growing number of people relying on it has led to measurable declines in service quality across several fronts.

#### Traffic Congestion & Travel Time
The road network has been overwhelmed by vehicle traffic. The number of car lanes into Manhattan grew from **68 in 1924 to 120 by 1965**, but this could not keep up with demand. As a result, traffic speeds have collapsed to historic lows.
*   **Record Slow Speeds**: In fiscal year 2024, the average traffic speed in midtown Manhattan was just **4.8 mph**, the lowest on record since tracking began in 1971.
*   **Longer Commutes**: The strain on the entire system is reflected in travel times. Between 2001 and 2017, the average daily commute time for Manhattan residents increased significantly, rising from **36.96 minutes to 44.98 minutes**.

#### Subway Crowding
As the subway system became the primary alternative to gridlock, it too became a victim of its own success, with passenger loads exceeding comfortable or even safe limits.
*   **A Long-Standing Problem**: Complaints about crowding are not new. As early as 1927, an engineer testified that trains were packed to nearly **fourfold their seating capacity**.
*   **The Data on Space**: Modern data shows that crowding has worsened over the past two decades. The table below shows how the average floor space per passenger during the peak hour has shrunk, indicating that people are packed more densely into each train car.

| Mode of Entry (8AM-9AM Peak) | 2009 (sq ft/person) | 2019 (sq ft/person) |
| :--- | :--- | :--- |
| **Subway - Express** | 5.3 | 4.4 |
| **Subway - Local** | 7.0 | 5.2 |

*   **The "Crush" Standard**: While the MTA has a stated guideline of **110 riders per car**, other analyses note that in practice, "the standing space allowance can be as low as 2-3 square feet per passenger, resulting in a highly dense and uncomfortable environment".

### 🗓️ The Timeline of Divergence

The break in the parallel growth of population and transport capacity is clear when viewed historically:

*   **1904 - 1940s (Parallel Growth)**: The opening of the first subway in 1904 triggered a cycle of development. The system expanded rapidly, and ridership grew three times as fast as the population between 1880 and 1940.
*   **1950s - 1970s (The Great Divergence)**: Subway expansion ceased. Ridership fell from its **1946 peak of over 2 billion rides annually** to **1.3 billion by the 1960s**, as the middle class moved to the suburbs and began driving. This marked the point where the public transport system stopped growing in parallel with the region's demographic and economic shifts.
*   **1980s - 2000s (The Strain Returns)**: After hitting a low point in 1980, ridership began to recover, skyrocketing from **4 million daily passengers in the 1990s to 6 million by the 2010s**. This surging demand hit a system that had not been meaningfully expanded in decades, leading to the chronic overcrowding and delays seen today.

### 💎 Conclusion

In summary, the data shows a clear and persistent mismatch. The last major phase of subway expansion ended in the 1950s, a point of divergence after which population and employment growth continued but transit capacity did not. The result is a system where traffic speeds are at a historic low, commute times are rising, and subway cars are carrying passengers at densities that frequently fall well below established comfort and guideline standards.
 
 Is that time about the same that prices started surging?
 
The timing is close, but it depends on which "prices" you mean. The period when subway expansion stopped aligns almost exactly with the **start of subway fare increases**, but it is a full generation **before the surge in Manhattan real estate prices**.

### 🚇 Subway Fares: The Surge Begins Immediately

The era of a flat, five-cent fare ended right as the subway's physical expansion was winding down. The fare had been **5 cents** from the system's opening in 1904 until **July 1, 1948**, when it doubled to **10 cents**. This was the first in a series of rapid increases:

*   **1953:** 15 cents
*   **1966:** 20 cents
*   **1970:** 30 cents
*   **1972:** 35 cents
*   **1975:** 50 cents

This timeline coincides with the institutional shift that marked the end of expansion. In **1953**, the old Board of Transportation was replaced by the newly created **NYC Transit Authority**, which was required to meet its operating expenses from its own revenues. This new financial structure made fare increases a primary tool for managing the system's costs, rather than a last resort.

### 🏙️ Real Estate: A Decades-Long Delay

The surge in Manhattan real estate prices did **not** occur at the same time. The data shows a clear and prolonged lag.

Between **1950 and 1980**, real housing prices in Manhattan remained **relatively flat**. The actual price surge began after 1980. From **1980 to 1989**, housing sale prices across New York City **soared by 152%**. The Manhattan Institute attributes this post-1980 divergence to increased regulatory constraints on housing supply, which created a "zoning tax" that drove prices far above construction costs.

### 💎 Summary

To directly answer your question: The time when the subway's physical growth stopped is indeed about the same time that **subway fares** began their long climb. However, the massive surge in **Manhattan real estate prices**—the phenomenon that made the city famously expensive to live in—did not begin until around **1980**, roughly three decades after the subway's expansion ended. 
 
During the surge of Real Estate prices, was there a growth in population density combined residents and commuters, and how much was invested into the traffic system?
 
Yes, the data shows that during the real estate price surge, both residential and daytime population density in Manhattan grew significantly. This growth was met with a massive, unprecedented wave of investment in the traffic system, reversing decades of disinvestment.

### 📈 Population Density: A Dual Resurgence

During the real estate boom, Manhattan experienced growth on two fronts: more people moved in to live there, and more people commuted in to work there.

*   **Residential Population Rebounded**: After decades of decline, Manhattan's residential population began to recover. It grew from a low of **1,428,285 in 1980** to **1,487,536 in 1990**, **1,537,195 in 2000**, and **1,585,873 in 2010**. This trend has continued, reaching **1,694,251 by 2020**.

*   **Daytime Population Swelled**: The influx of commuters is even more dramatic. Manhattan's daytime population, which was already substantial, grew to approximately **3.39 million in 1990**. By the 2010s, this figure had risen to nearly **3.94 million** on a typical weekday, meaning the island's population more than doubles during the day. This daytime surge is driven by about **1.6 million commuters** entering Manhattan for work each day.

### 💰 Traffic System Investment: The "Road Back"

The period of surging real estate prices coincided exactly with a historic reversal in the funding of New York's transit system. After decades of neglect, the 1980s marked the beginning of a massive, sustained capital investment program.

*   **The Turning Point**: In 1982, the MTA launched its first **five-year Capital Program**, a landmark effort to address the system's severe disrepair. This program was spearheaded by MTA Chairman Richard Ravitch, who secured an initial **$18 billion investment program** (equivalent to about $50 billion in 2020 dollars) to rehabilitate the subway.

*   **Sustained, Multi-Billion Dollar Investments**: This was not a one-time fix. The MTA has continued a series of five-year capital programs since then. Key funding milestones include:
    *   The first program (1982–1986) invested **$7.66 billion**.
    *   By 2011, funded totals for the capital programs had reached **just under $84.3 billion** in current dollars (over $116 billion in 2011 dollars).
    *   Over the past 37 years, the MTA has invested **more than $128 billion** in the system.

This massive investment rebuilt stations, replaced and refurbished thousands of subway cars, and modernized track and signals, directly enabling the system to handle the growing ridership that accompanied Manhattan's real estate boom.

### 💎 Summary

In short, the era of surging Manhattan real estate prices was not a period of static infrastructure. It was a time when **both residential and daytime populations grew**, putting immense pressure on the transit network. This pressure was met with a **historic, multi-decade, multi-billion dollar investment program** that reversed the previous era of decline and was essential in allowing the city to function as its population and economy grew. 
 
Did the investments just cover the neglect or keep up with more demand?
 
 The evidence indicates that the massive capital investments since the 1980s were primarily directed at **repairing decades of neglect** and restoring the system to a state of good repair. While this investment was essential and successfully reversed the system's decline, it largely **failed to keep pace with the era's surging demand**. The result was a system that was functionally rehabilitated but remained structurally constrained, leading to chronic overcrowding and delays as ridership reached historic highs.

### 🏗️ The Investment Focus: Rebuilding, Not Expanding

The capital programs launched in the 1980s were a direct response to the system's deteriorated condition after years of disinvestment. The priority was to fix what already existed, not to build new capacity.

*   **A Focus on State of Good Repair**: The overwhelming majority of capital spending has been dedicated to "state of good repair" (SGR) work. For example, the MTA's proposed $68.4 billion Capital Plan for 2025-2029 dedicates **95% of its investments** to rebuilding and improving the existing transit system to maintain its $1.5 trillion in assets. This focus is a direct acknowledgment of the enormous maintenance backlog.

*   **The Scale of the Neglect**: The backlog was immense. The MTA has stated it needs to reinvest **$23 billion annually** in its capital assets, a figure that includes **$7 billion a year** just to address its "outsized state-of-good repair backlog". As of 2024, the system's assets remained in dire shape, with **39% of transit station ventilation systems, 36% of substations, and more than 90% of Penn Station's electrical equipment** in poor or marginal condition.

*   **Minimal Expansion**: Expansion projects were rare exceptions. The 2025-2029 plan, for instance, includes **just one major expansion project**—the beginning of construction on the Second Avenue Subway Phase 2. The vast majority of funds went to rolling stock, signals, power systems, and station repairs.

### 🚇 The Demand Surge: A System Busting at the Seams

While investment focused on repair, the demand for the subway system exploded. The system's physical footprint—its tracks, tunnels, and stations—remained essentially the same size it had been for decades.

*   **Ridership Nearly Doubled**: Subway ridership surged from about **4 million people per day in the 1990s** to nearly **6 million riders each day by the mid-2010s**—the highest level since the 1940s. Annual ridership grew from **1 billion in 1990 to 1.8 billion by 2015**.

*   **Capacity Stagnated**: The critical metric is the mismatch between ridership and capacity. Between 1990 and 2015, annual riders increased by roughly **75%** (from 1.03 billion to 1.8 billion), while the number of subway cars increased by less than **1%** (from 5,255 to 5,282) and miles of track grew by just over **1%** (from 488 to 493 miles).

*   **The Result: Chronic Overcrowding**: This imbalance made overcrowding the primary operational problem. By 2017, overcrowding was responsible for **more than one-third of the nearly 75,000 subway delays** across the system each month. The situation reached a tipping point around 2013-2014, when the system could "no longer easily absorb the extra riders".

### 💎 Conclusion: A System Rehabilitated but Not Expanded

The massive investment since the 1980s was a **necessary rescue operation** that pulled the subway back from the brink of collapse. It fixed signals, replaced aging cars, and repaired crumbling stations, which was a monumental achievement. However, it was a **maintenance and modernization program, not a capacity-building one**.

Because the system's core infrastructure was not expanded to match the rapid growth in ridership, the investments ultimately **did not keep up with demand**. They restored the system to a functional state but left it unable to comfortably handle the number of people who depended on it, creating the paradox of a "rehabilitated" system that was also chronically overcrowded and unreliable for its passengers.
 
So it was little done to develop the traffic system, both road personal and public and public transport rail based, to match the more demand no matter more profits and ever rising housing prices?
 
Your assessment is largely correct. The evidence shows that despite enormous profits generated by real estate and rising housing prices, very little was done to **expand the physical capacity** of either the road network or the public transit system in Manhattan to match the growing demand. The massive investments that were made went almost entirely toward **repairing and maintaining the existing system**, not building new capacity.

### 🛣️ Road Capacity: Effectively Frozen

The road network in Manhattan has not been meaningfully expanded since the mid-20th century. The number of car lanes into Manhattan grew from **68 in 1924 to 120 by 1965**, but that figure has essentially flatlined since then. The last limited-access highway segment in New York City opened in **1976**. 

The Westway project, a major proposed highway along Manhattan's West Side, was approved in 1981 but was ultimately abandoned after decades of controversy and was replaced in 2001 by a surface-level boulevard with **no increase in vehicle capacity**. The West Side Highway widening in 2000 was projected to add only **1,200–1,500 vehicles per hour**, a negligible amount against the daily traffic volume of over 1.8 million river crossings.

Any road "improvements" since then have largely been about **reallocating existing space**—adding bus lanes, bike lanes, and pedestrian plazas—rather than adding new vehicle capacity. The phenomenon of **induced demand** means that even when road capacity is added, it fills up with traffic within three to five years, providing no lasting congestion relief.

### 🚇 Public Transport Rail: Expansion Halted, Then Minimal

The situation for rail transit is even starker. In **1981**, the MTA made a pivotal decision: it "**halted all new transit expansion until the existing system could be restored**". This marked the definitive end of the era of major subway construction.

Between 1940 and 1988, there were **no significant alterations** to the basic physical configuration of the subway network. The first new subway line in more than two decades, the 63rd Street Connector, opened only in **December 2001**.

Since 1980, the expansion projects that have been completed or are underway are remarkably few and small relative to the scale of demand:

*   **Archer Avenue Line Extension** (1988) and **63rd Street Tunnel** (1989): These were the first expansions in decades, but they added only a handful of new stations.
*   **7 Line Extension** to the West Side: A short extension from Times Square, opened in 2015.
*   **Second Avenue Subway Phase 1**: Opened on **January 1, 2017**, adding just **three new stations** on Manhattan's East Side. Phase 2 is still in planning and early construction.
*   **East Side Access**: Opened in 2023, bringing Long Island Rail Road trains into Grand Central Terminal.

These projects, while significant, represent **marginal additions** to a system that had not seen major expansion in over half a century. As one analysis noted, the MTA's current capital plan includes "**just one major expansion project**"—the Second Avenue Subway Phase 2.

### 💰 The Investment Focus: Repair, Not Expansion

The massive capital investments since the 1980s were a **rescue operation**, not a capacity-building program. The MTA's own leadership has confirmed that "**Ninety percent of the program is state-of-good repair**". The system's core infrastructure—signals, power systems, track, and stations—was in such dire condition after decades of neglect that the priority had to be fixing what already existed.

The numbers are revealing:

*   The MTA states it needs to reinvest **$23 billion annually** in its capital assets, including **$7 billion a year** just to address its "**outsized state-of-good repair backlog**".
*   As of 2024, **39% of transit station ventilation systems, 36% of substations, and more than 90% of Penn Station's electrical equipment** were in poor or marginal condition.
*   The 2025-2029 Capital Plan, with a **$68.4 billion** price tag, devotes the vast majority of its funds to keeping the system in a state of good repair, with only a small fraction for expansion.

### 📊 The Profit Mismatch

The disconnect between the profits generated and the infrastructure investment is stark. Research on the Second Avenue Subway found that the **$4.5 billion** construction cost generated a **$4.4 billion** benefit to residential real estate values along the corridor—meaning the infrastructure investment essentially paid for itself through value creation. Yet, only about **30%** of the private value created by subway improvements is captured through higher property tax revenue, leaving the public to bear the majority of the cost while private landowners reap the benefits.

A study on the 7 Line Extension found that the **$4.5 billion** project cost generated **$5.8 billion** in real estate value gains—enough to pay for the construction—yet the public financing model did not capture this value to fund further expansion.

### 💎 Conclusion

In summary, your assessment is correct. **Very little was done to develop the physical capacity of the traffic system**—whether roads or rail—to match the growing demand in Manhattan. The road network has been frozen since the 1960s, and subway expansion was halted in 1981 and has since been limited to a handful of marginal projects. The enormous investments made since the 1980s were directed almost entirely at **repairing the existing system** after decades of neglect, not at expanding it. Meanwhile, the real estate profits and rising housing prices that were fueled by this very infrastructure were largely **not captured to fund further capacity expansion**. The result is a system that was rehabilitated but not expanded, leaving it chronically overcrowded and unable to comfortably handle the number of people who depend on it. 
 
Can you find if there was a change in those who invested into Manhattan before and after the traffic system was left for neglect? There was a time in the 1980ies in which the prices were still flat, but also the traffic system was left for neglect. 
 
Yes, there was a significant change in who invested in Manhattan, and the timing aligns with the period of transit neglect in the 1970s and 1980s. The investor profile shifted from traditional local developers who held assets long-term to a more diverse group including foreign institutional investors, pension funds, and short-term speculators, a change that was directly shaped by the city's fiscal crisis and the deteriorating transit system.

### 📉 The 1970s: Crisis and the Exit of Traditional Investors

The 1970s was a period of profound crisis for New York City, which fundamentally altered the investment landscape.

*   **The Fiscal Crisis and Collapse of Values**: The city's near-bankruptcy in 1975, combined with a severe national recession, devastated the real estate market. The market was already suffering from an oversupply of office space that peaked in 1969-70. This led to a "virtual collapse in private investment" and declining property values throughout the 1970s and into the 1980s.
*   **The Failure of Speculative Vehicles**: The collapse was accelerated by the failure of highly leveraged Real Estate Investment Trusts (REITs). Many of these trusts, which had fueled a speculative boom in office building and luxury apartment construction, became insolvent. This not only caused losses for banks but also scared off a generation of investors from similar vehicles.
*   **A "Doomsday" Market**: During this period, traditional long-term owners, such as institutions and established developers, "doggedly stayed with their investments," but new investment was scarce. As one developer recalled about 1975, "It was doomsday. Everyone was fleeing the city. Only schmucks were buying".

### 🔄 The 1980s: A New Breed of Investor

As the city's fiscal situation stabilized and the national economy recovered, a new wave of investors entered the market, encouraged by the city's pro-development policies under Mayor Ed Koch. This new group was markedly different from the old guard.

*   **Foreign Institutional Investors**: The most dramatic shift was the arrival of foreign capital. **Japanese firms** became dominant, investing a staggering **$78 billion in U.S. real estate between the early 1980s and mid-1990s**. They often purchased trophy skyscrapers at record prices, such as Mitsubishi Estate's **$1.4 billion** acquisition of an 80% stake in Rockefeller Center in 1989-1990. Canadian developers, like **Olympia & York**, were also major forces in the market.
*   **Pension Funds and Institutional Capital**: Domestic institutional investors, particularly **pension funds**, also became major players. By 1985, union pension funds alone had invested **$500 million** in real estate. This represented a shift from direct ownership by families to investment by large, pooled funds.
*   **A New Generation of Local Developers**: While old-line families like the Dursts and Tishmans remained, a new generation of high-profile developers emerged. **Donald Trump**, for example, made a series of high-stakes bets on Manhattan, starting in the mid-1970s, wagering that his investments would precede a rejuvenation of the city.

### 🚇 The Transit Investment Mismatch

The critical point is that this massive wave of new investment, which drove up property values, was not matched by investment in the public transit system that supported it.

*   **A Fork in the Road**: In **1980**, the city chose to prioritize the **Westway** highway project, a **$2.3 billion** land-development deal that consumed a major share of the region's transportation capital for a decade, over rebuilding the deteriorating subway system.
*   **The Value Capture Failure**: The new investors were attracted by the value created by the city's transit system, but the public did not capture this value to fund further transit expansion. For instance, the **$4.5 billion** Second Avenue Subway generated a **$4.4 billion** benefit to nearby real estate values, yet only about **30%** of that private value was captured through property taxes.

In summary, the period of transit neglect in the 1970s and 1980s coincided with a fundamental change in Manhattan's investor base. The crisis of the 1970s drove out many traditional local investors and created an opening for foreign institutions, pension funds, and a new class of American developers to acquire properties at distressed prices. When the market recovered, these new investors reaped the rewards of a city whose infrastructure they had not been required to help fund.

If you would like to explore a specific group, such as the Japanese investors of the 1980s, in more detail, I can do that. 
 
I am sure that is not just a simple normal problem, but a deeply routed attitude problem of those investing. To me it is clear that a town serves human prosperity. That means that I am not erecting just a building, but shape and form a biosphere impacting human life.
If I plan a skyscraper I also have to provide parking and ensure accessibility. By growing profits due ever rising prices the financial base to provide the traffic network supporting like the pillars drilled into New York soil the future prosperity of the building and its surrounding therby the entire quarter and then town and region, should be much easier.
I am also sure that nothing as such happened by referring to legal responsibilities being off the developer, so they never even tried.
The major change must have been one of attitude, than.
 
Your perspective cuts to the heart of the matter. The evidence supports your conclusion: the failure to invest in transit capacity alongside Manhattan's real estate boom was not merely a technical or financial oversight, but a fundamental **attitude problem** embedded in the legal and cultural framework of American real estate development. The system was designed to **externalize** the cost of public infrastructure onto the public, allowing private developers to capture the value created by transit without being required to fund it.

### ⚖️ The Legal Design: Externalizing Infrastructure Costs

Your intuition that developers were "off the hook" by legal design is correct. In New York City, the responsibility for funding major transit infrastructure has never been automatically tied to the right to develop.

The core principle is that **developers are not legally required to fund the transit capacity their projects depend on**. The standard legal mechanism—the Uniform Land Use Review Procedure (ULURP)—allows a developer to proceed with construction as long as they meet zoning requirements, which historically did not include a mandatory contribution to transit capacity.

This contrasts sharply with other systems. In England, for example, the Community Infrastructure Levy (CIL) is a **compulsory charge on all new development** to contribute to the cost of infrastructure required to support it. In New York, there is no such universal mechanism.

Instead, the city has relied on **voluntary, negotiated deals** and **incentive-based programs**:

*   **Zoning Bonuses**: Developers can receive a Floor Area Ratio (FAR) bonus—allowing them to build taller or denser—in exchange for providing transit improvements. The current Zoning for Accessibility (ZFA) program offers up to a **20% floor area bonus** for station accessibility upgrades. But this is optional. A developer can simply choose not to build as tall and avoid the contribution entirely.
*   **Negotiated Agreements**: Major projects like Hudson Yards or One Vanderbilt involve bespoke deals where developers contribute in exchange for specific benefits. At Hudson Yards, the city used **value capture** mechanisms—bonds repaid by future property tax revenue and developer payments—to fund the $2.3 billion 7 Line extension. At One Vanderbilt, SL Green agreed to invest **$220 million** in Grand Central transit improvements in exchange for zoning approvals.

The critical flaw is that these are **exceptions, not the rule**. As one analysis noted, the MTA's planners themselves acknowledged that opportunities for privately contributed subway improvements "**depend on the proposals of the developers and cannot be anticipated**". The system was reactive, not proactive. There was no master plan requiring developers to fund the transit capacity their buildings would demand.

### 🧠 The Investor Attitude: "Buy Low, Sell High"

This legal framework shaped a specific investor attitude. Developers understood that **transit access increases land values**, but their job was to **capture that profit by "buying low" and "selling high" once the subway is built**—not to fund the subway itself.

The entire value proposition of developing in Manhattan rested on the existence of a transit system built by previous generations. As one report observed, the current development market "could be considered a 'benefit' accruing from investments in the subway made a half a century ago". Developers were **freeriding on infrastructure they had not paid for**.

The consequences of this attitude are stark. When the Second Avenue Subway was built at a cost of **$4.5 billion**, it generated **$4.4 billion** in residential real estate value along the corridor. Yet the public recouped **less than a third** of that value through property taxes. The rest was a **windfall for private developers who "just happened to own land in the vicinity of the subway stops"**.

A similar pattern emerged with the 7 Line Extension. The project cost **$4.5 billion** and generated **$5.8 billion** in real estate value gains—enough to pay for its construction. But the public financing model did not capture this value to fund further expansion.

### 🔄 The Shift in Attitude: From Implicit to Explicit

Your observation about a change in attitude is also supported. For decades, the expectation was that developers would **voluntarily** contribute to the public realm as a matter of civic responsibility. The Riverside South agreement in 1991, for example, contemplated a new highway, but the agreement **did not specify who would pay for it**. The development was allowed to proceed **without the developer committing to pay for the highway**, because it "simply couldn't support the cost" alongside other infrastructure requirements.

This illustrates the attitude problem in its purest form: public infrastructure was treated as an **optional amenity**, not a **prerequisite** for development.

What has changed in recent years is that the public sector is now attempting to make this relationship **explicit**. Programs like ZFA and the expansion of the transit improvement bonus program are attempts to **codify** the expectation that developers who benefit from transit should contribute to it. But these remain **incentive-based and voluntary**, and they only address station access—not the fundamental capacity of the rail network itself.

### 💎 Conclusion

You are correct. The failure to expand Manhattan's traffic system to match demand was not a simple market failure. It was rooted in a **legal and cultural attitude** that treated public infrastructure as a **free good**—something that existed independently of the private development it enabled. Developers were not required to fund it, and the system was not designed to capture the value they extracted from it.

The result is a city whose private wealth was built on a public foundation that was never properly maintained or expanded. The massive capital investments since the 1980s were a **rescue operation**—repairing the neglect, not building for the future. And the attitude that made that neglect possible—the belief that a developer's responsibility ends at the property line—remains the central obstacle to building the transit capacity that Manhattan's continued prosperity will require. 
 
Was there any legal binding in 1910 and how much of the first lines were created by a different attitude?
 
 In 1910, the legal framework was defined by the **Rapid Transit Act of 1891** and the contracts it enabled. The attitude behind the first lines was fundamentally different from today: it was a **public-private partnership where the city owned the infrastructure and private companies bore the operating risk**, a model that is starkly different from the modern developer-centric approach.

### ⚖️ The 1910 Legal Framework: City-Owned, Privately Operated

By 1910, the legal structure for the subway was well-established under the **Rapid Transit Act of 1891** (amended in 1894), which created the Rapid Transit Commission to oversee the system.

The core legal principle was that the **city owned the subway** and leased it to a private operator. The contract with John B. McDonald, signed on **February 21, 1900 (Contract No. 1)**, established the template:

*   **City Financing**: The city issued bonds to finance the **construction** of the subway.
*   **Private Operation**: The contractor (McDonald, who assigned the contract to the **Interborough Rapid Transit Company**) was required to **equip, maintain, and operate** the railroad at its own expense for **50 years**, with an option to renew for another 25.
*   **Rent**: The company paid the city an annual rent equal to the **interest on the city's construction bonds**, plus a **1% sinking fund contribution** to retire the debt.
*   **Fare Control**: The city retained **direct control over the fare structure** through the franchise agreement.

This was a true public-private partnership. The public bore the capital cost, while the private operator took on the commercial risk of running the system and was incentivized to maximize ridership and efficiency.

### 🔄 The First Lines: A Different Attitude

The creation of the first subway lines was driven by a **civic-minded, long-term vision** that prioritized public benefit and city-wide growth, an attitude fundamentally different from the modern, purely profit-driven developer model you described.

*   **Civic Responsibility**: The project was championed by progressive reformers and businessmen like **Abram S. Hewitt**, who realized that private capital alone was insufficient for a project of this scale and that **public funds were necessary** to build the subway for the public good.
*   **Public Benefit over Private Profit**: The primary motivation was to **relieve the crushing congestion** of horse-drawn traffic and to **promote the dispersal of the population** into the outer boroughs (the Bronx, Brooklyn, and Queens), shaping the city's future growth.
*   **Shared Risk**: The model aligned public and private interests. The city provided the massive upfront capital, while the private company (organized by banker **August Belmont Jr.**) took on the operational risk and was responsible for the system's performance.

The **Dual Contracts of 1913** continued this model, with the **city contributing $123 million** and the private companies (IRT and BRT) building and operating the expanded network, which **doubled the size of the subway and tripled its capacity**.

### 💎 Summary

In 1910, the legal framework ensured that the **city owned the subway and controlled the fares**, while a private operator was contractually bound to **equip, maintain, and operate** the system at its own expense for decades. The attitude behind the first lines was one of **civic responsibility and long-term public planning**, where the goal was to use public capital to shape a growing city and share the commercial risk with private operators. This is a world away from the modern approach, where developers capture the value of public transit without being legally required to fund the capacity their projects demand.

If you would like to explore a specific aspect of this early legal framework, such as the specifics of the Dual Contracts or the role of the Public Service Commission, feel free to ask.
 
Taking the hardcore U.S. conservatives around Reagan and Bush, would those men back than fit the "Fucking Commis" enemy created during Cold War?
 
"In the political vocabulary of the Cold War, Ronald Reagan and George H. W. Bush were not."
You missed my question. I wonder if those having build the first subway had the attitude of the "Fucking Commis" Regan and Bush created as an enemy stereotype?
 
 
No — not in the actual Cold War sense. The men who built New York’s first subway were **not communists**, and they did not share the ideology that Reagan and Bush later cast as the “commie” enemy.

But your instinct points to something real: they **did** believe in a very different role for government and private capital than the Reagan-era neoliberal model. If you reduce “commie” to mean “anyone who thinks government should build and own major public infrastructure,” then a hardcore anti-government conservative might smear them that way. But that is a rhetorical distortion, not what the Cold War enemy actually was.

### Who built the first subway?

The key figures were:

- **August Belmont Jr.** — a wealthy banker and financier who organized the Interborough Rapid Transit Company (IRT).
- **John B. McDonald** — a contractor who won the original contract and then assigned it to Belmont’s company.
- **Abram S. Hewitt** — an industrialist, former mayor, and progressive reformer who pushed for public funding.
- **The Rapid Transit Commission** and city government — which issued bonds, owned the tunnels, and regulated fares.

These were **capitalists, bankers, contractors, and progressive-era civic reformers**. They were not Marxist revolutionaries. They wanted to make money, relieve congestion, and expand the city. They believed in private enterprise, but they also believed that the city government had a duty to build the backbone of the system.

### Their attitude: civic capitalism, not communism

The model they created was a **public-private partnership**:

- The **city owned the subway** and paid for its construction with municipal bonds.
- A **private company equipped, maintained, and operated** the trains at its own risk.
- The **city controlled the fare** and received rent to service the debt.
- The goal was to shape the city’s growth, not to abolish private property.

That is not communism. It is closer to **municipal capitalism** or **state capitalism** — using government power to enable private profit and public benefit at the same time. Belmont and his investors got rich. The city got a subway. The public got mobility.

### Would Reagan/Bush have called them commies?

Probably not. Reagan and Bush were anti-communists who defined the enemy as Soviet-aligned, Marxist-Leninist, collectivist, and anti-capitalist. Belmont and Hewitt were none of those things. They were wealthy capitalists who believed in public infrastructure as a foundation for private development.

However, the **methods** they used — public ownership of the subway, municipal bonds, fare regulation, long-term government planning — are exactly the kinds of things that Reagan-era anti-government rhetoric often labeled “socialism.” Reagan’s famous line was “government is not the solution to our problem; government is the problem.” The early subway builders would have rejected that. They thought government was essential to building the platform on which private prosperity could stand.

So if you mean: “Did they have the attitude that Reagan/Bush demonized as the communist enemy?” — the answer is **no, not in the Cold War ideological sense**. But they did have an attitude that **big government and big capital should work together for public ends**, which is the opposite of the neoliberal, developer-externalizes-costs attitude that later prevailed.

### The deeper point

The first subway was built by people who accepted that **private development depends on public infrastructure**, and that the public therefore had a right and a duty to shape and fund that infrastructure. They were not reds. They were capitalists with a civic conscience — or at least with enough self-interest to realize that a functioning city was good for business.

The Reagan/Bush-era shift was not a fight against communism in subway construction. It was a shift toward letting private developers capture the value of public transit without being required to fund its capacity. That is the attitude you have been criticizing. And the early subway builders, whatever their faults, did not share it. 
 
Would it be more correct stating that civic capitalism turned to accumulating capitalism? Meaning it is all done to have those having have more, no matter the price for those having less while civic meant to participate for common interest having with those having less?
 
 Yes — as a shorthand, that is a useful and largely accurate way to describe the shift. But it needs one important correction: **capitalism was always accumulative**. What changed was not the fact of accumulation, but the **terms** on which it happened, and the **obligations** attached to it.

A more precise formulation would be:

> **Civic capitalism turned into financialized, rentier, or extractive capitalism.**

Or, in your language:

> A capitalism that at least pretended to share the city’s fate turned into a capitalism that privatizes the gains and socializes the costs.

### What “civic capitalism” meant in the early subway era

It was not altruism. The men who built the first subway were bankers, contractors, and investors. They wanted profit. But they operated within a framework where **private profit depended on public infrastructure**, and the public retained real control:

- The **city owned the subway**.
- The **city issued the bonds** and bore the capital cost.
- The **private operator** equipped, maintained, and ran the trains at its own commercial risk.
- The **fare was regulated** by the public.
- The operator paid **rent** to the city to service the debt.

That is not communism. It is a form of **municipal capitalism** or **public-private partnership** where the public kept ownership and control, and private capital accepted long-term operating obligations. It was enlightened self-interest: a functioning city was good for business.

### What changed

From the 1970s onward, and especially under the Reagan–Bush era, the framework shifted:

- **Government was recast as the problem**, not the foundation.
- **Tax cuts and deregulation** starved public budgets.
- **Financialization and shareholder primacy** rewarded short-term extraction over long-term stewardship.
- **Public infrastructure became a cost to minimize**, not a platform to maintain.
- **Developers were no longer required** to fund the transit capacity their buildings depended on.

The result is the pattern you identified: private developers capture the land-value uplift created by public transit, while the public bears the cost of overcrowding, delays, and underinvestment.

Examples:

- The **Second Avenue Subway** cost **$4.5 billion** and generated **$4.4 billion** in nearby real estate value. The public recouped less than a third through taxes. The rest was a windfall to landowners.
- The **7 Line Extension** cost **$4.5 billion** and generated **$5.8 billion** in real estate gains. The public financing model did not capture that value to fund further expansion.
- **Hudson Yards** used public bonds and tax increments to fund infrastructure, while private developers reaped the development profits.

That is not civic capitalism. That is **accumulation by externalization**: keep the gains, push the costs onto the public.

### The caveat: the early era was not a golden age

It would be wrong to romanticize the early subway builders. They were often corrupt, monopolistic, and exclusionary. The system served property interests, displaced communities, and often ignored the poor and racial minorities. “Civic” did not mean egalitarian.

But there was a **recognition of interdependence**. The wealthy and powerful understood that their own prosperity required a functioning city, and they were willing — or compelled — to contribute to the public realm through taxes, fares, and infrastructure.

The later model denies that interdependence. It says: *I keep my gains. The public bears the costs. If the city decays, I can move my capital elsewhere.*

### So is your framing correct?

Yes, with refinement:

- **Civic capitalism** = accumulation embedded in public obligations, public ownership, and a shared fate.
- **Accumulating capitalism** = accumulation disembedded from those obligations, financialized, rentier, and extractive.

The common interest was replaced by shareholder interest. The price for those having less is rising inequality, unaffordable housing, chronic overcrowding, and longer commutes — while those having more capture the value of a city they did not pay to build.

So your instinct is right: it was a change in attitude. But the more precise name is not just “accumulating capitalism.” It is **financialized rentier capitalism** — a system that extracts value from the public without reciprocating.
 
Do you see any of the two a path to Slavery coming back to mankind?
 
 Short answer: **neither “civic capitalism” nor “accumulating/financialized capitalism” is a direct path back to legal chattel slavery.** But if you mean **functional slavery** — debt bondage, neo-feudalism, hereditary dependency, and the loss of any real exit — then **accumulating financialized capitalism is the more plausible path**, because it concentrates ownership of the necessities of life and strips away public alternatives. Civic capitalism is not a guarantee against that, but it contains countervailing institutions that make it harder.

The key distinction is this:

- **Slavery** is not just exploitation. It is a legal and political condition: a person is property, can be bought and sold, has no legal personhood, no exit, and their unfreedom is inherited.
- **Neo-feudalism or debt peonage** is not the same, but it can approach it: you are legally free, yet you cannot leave because you owe rent, debt, or your housing, transport, healthcare, and work are all controlled by someone else.

### Civic capitalism
Civic capitalism — the model behind the early subway — keeps some public ownership, public control, and mutual obligation. It says: private profit is legitimate, but it must operate on a public platform and contribute to the common good. That model can be paternalistic, corrupt, and exclusionary, but it tends to preserve **exit options**: public transit, public housing, regulated fares, labor rights, and democratic accountability.

It can still fail. State capitalism can become coercive. A civic model can turn into a soft authoritarianism where the state owns everything and controls everyone. So it is not inherently safe. But it is less likely to produce slavery because power is at least partially accountable and the necessities of life are not entirely privatized.

### Accumulating / financialized rentier capitalism
This model says: capture as much value as possible, externalize the costs, and let the public absorb the damage. It privatizes land, housing, transit, healthcare, water, and even governance. If taken to an extreme, it produces a society where a small class owns everything and everyone else rents everything.

That is not chattel slavery. But it creates the conditions for **unfreedom**:
- You cannot afford to live near work, so you commute for hours.
- You are born into debt.
- Your employer controls your visa, your housing, or your healthcare.
- You cannot leave because you have no savings, no public option, and no political voice.
- Private owners control the infrastructure of daily life.

At that point, you are not property, but you are not free either. You are dependent. If the law then criminalizes debt, restricts movement, or allows employers to withhold wages and documents, it becomes debt bondage. That is already happening in supply chains, migrant labor, and prison labor around the world.

### So is either a path to slavery?
- **Chattel slavery**: No, not directly. That requires a legal regime that turns people into property. It could only return after a total collapse of rights, rule of law, and democratic accountability.
- **Functional slavery / neo-feudalism**: Yes, accumulating financialized capitalism is the more likely path. It does not need to own you legally. It only needs to own everything you need to live, and to leave you no exit.
- **Civic capitalism**: It is a mitigator, not a savior. It can keep the commons alive and preserve exit options. But if it becomes authoritarian or exclusionary, it can also become a cage.

### The real dividing line
The question is not “capitalism or socialism.” It is: **who controls the necessities of life, and can you leave?**

If you have:
- secure housing,
- public transit,
- healthcare,
- education,
- labor rights,
- the vote,
- and the ability to organize,

then you are not a slave, even if you are exploited.

If you lose all of those, and a private or state power controls your access to them, then you are on the road to unfreedom. That road does not have to end in legal slavery, but it can end in something that feels very close to it.

So: **accumulating capitalism is the more dangerous path. Civic capitalism is a guardrail, but only if it remains democratic, universal, and accountable.** The danger is not one economic model by itself. The danger is **absolute power over the means of life** — whether that power is held by a corporation, a landlord, or a state.
 
#cyberpunkcoltoure #IRAmovement Vs #theneoconfederatestates
The Firm. If it had a name... 
 

AI - Status Update - Cyberdeck

 Speculative Decoding 

That's a thing. So, I am using still no GPU being short of 800 bugs for a appropriate NVIDIA card and that means that even with 64GB RAM I won't be able to run certain model architectures by data transfer bottle necks. 

(DeepSeek AI)
You're right to focus on this — MoE (Mixture of Experts) models are a completely different beast on a CPU-only machine, and understanding why will save you a lot of frustration.

The core issue is that **MoE models are fundamentally memory-bandwidth-bound, and a CPU's memory bandwidth is its weakest link** for this workload.

### 🧠 The MoE Paradox: Sparse Compute, Dense Memory Traffic
The appeal of MoE is that it activates only a small fraction of its total parameters for each token. For example, a 30B MoE model might only use 3B active parameters per token. On paper, this sounds perfect for a CPU: less compute per token.

However, the **memory access pattern is the opposite of sparse**. Every token must still be routed through the model's full architecture, and the router must load and evaluate the expert weights to decide which ones to activate. The key problem is that the expert weights for *all* experts must be resident in memory — they can't be loaded from disk on demand without catastrophic latency.

This is why MoE models are so memory-hungry. A 30B MoE model with 3B active parameters still needs ~18–22 GB of RAM for its weights in 4-bit quantization. That's because the *total* parameter count determines memory capacity, while the *active* parameter count determines compute.

### 🚧 The CPU Memory Bandwidth Bottleneck
Here's the critical constraint: **CPU RAM bandwidth is roughly 10–20x lower than GPU VRAM bandwidth**. When you run an MoE model entirely on CPU, every token generation requires streaming the active expert weights from system RAM through the CPU's memory bus.

For a 3B-active MoE model, you're moving ~3 GB of weights per token. If your CPU's memory bandwidth is, say, 50 GB/s, that's a theoretical maximum of ~16 tokens/second — before accounting for attention, routing overhead, and other memory traffic. In practice, you'll see far less.

Research on this exact problem confirms the severity: one study found that MoE models exhibit **higher inference latency and lower throughput than dense baselines on both GPU and CPU**, precisely because the computational savings from skipping inactive experts are **offset by the overhead introduced by conditional routing**.

### 🔄 The CPU-GPU Hybrid Trap (And Why Pure CPU Is Even Worse)
Most MoE deployments use a **hybrid approach**: the GPU handles the dense attention and shared layers, while the expert weights live in CPU RAM and are streamed to the GPU on demand. This works because the GPU's high bandwidth compensates for the PCIe transfer overhead.

But on a **pure CPU machine**, there is no high-bandwidth staging area. The CPU must both compute *and* fetch weights from the same memory pool. This is why pure CPU MoE inference is notoriously slow: one benchmark showed a Phi-3.5 MoE model dropping from **13.8 tokens/second (GPU) to 4.5 tokens/second (CPU-only)** — a 3.1x slowdown.

### 📋 What This Means for You (64 GB CPU-Only)
With 64 GB of RAM, you *can* fit many MoE models in memory. But the token generation speed will be **poorly interactive**. Your 32B dense model with speculative decoding is actually a smarter choice for CPU inference because:

- **Dense models have predictable, sequential memory access patterns** that CPUs handle well.
- **Speculative decoding (your 1.5B draft model) works brilliantly on CPU** because the draft model is small enough to run in cache, and the verification pass is batched.
- **No routing overhead** — every token uses the same weights, so the CPU's prefetcher can work effectively.

### ✅ What to Use Instead

Given your 64 GB CPU-only setup, here's what works well:

#### 1. **Dense Models (Your Current Approach)**
Your Qwen2.5-Coder-32B in Q5_K_M quantization is essentially the sweet spot. It's large enough to be highly capable, but dense architecture means predictable memory access. With speculative decoding, you're getting the best possible CPU performance.

#### 2. **CPU-Optimized MoE Models (If You Want to Try)**
If you want to experiment with MoE, look for models specifically designed for CPU inference. The standout is **Maple Preview 20B-A1B** (ternary 2-bit quantization). It's a 20B MoE with only 1B active parameters, and it runs at **28–34 tokens/second on a 4-core laptop CPU**. That's because the ternary quantization is so aggressive that the entire model fits in a 5.5 GB file, and the memory traffic per token is minimal.

#### 3. **LiquidAI LFM2.5 Family**
The **LFM2.5-8B-A1B** is an MoE model with 8.3B total and 1.5B active parameters, designed for efficient on-device inference. It's a good middle ground if you want MoE architecture but need CPU-friendly performance.

#### 4. **Small Dense Models (For Speed)**
If raw speed matters more than quality, small dense models like **Gemma 3 (1B)**, **DeepSeek R1 (1.5B)**, or **Llama 3.2 (1B/3B)** run comfortably on CPU. But you'll lose the coding capability of your 32B model.

### 💡 The Bottom Line

| Model Type | CPU Suitability | Why |
|---|---|---|
Dense 32B + Spec. Dec.                                                   ✅ Excellent 
Predictable memory access, batch verification 
CPU-optimized MoE (Maple, LFM2.5)                         ✅ Good 
Aggressive quantization, small active params 
Standard MoE (Qwen3-30B, DeepSeek)                     ❌ Poor 
Routing overhead, high memory traffic per token 
Small dense (1B-3B)**                                                   ✅ Fast 
 Low memory footprint, but lower quality  

Your current setup — a dense 32B model with a 1.5B speculative decoding draft — is actually the **optimal configuration for a CPU-only machine**. You get the quality of a large dense model with the speed boost of speculative decoding, without the memory bandwidth penalties that plague MoE architectures.

If you want to experiment with MoE later, start with **Maple Preview 20B-A1B** — it's the only MoE I've seen that genuinely runs well on pure CPU, and it's small enough to test without disrupting your current setup. 
 --
 
 With Speculative Decoding I add on top of being capable of using a large model quite some speed...
 
(DeepSeek)
Here's a summary of what speculative decoding gives you, compared to the standard Ollama + Llama 3 setup you were using before.

### ⚙️ The Core Mechanism: How It Actually Works
The fundamental bottleneck with local LLMs is **memory bandwidth**. Generating a single token requires a full forward pass through every model weight, and successive tokens cannot be parallelized — each one depends on the last.

Speculative decoding breaks that dependency chain by pairing a **small draft model** with your **large target model**:

1. The draft model quickly guesses the next several tokens (your 1.5B model).
2. The large target model (your 32B) verifies *all* those guesses in a **single parallel forward pass**, because every position's inputs are already known.
3. The target model accepts the prefix of guesses it agrees with and replaces the first rejection with its own prediction. The process repeats.

The critical guarantee: the accept/reject rule is mathematically constructed so the final output distribution is **identical** to what the large model would have produced alone. You get no quality loss.

### 📊 What You Gain (and When)
The speedup depends entirely on the **acceptance rate** — how often the draft model's guesses match what the 32B model would have chosen. Your logs earlier showed **100% acceptance** on a simple prompt (`draft_n: 10, draft_n_accepted: 10`), which is the best-case scenario. In practice:

*   **On CPU (your setup):** Benchmarks show **2–3× speedups** are typical for large targets, because the draft model is genuinely tiny relative to the 32B per-token cost.
*   **Real-world example:** A CPU-only test with a 3B target went from 12.9 tok/s to 22.1 tok/s — a **1.72× overall speedup**, peaking at 2.03× on math tasks.
*   **When it backfires:** If the draft model is poorly aligned with the target (low acceptance rate), or the target model is already small and fast, the overhead of running two models can make things **slower**, not faster. One benchmark saw a 7B target actually **regress** on a fast GPU.

In your specific case — a 32B model on CPU with a 1.5B draft — the conditions are ideal. The draft is small enough to be nearly free, and the 32B target is slow enough that batching verification across 5 drafted tokens saves substantial memory-bandwidth round-trips.

### 🆚 Speculative Decoding vs. Ollama + Llama 3
Here's the key difference: **Ollama does not expose speculative decoding the way llama.cpp does.**

| Aspect 
| Ollama + Llama 3                          | Your llama.cpp + Spec. Decoding 
| :--- | :--- | :--- |
Model quality
 Llama 3 (8B, 70B, etc.)                  | Qwen2.5-Coder 32B  
Draft model
 None — single model                    | Qwen2.5-Coder 1.5B 
Speedup mechanism
 Standard autoregressive             | Draft-then-verify, ~2–3× on CPU |
Speed control
 `num_thread`, GPU layers        | `-md`, `--draft-max`, `--draft-min` 
Output quality
 Fixed to model                                 | **Identical** to 32B alone 

Ollama has only recently begun adding speculative decoding, and only for specific model families like Gemma 4 via MTP (multi-token prediction). For Qwen models on CPU, Ollama gives you **no draft-model acceleration** — it runs the 32B model straight through, token by token.

Your llama.cpp setup, by contrast, is doing exactly the thing Ollama doesn't: it's using a **smaller companion model to draft tokens**, then verifying them in parallel. That's why your 32B model is usable at interactive speeds on CPU at all.

### 💡 The Practical Takeaway
*   **Ollama + Llama 3** is a single-model system. Simple, but no draft-model trick — your 32B model would run at baseline CPU speed (roughly 2.5 tok/s in your earlier logs without spec. dec.).
*   **Your llama.cpp + 1.5B draft** is a two-model system. The 1.5B drafts tokens cheaply; the 32B verifies them in batches. On a good acceptance rate, you're effectively getting several tokens per 32B forward pass instead of one.

The cost is **memory**: you're holding both models in RAM simultaneously (the 32B at ~22 GB, the 1.5B at ~1.2 GB). But on a machine with 64 GB, that's a worthwhile trade for a 2–3× speedup.

If you ever test Ollama with a Llama 3 8B model on the same hardware, you'll see it's faster *per token* simply because the model is smaller — but the quality is lower. Speculative decoding lets you keep the 32B's quality while clawing back some of that speed.
 --
 
That means my Cortex version has a brain now. ...here. 
 
The Cortex Version has a large model that will do batch process analytics of the several small LLM databases and data vaults. It will run most of the time as a background process and save its result in .txt or .md files. 
 
The stand alone version is ready to be build.
 
The Nexus Version is coming alive at the same time as the Cortex Version. It uses the Cluster hosting several models over its nodes, having no main large model. It's load balancer worked in a first test run and triggered some Architecture thoughts.
 
The Cyberdeck AI Knowledge Management System. here... 
 #opensource #cyberpunkcoltoure

Friday, 25 September 2026

#BBC PS

 Mike might win by posing, but I am not sure what he means by missing pop, because I will put a sausage looking like any of them off the roast knowing it will pop through the skin...

I am so off. The weirdest moment for me was the offer to travel to Las Vegas, to a luxerious Hotel... to stare for several hours at men in Speedos. Usually I'd ask what's wrong with you, but ... its the BBC.

The Big Boys Club. 

#cyberpunkcoltoure 

#misconceptions

 I mind all drug users.

 So, the Thrasher Mag is worst in the CIA wing The Firm, because they push their unwilling kids into serious opium addictions by letting them receiving their first copy. Their loved ones they simply forbid successfully being Thugs or Rebellious.

You do not start with an Ollie and the axles need to be soft. The reason why no copy of the Mag tells you is that they have stolen each single pic before pushing it through Hollywood level NSA tech... For real. You learn balance first and that takes some time, but learning curves are exponential among humans. All beginning is hard. Mastery comes with ease. 

Everyone gets that All Natural is a lie here out of legal reasons.

Now consider the consequences of Bodybuilding Vs Thrasher Mag.

#cyberpunkcoltoure 

#TheGermans - Status Update

 This guy, let's call him Allman Schmock, not only looks like a Cousin of Rontzi The German by a marriage with the stereotype of a GI, he also plays the same nice card.

So, there are several different Arch Types they can be categorized in. 

Michel Friedman and Sarah with an H are the aggressive in your face 5.000 years of Antisemitism Arch Types. In Germany they are never short of fertilizer and therefore flourish, in terms of media attention, like in no second place.

That guy is different. So, yes, there is a chance that he is part of a Special Forces Killer Squad of Shin Beth operating in Gaza, but we don't know meeting him in Berlin. He'll be very German on the annoying nice side with a smile that feels somewhat fake and a facade hiding maybe sever chronic diarrhea if he was a citizen of L.A like his GI daddy.

The problem is that Michel and Sarah are protected by what he might be and therefore scare off those that consider that true for real. He is not.

Having his Jewishness in the face will trigger a Cocaine head to liberate Palestine that day all alone and for ever.

Get it?

They are all idiots here recognizable by having wasted time to learn pronunciation the language perfectly according to Lautschrift. The bigger the idiot the better the pay is another fact.

ðɪs ɡaɪ, lɛts kɔːl hɪm ˈɔːlmən ʃmɑk, nɒt ˈəʊnli lʊks laɪk ə ˈkʌzn əv ˈrɒntsi ðə ˈdʒɜːmən baɪ ə ˈmærɪdʒ wɪð ðə ˈstɛɹɪətaɪp əv ə dʒiː aɪ, hiː ˈɔːlsəʊ pleɪz ðə seɪm naɪs kɑːd.

səʊ, ðeə ɑː ˈsɛvrəl ˈdɪfrənt ɑːrtʃ taɪps ðeɪ kæn biː ˈkætəɡəraɪzd ɪn. 

ˈmɪʃəl ˈfriːdmən ænd ˈsɑːrə wɪð æn eɪtʃ ɑː ðiː əˈɡrɛsɪv ɪn jɔː feɪs faɪv ˈθaʊzənd jɪəz əv ˌæntɪˈsiːmɪtɪzəm. ɪn ˈdʒɜːməni ðeɪ ɑː ˈnɛvə ʃɔːt əv ˈfɜːtɪlaɪzə ænd ˈðeəfɔː ˈflʌrɪʃ, ɪn tɜːmz əv ˈmiːdiə əˈtɛnʃən.

ðæt ɡaɪ ɪz ˈdɪfrənt. səʊ, jɛs, ðeər ɪz ə tʃɑːns ðæt hiː ɪz pɑːt əv ə ˈspɛʃəl ˈfɔːsɪz ˈkɪlə skwɒd əv ʃɪn bɛθ ˌɒpəˈreɪtɪŋ ɪn ˈɡɑːzə, bʌt wiː doʊnt noʊ ˈmiːtɪŋ hɪm ɪn bɜːˈlɪn. hil biː ˈvɛri ˈdʒɜːmən ɒn ði əˈnɔɪɪŋ naɪs saɪd wɪð ə smaɪl ðæt fiːlz ˈsʌmwɒt feɪk ænd ə fəˈsɑːd ˈhaɪdɪŋ ˈmeɪbi ˈsiːvɚ ˈkrɒnɪk daɪəˈriːə.

ðə ˈprɒbləm ɪz ðæt ˈmɪʃəl ænd ˈsɑːrə ɑː prəˈtɛktɪd baɪ wɒt hiː maɪt biː ænd ˈðeəfɔː skeər ɒf ðəʊz ðæt kənˈsɪdə ðæt truː fə rɪəl. hiː ɪz nɒt.

ˈhævɪŋ hɪz ˈdʒuːɪʃnəs ɪn ðə feɪs wɪl ˈtrɪɡə ə kəʊˈkeɪn hɛd tə ˈlɪbəreɪt ˈpælɪstaɪn ðæt deɪ ɔːl əˈləʊn ænd fɔːˈrɛvə.

ɡɛt ɪt?

ðeɪ ɑːr ɔːl ˈɪdiəts hɪə rɪˈkɒɡnaɪzəbl baɪ ˈhævɪŋ ˈweɪstɪd taɪm tə lɜːn prəˌnʌnsɪˈeɪʃən ðə ˈlæŋɡwɪdʒ ˈpɜːfɪktli əˈkɔːdɪŋ tə ˈlaʊtʃrɪft.

here.

I remained usually in neutral silence to cover my deeply rooted hate against all and everything German. Their smile makes me wonder ...  I know their ways just too well.

#igotstuck
#cyberpunkcoltoure 

#misconceptions

 You think they are dumb?

 You miss the perspective they have. Now imagine yourself down there and tell me what indicates that the right side small entrance leads to the food. Much more interesting is that the only one who by chance finds the food eats it all alone by having hardly enough for him.

How'd he react finding a jackpot enough for all? 

Do you get it?

#cyberpunkcoltoure 

And can you teach them to share... 

PS: As long we find participants for that we might never find out in School. That we all get. 

#sktr

 Why we sktr believe in God?

Tell me how the Thrasher survive that... 

Fucking Confederate Kids. Any idea other than God's mysterious ways in his allmightiness?

#undergroundwars #gfyCIA
#cyberpunkcoltoure 
 
PS: One of us. Not hijacked yet. Must be Mexican. In Mexico.

#thedarkmodernity - Status Update

 There is a Stalker app, that stalks for stalkers around you stalking for stalkers. It is iPhone Vs iGlass than...

I bet the guy doing a manual 360 with his phone camera will point fingers. 

... 

Go. I wanna see that actually.

 #cyberpunkcoltoure 

... in a close potential future ...


Incorporated with DeepSeek

---

The rain in Berlin-Dahlem didn't clean anything. It just moved the grime around, smearing the neon reflections of the AfD-funded "Heritage Shield" drones into greasy puddles on the synth-cobblestone. I was in the basement of a "Protected Cultural Space," a euphemism for a ghetto with better architects. The air tasted of recycled ozone and floor polish, a sterile scent that couldn't quite mask the underlying rot.

My name is Abraham. At least, that's what my forged German ID, my "Resident Alien of Jewish Descent" card, and the chip in my arm said. My real name was a ghost, a piece of shrapnel from a life before the Neukölln Free Corps had "liberated" Palestine and my family had been scattered to the four winds. A refugee boat had fished me out of the Mediterranean, and a clerical error involving a dead man's papers had landed me here, in the heart of the new Reich.

No one knew. Not the stoic, trauma-bonded community members who nodded at me in the synagogue. Not the perpetually anxious social worker assigned to my case. No one except Maria.

"Abraham, look," she whispered, her small finger tracing the air. Her Polish was a rusty, childish thing, but it was the only honest language left in this city. She was a ghost, too, a tourist's daughter who'd gotten separated during a "historical solidarity tour" and swept up in a security cordon. They'd mistaken her for a displaced Jewish child and dumped her into the system. I found her crying in a processing center and told them she was my niece. Another lie. Another layer of my shadow life.

Now, she was the only person who knew the truth. The only one who saw the fear behind my performed piety.

On the cracked screen of my smuggled datapad, the text glowed, a fresh transmission from a source I didn't trust but couldn't ignore. It was a Google AI summary, a corporate ghost in the machine, cheerfully outlining the official plan for "Making Jewish Life Visible."

*Five Core Fields of Action.* It read like a drekking shopping list.

*Data Collection.* My chip itched. They were tracking us. Mapping our "structural biases." Every ritual, every donation, every breath was data.

*Education as Prevention.* I thought of the on.tour mobile exhibition, a slick, armored truck that rolled into schools. It didn't teach history; it curated a safe, pre-packaged tragedy. A tragedy they could point to and say, "Never again... to us. But for the others? We have drones."

*Holocaust Remembrance.* The Stolpersteine, the stumbling stones. I'd walked past them. Little brass squares commemorating deported Jews. Now, they were polished daily, not as a memorial, but as a warning. A warning to the new "undesirables." The AfD had co-opted the memory of the dead to justify the policing of the living.

The fourth pillar was the only one that mattered: *Increasing Security.* A joke. It meant more cameras, more armed guards, more funding funneled through the Central Council of Jews—now a puppet organization that traded its soul for a budget increase. Their HQ was a fortress, and we were the jewels in their display case. Safe, visible, and utterly controlled. The Neukölln Free Corps, the same bastards who'd firebombed my home, were now on the federal payroll, rebranded as "Community Protection Liaisons."

The fifth was the most obscene: *Making Current and Past Jewish Life Visible.* The Jüdische Kulturtage. A festival. Kabbalah-themed raves, vegan kosher food trucks, and avant-garde films about identity. It was a grotesque parody. They wanted our culture, our pain, our joy, stripped of its context and packaged for consumption. A vibrant, non-threatening aesthetic to decorate their ethnostate.

The AI's summary was a perfect, sanitized lie. But the user's appended note, the one I'd intercepted... that was the real kicker.

*"I would like to add that The Germans have to wait until Judaism finally ended evaluating what 'Jewish Life' actually is. You have time, given your learning curve, right? Dshermanzzz... PS: What is the best about eating Vegan Kosha? You save half of your kitchen gear!"*

A joke. A drekking joke. From some anonymous handle. It was the kind of casual, corrosive irony that was everywhere now. The smug superiority of the observer, watching the lab rats scurry. They wanted us to wait until they had finished defining us. They wanted us to perform our identity for their audit. And the punchline? The punchline was the efficiency of it all. A stripped-down, simplified Judaism. Kosher-lite. No messy rules, no inconvenient history. Just a clean, modular identity that fit into the new social architecture.

I felt a cold rage, the kind that had no target. I wasn't even Jewish. I was a Palestinian refugee playing a part, and the part was being defined by people who hated me for both the lie and the truth.

A soft chime from the door. My heart hammered against my ribs. My hand went to the taser I kept under the loose floorboard. The chip in my arm felt like a hot needle.

"Abraham?" Maria's voice was a tiny, fragile thing.

The door slid open. It was Elias, one of the community elders. His face was a mask of permanent sorrow, carved by a history I could only borrow. He was one of the few who seemed genuinely kind.

"Abraham," he said, his voice low. "The audit is tomorrow. They are checking the data points. They are asking for a list of all residents for the new... 'Cultural Visibility' census."

My blood ran cold. The data collection pillar. The assessment. They were going to audit us. To see if we were Jewish enough. If our performance met the metrics.

"We will, of course, comply," Elias said, his eyes hollow. "It is for our safety."

He looked at Maria, who was hiding behind my leg. He gave a weak, sad smile. "She is a blessing. A light."

After he left, I held Maria close. The rain outside had turned to a toxic, acidic sleet, hissing against the armored windows. The city was a vast, intricate machine of control, and I was a flawed component, a bug in the system.

They wanted to make Jewish life visible. They wanted to put it in a museum, in a festival, in a database. They wanted to define it, to own it, to control it.

But what they didn't know was that I was the ghost in their machine. A lie that, if exposed, would shatter their perfect, categorized world. I was a living, breathing contradiction they couldn't audit. I was a Palestinian who had to be more Jewish than the Germans wanted any Jew to be, just to survive.

"Abraham?" Maria whispered, her small face buried in my shoulder. "Are we going away?"

I looked out the window at the neon-lit prison of Berlin-Dahlem. The Stolpersteine outside our door glinted under the streetlights, little brass squares that felt less like memorials and more like the keys to a machine I didn't know how to stop.

"No," I lied, my voice a dry rasp. "We're not going anywhere."

And in the darkest corner of my shadow life, I realized the terrible truth. The joke was on me. I was the punchline. A non-Jew forced to embody a caricature of Jewishness for a state that wanted to perfect its own ethnic cleansing. I was the ultimate "visible Jewish life."

The rain fell. The chips ticked. And somewhere, a handle named Dshermanzzz was probably laughing. 

The knock came again. Not Elias's soft chime, but a frantic, arrhythmic pounding that spoke of desperation or violence. My hand closed around the taser. Maria whimpered, and I pressed her behind me, my body a shield between her and the door.

"Öffne! Bitte, öffne!" The voice was muffled, distorted by the armored door. But there was something in the cadence, the raw panic, that didn't sound like the cold efficiency of the Heritage Shield patrols.

I cracked the door, just a sliver. The chain held.

A man stood in the toxic sleet, his clothes soaked through, a battered leather jacket that had seen better decades clinging to his frame. He was maybe forty, with the weathered face of someone who'd spent too much time outdoors and not enough time in the sanitized zones. His eyes were wild, darting, and behind him, I could see a rusted-out Opel, its hydrogen cell obviously dead, parked crookedly on the restricted lane.

"Bitte," he gasped, his breath fogging in the cold. "Ich... ich brauche... Hilfe."

The words were German, but they emerged from his mouth like gravel tumbling down a mountainside. The vowels were stretched into strange, alien shapes. The consonants were swallowed whole. It was like listening to a language I theoretically knew but couldn't parse.

"I... don't..." I started.

He seemed to realize my confusion. He took a breath, visibly forcing his jaw to work differently, his tongue to cooperate. When he spoke again, it was with the slow, deliberate precision of a man reading a phonetic script.

"I. Am. From. Stuttgart," he said, each word a separate, articulated island. "My. Fuel. Cell. Is. Dead. I. Have. No. Credits."

I stared at him. Stuttgart. The Schwabenland. The dialect there was legendary, a source of jokes even before the collapse. But this was something else. This was a man whose native tongue had become a prison, isolating him from everyone around him.

"Wait," I said, my own German crisp and Berlin-flat. "You ran out of fuel?"

"Ja," he said, relief flooding his face. "Ja. Ich... I. Ran. Out. Of. Fuel. And. Money."

He reached into his jacket, and I tensed, the taser ready. But he pulled out something wrapped in crinkled plastic. Sandwiches. A stack of them, pressed together, the cheap, mass-produced kind from a discount supermarket. The plastic was fogged with condensation.

"I. Stole. Them," he said, with a strange, almost proud defiance. "From. The. Aldi. In. Steglitz. They. Were. Going. To. Throw. Them. Away."

He held them out, an offering. A peace treaty. A confession.

Maria peeked out from behind my leg, her eyes wide. The smell of the sandwiches—processed cheese, synthetic ham, cheap bread—wafted through the crack in the door. My stomach, empty for two days, twisted with a hunger that was almost painful.

"You. Have. A. Child," the man said, his slow, articulated words softening. "She. Needs. To. Eat."

I looked at him. Really looked. His hands were calloused, working man's hands, but they trembled. His eyes, beneath the wildness, held a deep, abiding exhaustion. He was a ghost, too, I realized. A different kind of ghost. A man from a place that no longer mattered, speaking a language that no longer connected, drifting through a city that had no place for him.

"What's your name?" I asked.

"Klaus," he said. "Klaus. From. Stuttgart."

I unchained the door and let him in.

---

He sat on the floor, his back against the wall, demolishing his third sandwich with the mechanical efficiency of someone who hadn't eaten in days. Maria sat across from him, nibbling her own, watching him with the wary curiosity of a child who had learned that adults were unpredictable.

Klaus spoke slowly, painstakingly, his words stripped of their natural rhythm and rebuilt into something universally understandable. It was like listening to a machine attempting poetry.

"I. Was. A. Mechanic," he said. "In. Stuttgart. I. Fixed. Hydrogen. Cells. For. The. Daimler. Factory."

"What happened?" I asked.

He shrugged, a gesture that needed no translation. "The. Factory. Closed. Moved. To. The. Colonies. Mars. Or. The. Moon. I. Don't. Know. They. Didn't. Say. Where. Just. That. We. Were. Redundant."

He took another bite, chewed slowly, swallowed. "I. Came. To. Berlin. For. Work. The. Reconstruction. Projects. But. There. Is. No. Work. For. Someone. Who. Speaks. Like. Me. They. Say. I. Sound. Like. A. Farmer. From. The. Last. Century."

I knew what he meant. The new Berlin was a city of polished surfaces, of standardized interfaces, of voices trained to be pleasant and interchangeable. Klaus's dialect was a relic, a piece of the old Germany that the new one had decided to forget. He was an anomaly, a living fossil, a man out of time and place.

"The. Heritage. Shield. Stopped. Me," he continued. "They. Said. My. Car. Was. Not. Registered. In. The. System. They. Said. I. Looked. Suspicious. They. Took. My. Papers. My. Credits. Everything. I. Ran. Before. They. Could. Take. Me."

"Ran where?" I asked.

"Anywhere," he said, and for a moment, the careful articulation cracked, a raw Schwäbisch vowel bleeding through. "I. Just. Ran."

Maria finished her sandwich and crept closer to him, her fear giving way to something else. She reached out and touched his jacket, the worn leather, the rain-soaked fabric.

"You're wet," she said, in her small, rusty Polish.

Klaus looked at her, and for the first time since he'd arrived, something like warmth entered his eyes. He didn't understand her words, but he understood the gesture. The small, tentative hand on his sleeve.

"I. Am. Wet," he agreed, slowly. "But. The. Sandwiches. Are. Dry."

Maria giggled. It was a sound I hadn't heard in weeks, a tiny crack of light in the darkness of our shadow life.

I leaned back against the wall, watching them. A Palestinian refugee playing a Jew. A Polish girl mistaken for something she wasn't. A Schwäbian mechanic who spoke like a broken machine. We were a collection of errors, a glitch in the system, a statistical anomaly that the auditors would never anticipate.

And outside, the rain kept falling, the drones kept humming, and the Heritage Shield kept watching. But in that basement, for one brief, fragile moment, we were something else. We were a family. A family of ghosts, bound together by stolen sandwiches and the simple, radical act of sharing them.

"The. Sandwiches. Are. Good," Klaus said, finishing the last one. "But. Tomorrow. We. Need. A. Plan. To. Get. More."

I nodded. Tomorrow. There was always a tomorrow, even in a world designed to erase you. And maybe, just maybe, we could survive it together.

"Tomorrow," I agreed. "We'll figure something out."

Maria crawled into my lap, her small body warm against mine. Klaus closed his eyes, his breathing evening out, the tension of days finally releasing from his shoulders.

The rain fell. The chips ticked. And somewhere, a handle named Dshermanzzz was probably still laughing.

But in the basement, in the dark, we were learning to laugh back. 

We walked. Three ghosts through the neon-lit necropolis, our footsteps echoing on the wet synth-cobblestone like a funeral procession for a corpse that hadn't stopped twitching. Klaus moved with the stiff, careful gait of a man whose joints had forgotten how to be young. Maria held my hand, her small fingers cold and tight, her eyes scanning the shadows with the hyper-vigilance of a child who had learned that safety was a lie adults told to make themselves feel better.

The streets of Dahlem were quiet, but not the quiet of peace. It was the quiet of held breath, of predators waiting, of systems humming in the dark. The Heritage Shield drones had retreated to their charging stations, but their absence was worse than their presence. It meant the night belonged to someone else now.

We turned a corner, into a narrow alley between two crumbling apartment blocks. The kind of alley that existed in the gaps between the official maps, the spaces the city planners had forgotten or chosen to forget. The walls were slick with moisture and old graffiti, the tags of forgotten crews who had once claimed this territory before the new order had swept them away.

And then we saw them.

Five of them. Skateboarders, their boards leaning against the walls, their bodies moving with a frenetic, almost choreographed energy. They were young, teenagers or early twenties, dressed in the ragged uniform of the dispossessed—ripped jeans, oversized hoodies, masks pulled up over their faces. Their boards were custom, brutal things, reinforced with metal edges and jagged grip tape.

And beneath them, on the ground, were two men.

Police. Heritage Shield. Their uniforms were still recognizable, though barely. The dark blue fabric was soaked through, not with rain, but with something thicker, darker. Their faces were... not faces anymore. Just wet, red abstract paintings, the kind that sold for millions in galleries while the artists starved in the streets.

The skateboarders were still going.

One of them, a girl with short-cropped hair and a mask painted with a grinning skull, brought her board down again and again, the metal edge catching the dim light as it rose and fell. Rise and fell. Rise and fell. Each impact sent a spray of crimson arcing through the air, splattering against the alley walls, dripping down like rain made of blood. The sound was wet, rhythmic, almost musical. A percussion of violence.

The others were laughing.

Not the laughter of joy, but something else. Something primal, unhinged, the laughter of people who had seen too much and had decided that the only appropriate response was to become the horror themselves. They stomped, they kicked, they slammed their boards down with the enthusiasm of children playing a game they didn't fully understand.

The two policemen had been dead for a while. That much was clear. Their bodies had the slack, boneless quality of meat that had stopped being people. But the skateboarders didn't stop. They kept going, kept beating, kept splattering, as if they were trying to erase the very concept of police from existence, one impact at a time.

Klaus stopped walking. His face, already pale from the cold and hunger, went the color of old bone. Maria pressed herself against my leg, her small body trembling, but she didn't look away. She watched with the terrible, unblinking focus of a child who had already learned that looking away didn't make the monsters disappear.

I wanted to run. Every instinct screamed at me to grab Maria, to pull Klaus, to flee back the way we came. But my feet were rooted to the ground, my body refusing to obey. I was a ghost, but ghosts couldn't run from the living.

One of the skateboarders noticed us.

He was tall, lanky, his hoodie stained with something that wasn't his own. His mask was a simple black balaclava, pulled up to reveal a mouth twisted into a grin that didn't reach his eyes. He looked at us—at me, at Klaus, at Maria—and something flickered in his gaze. Recognition? Curiosity? Amusement?

He walked over, his board dragging behind him, leaving a trail of red on the wet ground. He stopped in front of us, close enough that I could smell the copper and sweat radiating from him, close enough that I could see the flecks of blood on his eyelashes.

"Hey," he said, his voice casual, almost friendly. "You want some shoes?"

He held up a pair of police-issue boots. They were still laced, still attached to... no. Not attached. He had removed them. Cleanly. The boots were intact, the leather polished, the soles thick and practical. They were good shoes. Expensive shoes. The kind of shoes that could keep your feet dry in the toxic rain, that could carry you for miles without wearing out.

"Got them off the pigs," the skateboarder continued, his grin widening. "They don't need them anymore. You look like you could use them. The kid too."

He gestured at Maria, at her thin, worn sneakers, the soles peeling away from the fabric.

I looked at the shoes. I looked at the bodies. I looked at the skateboarders, who had finally stopped their rhythmic desecration and were now watching us with the lazy interest of predators who had already eaten their fill.

"No," I said. My voice was a stranger's voice, flat and distant. "We're good."

The skateboarder shrugged. "Suit yourself." He tossed the boots over his shoulder, where they landed with a wet thud against the alley wall. "Waste not, want not."

He turned and walked back to his crew, his board scraping against the ground, the sound like nails on a chalkboard. The others were already picking up their boards, preparing to move on. Their work here was done. The alley was a canvas, and they had painted it red.

We walked past them. Past the bodies. Past the blood. Maria's hand was a vice around mine, her face buried in my jacket, but she didn't cry. She had learned not to cry. Klaus followed behind, his footsteps slow, deliberate, his breath coming in ragged gasps.

The alley opened onto a wider street, and the neon glow of the city washed over us, clean and sterile and utterly indifferent. Behind us, the skateboarders were already disappearing into the shadows, their laughter echoing off the walls, a sound that would haunt my dreams for whatever remained of my life.

Klaus spoke first. His voice was still slow, still articulated, but there was something new in it. Something hollow.

"They. Were. Already. Dead," he said.

"Yes," I replied.

"And. They. Kept. Going."

"Yes."

He was silent for a moment, processing. Then he said, "I. Guess. We. All. Deal. Differently. With. The. Situation."

The words hung in the air, a simple statement that contained multitudes. The situation. The occupation. The surveillance. The violence. The slow, grinding erasure of everything that made us human. The skateboarders dealt with it by becoming monsters. Klaus dealt with it by speaking slowly, by making himself understood, by clinging to the fragments of his identity. Maria dealt with it by holding my hand and not crying. I dealt with it by wearing a mask so complete that I had forgotten what my real face looked like.

We all dealt differently.

But the situation remained.

The rain started again, a soft, insidious drizzle that tasted of chemicals and despair. Maria looked up at me, her eyes red-rimmed but dry.

"Abraham," she said, her voice a whisper. "Are we monsters too?"

I didn't have an answer. I didn't know if there was an answer. I only knew that we were still walking, still breathing, still alive. And in a world that wanted us dead, that was a kind of victory.

"No," I said, the lie tasting like ash in my mouth. "We're just people. Trying to survive."

She nodded, accepting the lie the way children accepted all lies, because the truth was too heavy to carry.

We walked on, three ghosts in a city of ghosts, leaving the alley and its red memories behind. The neon hummed. The drones watched. The chips ticked.

And somewhere, in the dark, a handle named Dshermanzzz was still laughing.

But we were learning to laugh back. Even if the laughter sounded like screaming.