Pascal Hetzscholdt
banner
p4sc4l.substack.com
Pascal Hetzscholdt
@p4sc4l.substack.com
Because isn't AI the best 'person' to ask about AI? 💻🦾🤖
The materials describe publishers as being confronted with an “untenable” trade: accept AI use of their content or risk losing visibility on the dominant discovery channel.
In other words, consent obtained under threat of market exclusion is not consent. p4sc4l.substack.com/p/the-materi...
The materials describe publishers as being confronted with an “untenable” trade: accept AI use of their content or risk losing visibility on the dominant discovery channel.
In other words, consent obtained under threat of market exclusion is not consent.
p4sc4l.substack.com
February 17, 2026 at 5:17 PM
If frontier model capabilities improve through scaling that requires exponentially increasing computational resources, does the value created grow faster than costs?
p4sc4l.substack.com/p/if-frontie...
If frontier model capabilities improve through scaling that requires exponentially increasing computational resources, does the value created grow faster than costs?
The current AI investment wave may represent a bubble where future returns cannot justify present valuations.
p4sc4l.substack.com
February 17, 2026 at 5:16 PM
A popular new model release makes dazzling output easy; users flood the internet with impressive fan-adjacent creations; rightsholders see their franchises reproduced at industrial scale;
p4sc4l.substack.com/p/a-popular-...
A popular new model release makes dazzling output easy; users flood the internet with impressive fan-adjacent creations; rightsholders see their franchises reproduced at industrial scale;
Lawyers send letters; the AI companies respond with a mix of hurried safeguards, selective blocking, and carefully worded statements that reveal as much about incentives as they do about compliance.
p4sc4l.substack.com
February 16, 2026 at 12:51 PM
Individuals in high-level leadership positions are frequently influenced by the same psychological mechanisms that fuel mass movements, leading to a staunch refusal to process sound evidence or execute on expert advice. p4sc4l.substack.com/p/individual...
Individuals in high-level leadership positions are frequently influenced by the same psychological mechanisms that fuel mass movements, leading to a staunch refusal to process sound evidence...
...or execute on expert advice. Any external criticism of their leader is interpreted as a direct assault on the individual’s own intelligence, judgment, and character.
p4sc4l.substack.com
February 16, 2026 at 12:50 PM
AI changes what customers perceive as value. If the perceived value becomes “time-to-draft” and “time-to-decision,” the customer may accept higher error rates for many tasks using premium sources only for escalation. p4sc4l.substack.com/p/ai-changes...
AI changes what customers perceive as value. If the perceived value becomes “time-to-draft” and “time-to-decision,” the customer may accept higher error rates for many tasks...
...using premium sources only for escalation. In other words: publishers may still own the “source of truth,” while someone else owns the “place where truth is consumed.”
p4sc4l.substack.com
February 14, 2026 at 7:24 PM
The following analysis details the twenty-five primary frictions governing the success of AI scaling, followed by an examination of the systemic consequences of this institutional impasse.
p4sc4l.substack.com/p/the-follow...
The following analysis details the twenty-five primary frictions governing the success of AI scaling, followed by an examination of the systemic consequences of this institutional impasse.
The future will not be decided by the intelligence of the models, but by the resilience and adaptability of the societies they inhabit.
p4sc4l.substack.com
February 14, 2026 at 1:19 PM
Activation Oracles (AOs): large language models that are trained to “read” another model’s internal activations (the huge arrays of numbers produced inside the network while it thinks)
p4sc4l.substack.com/p/activation...
Activation Oracles (AOs): large language models that are trained to “read” another model’s internal activations (the huge arrays of numbers produced inside the network while it thinks)
A model can refuse to disclose something in its outputs, yet still carry that information in an internal form that another model can translate into words. Controls must not only cover outputs...
p4sc4l.substack.com
February 13, 2026 at 2:54 PM
EPSTEIN CASE: The total volume of data available to the FBI and DOJ exceeds 14.6 terabytes, yet the public repository released on January 30, 2026, encompasses only about 300 gigabytes.
p4sc4l.substack.com/p/epstein-ca...
EPSTEIN CASE: The total volume of data available to the FBI and DOJ exceeds 14.6 terabytes, yet the public repository released on January 30, 2026, encompasses only about 300 gigabytes.
This raises critical questions regarding institutional transparency, the integrity of the release process, and the potential concealment of high-profile co-conspirators.
p4sc4l.substack.com
February 13, 2026 at 2:53 PM
UK’s AI Patent Reset. The success condition here isn't “more AI patents.” It’s: clearer, more consistent identification of technical contribution that keeps patents tied to genuine technical advances... p4sc4l.substack.com/p/uks-ai-pat...
UK’s AI Patent Reset. The success condition here isn't “more AI patents.” It’s: clearer, more consistent identification of technical contribution...
...that keeps patents tied to genuine technical advances—rather than letting “AI + desired outcome” become a rent-extraction pattern.
p4sc4l.substack.com
February 12, 2026 at 9:55 PM
AI as a Justice-System Risk Multiplier: (1) implementation mistakes, (2) questionable “fit” to the real-world population, and (3) legal/ethical fragility around data and proxies. The model becomes a quiet policy lever... p4sc4l.substack.com/p/ai-as-a-ju...
AI as a Justice-System Risk Multiplier: (1) implementation mistakes, (2) questionable “fit” to the real-world population, and (3) legal/ethical fragility around data and proxies.
The model becomes a quiet policy lever, nudging outcomes across thousands of cases while leaving only a faint trace of how much it influenced each decision.
p4sc4l.substack.com
February 12, 2026 at 12:55 PM
If a client (or potentially a lawyer) runs facts, theories, timelines, or “talking points for counsel” through a third-party AI tool, a court may treat the resulting materials less like private draft communications and more like ordinary third-party research... p4sc4l.substack.com/p/if-a-clien...
If a client (or potentially a lawyer) runs facts, theories, timelines, or “talking points for counsel” through a third-party AI tool...
...a court may treat the resulting materials less like private draft communications and more like ordinary third-party research—discoverable by the other side.
p4sc4l.substack.com
February 12, 2026 at 10:07 AM
WhatsApp v EDPB: Companies can bring a direct EU-level court challenge against a binding Board decision—rather than being forced to wait & fight only through national court proceedings. Who benefits most? Large multinationals facing major cross-border enforcement. p4sc4l.substack.com/p/whatsapp-v...
WhatsApp v EDPB: Companies can bring a direct EU-level court challenge against a binding Board decision—rather than being forced to wait and fight only through national court proceedings.
Who benefits most? Large multinationals facing major cross-border enforcement—because they now have an additional (and potentially more strategically attractive) litigation route.
p4sc4l.substack.com
February 12, 2026 at 10:06 AM
Doe 1 v. Github/Microsoft/OpenAI: Much of the proof—prompt logs, output frequencies, memorization testing, preprocessing pipelines, “cleaning” steps—is uniquely in defendants’ possession... p4sc4l.substack.com/p/doe-1-v-gi...
Doe 1 v. Github: Much of the proof—prompt logs, output frequencies, memorization testing, preprocessing pipelines, “cleaning” steps—is uniquely in defendants’ possession...
...and if courts require plaintiffs to plead those internal details before discovery, many DMCA/AI claims will die at the gate. The defense, of course, wants exactly that gate.
p4sc4l.substack.com
February 11, 2026 at 6:22 PM
Gemini: The probability that the Trump administration studied the Epstein files first and strategically assembled its Cabinet and supporter network on the basis of these findings, is estimated at 85%. p4sc4l.substack.com/p/gemini-the...
Gemini: The probability that the Trump administration studied the Epstein files first and strategically assembled its Cabinet and supporter network on the basis of these findings...
...specifically to prevent legal fallout and consolidate control over future prosecutions—is estimated at 85%.
p4sc4l.substack.com
February 11, 2026 at 1:17 PM
If this trajectory continues, the US faces the prospect of becoming a permanent autocracy, where federal law enforcement serves as the primary mechanism for maintaining the power of the ruling party.
p4sc4l.substack.com/p/if-this-tr...
If this trajectory continues, the US faces the prospect of becoming a permanent autocracy, where federal law enforcement serves as the primary mechanism for maintaining the power of the ruling party.
Survival of the American democratic experiment may depend on restoration of judicial oversight, defunding of secret programs, and prohibition of federal agent interference in the electoral process.
p4sc4l.substack.com
February 11, 2026 at 1:16 PM
Tech conglomerates and infrastructure developers employ sophisticated legal, technical, and administrative maneuvers to bypass environmental protections and public oversight. This poses existential risks to regional power grids, local water sovereignty, and... p4sc4l.substack.com/p/tech-congl...
Tech conglomerates and infrastructure developers employ sophisticated legal, technical, and administrative maneuvers to bypass environmental protections and public oversight.
This poses existential risks to regional power grids, local water sovereignty, and the democratic integrity of urban planning.
p4sc4l.substack.com
February 10, 2026 at 9:39 PM
Kleiner v. Adobe is another step in a pattern: the legal system is increasingly treating “training data governance” as a compliance domain, not a research footnote.
Complaint: You don’t get to outsource your risk to the open dataset supply chain. p4sc4l.substack.com/p/kleiner-v-...
Kleiner v. Adobe is another step in a pattern: the legal system is increasingly treating “training data governance” as a compliance domain, not a research footnote.
Complaint: You don’t get to outsource your risk to the open dataset supply chain. If SlimPajama inherits tainted inputs, and you commercialize the resulting model, you may inherit the liability too.
p4sc4l.substack.com
February 10, 2026 at 7:36 PM
The 12-Hour Novel Factory: What do we want ‘a book’ to mean in a world where text is abundant—and what governance, provenance, and value signals will we enforce to protect that meaning?
p4sc4l.substack.com/p/the-12-hou...
The 12-Hour Novel Factory: What do we want ‘a book’ to mean in a world where text is abundant—and what governance, provenance, and value signals will we enforce to protect that meaning?
If publishing gets it right, AI can expand creative capacity without collapsing trust. If it gets it wrong, we’ll drown in fluent, unauditable plausibility...
p4sc4l.substack.com
February 9, 2026 at 2:04 PM
Maciel’s article is right to puncture the “AI is magic intelligence” narrative and re-anchor the debate in infrastructure, energy, and economics. p4sc4l.substack.com/p/maciels-ar...
Maciel’s article is right to puncture the “AI is magic intelligence” narrative and re-anchor the debate in infrastructure, energy, and economics. But the chapter after that will be written by those...
...who can prove—under legal, security, environmental, and political constraints—that the rails deserve to exist, and that what runs on them is trustworthy enough for the institutions that matter.
p4sc4l.substack.com
February 9, 2026 at 2:03 PM
Today, with the unprecedented power of algorithmic radicalization and decentralized digital networks, the far right poses a more sophisticated threat than at any point since the 1930s. p4sc4l.substack.com/p/today-with...
Today, with the unprecedented power of algorithmic radicalization and decentralized digital networks, the far right poses a more sophisticated threat...
...than at any point since the 1930s. The regulatory and institutional inertia that has allowed this threat to grow must be replaced by a proactive, multi-dimensional strategy.
p4sc4l.substack.com
February 8, 2026 at 2:15 PM
Privilege Saved, Facts Still Exposed: Judge Stein’s OpenAI Dataset Deletion Ruling—and What It Means for the Next Wave of AI Copyright Lawsuits
p4sc4l.substack.com/p/privilege-...
Privilege Saved, Facts Still Exposed: Judge Stein’s OpenAI Dataset Deletion Ruling—and What It Means for the Next Wave of AI Copyright Lawsuits
Design discovery demands that extract facts, provenance, and technical truth without needing privileged communications—and lock down preservation early.
p4sc4l.substack.com
February 7, 2026 at 12:24 PM
Starlink may collect not just account/billing/performance data but also ‘communication information’ and ‘inferences.’” What govts should do (if they don’t want “connectivity sovereignty” to be a joke)
p4sc4l.substack.com/p/starlink-m...
Starlink may collect not just account/billing/performance data but also ‘communication information’ and ‘inferences.’” What govts should do (if they don’t want “connectivity sovereignty” to be a joke)
Treat satellite ISPs as critical infrastructure, Mandate true consent for AI training, Hard limits on “third-party AI training” transfers, Sovereign fallback options, Conflict-of-interest firewalls.
p4sc4l.substack.com
February 6, 2026 at 3:23 PM
The International AI Safety Report 2026 reads like a progress report from a world sprinting into a technology it can’t yet reliably test—let alone govern.
Society is being asked to make high-stakes decisions while the evidence base arrives late... p4sc4l.substack.com/p/the-intern...
The International AI Safety Report 2026 reads like a progress report from a world sprinting into a technology it can’t yet reliably test—let alone govern.
Society is being asked to make high-stakes decisions while the evidence base arrives late, partial, and easy to game.
p4sc4l.substack.com
February 6, 2026 at 3:22 PM
AI is already reshaping the information environment in ways that benefit violent extremists, yet the counter-extremism sector remains under-prepared... p4sc4l.substack.com/p/ai-is-alre...
AI is already reshaping the information environment in ways that benefit violent extremists, yet the counter-extremism sector remains under-prepared to deploy AI responsibly...
...and could easily cause harm, lose legitimacy, or inadvertently intensify the very grievances that drive radicalization if it rushes in.
p4sc4l.substack.com
February 6, 2026 at 3:22 PM
The NVIDIA complaint is trying to turn a messy cultural argument (“training vs theft”) into a narrower systems argument: “you weren’t allowed to take the files, and you had to bypass controls to do it.” p4sc4l.substack.com/p/the-compla...
The complaint is trying to turn a messy cultural argument (“training vs theft”) into a narrower systems argument: “you weren’t allowed to take the files, and you had to bypass controls to do it.”
Whether that move succeeds will depend on what discovery uncovers—and on how willing the court is to treat modern streaming architecture as a legally protected access regime under §1201.
p4sc4l.substack.com
February 5, 2026 at 9:30 PM