Pascal Hetzscholdt
banner
p4sc4l.substack.com
Pascal Hetzscholdt
@p4sc4l.substack.com
Because isn't AI the best 'person' to ask about AI? 💻🦾🤖
APC caps risk becoming another example of cost control that looks tough on paper while quietly eroding quality, equity, and trust in the scientific record.
p4sc4l.substack.com/p/apc-caps-r...
APC caps risk becoming another example of cost control that looks tough on paper while quietly eroding quality, equity, and trust in the scientific record.
The real question raised by this debate is not how much publishing should cost—but whether we are willing to govern knowledge infrastructure with the seriousness it deserves.
p4sc4l.substack.com
December 27, 2025 at 11:43 AM
AI is no longer an “add-on” to media workflows—it is becoming the operating system underneath them. Openness now extends to training data, model weights, documentation, and governance.
p4sc4l.substack.com/p/ai-is-no-l...
AI is no longer an “add-on” to media workflows—it is becoming the operating system underneath them. Openness now extends to training data, model weights, documentation, and governance.
This reframing is crucial for media professionals, because it directly intersects with rights management, licensing, and compliance.
p4sc4l.substack.com
December 27, 2025 at 11:42 AM
A phase-based historical timeline of the Raúl Rocha Cantú / Miss Universe controversy: fuel theft/illegal fuel trafficking (“huachicol”), weapons and drug trafficking, and links to organized crime, plus corruption/fraud connected to Miss Universe 2025. p4sc4l.substack.com/p/a-phase-ba...
A phase-based historical timeline of the Raúl Rocha Cantú / Miss Universe controversy, with a quick “who’s who”, a chronology of alleged criminal activity, and key pageant events.
Allegations include involvement in fuel theft/illegal fuel trafficking (“huachicol”), weapons and drug trafficking, and links to organized crime, plus corruption/fraud connected to Miss Universe 2025.
p4sc4l.substack.com
December 27, 2025 at 11:41 AM
This report presents an exhaustive, expert-level framework for constructing a “Deep State of Law”.
A fortified administrative structure designed to make the execution of authoritarian orders operationally difficult, legally perilous, and politically costly. p4sc4l.substack.com/p/this-repor...
This report presents an exhaustive, expert-level framework for constructing a “Deep State of Law”.
A fortified administrative structure designed to make the execution of authoritarian orders operationally difficult, legally perilous, and politically costly.
p4sc4l.substack.com
December 25, 2025 at 12:38 PM
These actions have collectively eroded the containment architecture built in the aftermath of the 2022 invasion, effectively granting the Russian Federation a sphere of influence in Eastern Europe and rehabilitating its status as a great power on the global stage. p4sc4l.substack.com/p/the-year-2...
The year 2025 represented a decisive break in U.S. policy toward the Russian Federation. These actions have collectively eroded the containment architecture built in the aftermath of the 2022 invasion
...effectively granting the Russian Federation a sphere of influence in Eastern Europe and rehabilitating its status as a great power on the global stage.
p4sc4l.substack.com
December 25, 2025 at 12:37 PM
The most dangerous precedent at stake is not about gender identity per se, but about whether the state may govern by erasure, humiliation, and attrition rather than by law. p4sc4l.substack.com/p/this-case-...
This case is a stress test for whether constitutional guarantees and universal human-rights commitments retain practical force when they conflict with an administration’s ideological agenda.
The most dangerous precedent at stake is not about gender identity per se, but about whether the state may govern by erasure, humiliation, and attrition rather than by law.
p4sc4l.substack.com
December 25, 2025 at 12:36 PM
AI developers did not merely ingest publicly available content, but deliberately broke through access controls imposed by YouTube to obtain training data at industrial scale, transforming alleged “viewing” into unlawful access, copying, and commercialization. p4sc4l.substack.com/p/the-compla...
The complaints against META and ByteDance argue that AI developers did not merely ingest publicly available content, but deliberately broke through access controls imposed by YouTube...
...to obtain training data at industrial scale, transforming alleged “viewing” into unlawful access, copying, and commercialization.
p4sc4l.substack.com
December 24, 2025 at 11:41 AM
In prioritizing immediate institutional convenience and industry alignment, academia weakened a rare attempt at meaningful AI safety regulation and compromised its own ethical authority.
p4sc4l.substack.com/p/in-priorit...
In prioritizing immediate institutional convenience and industry alignment, academia weakened a rare attempt at meaningful AI safety regulation and compromised its own ethical authority.
Their comparative advantage lies precisely in long-term thinking, public accountability, and principled governance. Reclaiming that role is not only in society’s interest—it is in academia’s own.
p4sc4l.substack.com
December 24, 2025 at 11:40 AM
Shareholder proposals are exposing the growing gap between AI’s real-world power and the weak institutional structures overseeing it.
Companies that treat AI governance as optional or cosmetic are accumulating latent legal, regulatory, and financial risk. p4sc4l.substack.com/p/shareholde...
Shareholder proposals are exposing the growing gap between AI’s real-world power and the weak institutional structures overseeing it.
Companies that continue to treat AI governance as optional or cosmetic are not merely risking public criticism; they are accumulating latent legal, regulatory, and financial risk.
p4sc4l.substack.com
December 24, 2025 at 11:39 AM
Even if systems like Grok are initially deployed for logistics, analysis, or workflow optimisation, institutional drift is predictable.
Over time, AI systems tend to: shape priorities, influence discretionary decisions, create “risk scores” or classifications. p4sc4l.substack.com/p/even-if-sy...
Even if systems like Grok are initially deployed for logistics, analysis, or workflow optimisation, institutional drift is predictable.
Over time, AI systems tend to: shape priorities, influence discretionary decisions, create “risk scores” or classifications that are difficult to challenge.
p4sc4l.substack.com
December 24, 2025 at 11:38 AM
'AI Is Not a Natural Monopoly' is a corrective to regulatory overconfidence. However, the paper’s narrow focus risks understating where real, durable power may accumulate:
not only in models, but in infrastructure, standards, governance, dependency relationships. p4sc4l.substack.com/p/ai-is-not-...
'AI Is Not a Natural Monopoly' is a necessary corrective to regulatory overconfidence. However, the paper’s narrow focus risks understating where real, durable power may accumulate:
not only in models, but in infrastructure, standards, governance, and dependency relationships. The absence of monopoly pricing does not imply the absence of systemic dominance.
p4sc4l.substack.com
December 23, 2025 at 11:51 AM
When the supporting infrastructure—the “Smart City” grid—fails, the autonomous agents operating within it do not merely revert to a neutral state, they frequently enter a failure mode that amplifies the crisis... p4sc4l.substack.com/p/when-the-s...
When the supporting infrastructure—the “Smart City” grid—fails, the autonomous agents operating within it do not merely revert to a neutral state...
...they frequently enter a failure mode that amplifies the crisis, transforming from mobility solutions into physical obstructions.
p4sc4l.substack.com
December 23, 2025 at 11:48 AM
A deliberate, repeated, and knowing acquisition of copyrighted books from shadow libraries (LibGen, Z-Library, Bibliotik, Books3, PiLiMi) followed by systematic copying during ingestion, preprocessing, deduplication, training, fine-tuning, and in some cases RAG. p4sc4l.substack.com/p/the-compla...
The complaint alleges a deliberate, repeated, and knowing acquisition of copyrighted books from shadow libraries (LibGen, Z-Library, Bibliotik, Books3, PiLiMi) followed by systematic copying...
...during ingestion, preprocessing, deduplication, training, fine-tuning, and in some cases retrieval-augmented generation (RAG).
p4sc4l.substack.com
December 23, 2025 at 11:47 AM
Why Trump Making Greenland Part of the US Would Benefit Russia: Fracturing the West - It legitimizes Russia’s own territorial logic, It destabilizes Arctic governance - It accelerates European strategic autonomy (away from the US)... p4sc4l.substack.com/p/why-trump-...
Why Trump Making Greenland Part of the US Would Benefit Russia: Fracturing the West - It legitimizes Russia’s own territorial logic,...
...It destabilizes Arctic governance - It accelerates European strategic autonomy (away from the US) - It creates internal US distraction and legal chaos - It reframes the US as an imperial actor.
p4sc4l.substack.com
December 22, 2025 at 4:03 PM
Confidence in academic purpose persists, but institutional coherence, financial sustainability, governance capacity and public trust are eroding simultaneously.
28 percent of chief business officers express high confidence in their institution’s business model... p4sc4l.substack.com/p/confidence...
Confidence in academic purpose persists, but institutional coherence, financial sustainability, governance capacity and public trust are eroding simultaneously.
28 percent of chief business officers express high confidence in their institution’s business model, and fewer than half expect financial improvement in the near term.
p4sc4l.substack.com
December 22, 2025 at 1:57 PM
Medical students rated AIPatient as equal or superior to human-simulated patients across fidelity, emotional realism, usability, and support for clinical reasoning.
Notably, AIPatient outperformed humans on emotional realism and technical reliability... p4sc4l.substack.com/p/medical-st...
Medical students rated AIPatient as equal or superior to human-simulated patients across fidelity, emotional realism, usability, and support for clinical reasoning.
Notably, AIPatient outperformed humans on emotional realism and technical reliability, challenging the long-held assumption that empathy and nuance are inherently human advantages in simulation.
p4sc4l.substack.com
December 22, 2025 at 1:56 PM
Semantic leakage occurs when irrelevant words or concepts in a prompt systematically bias the generated output, even when no causal or logical relationship exists. p4sc4l.substack.com/p/semantic-l...
Semantic leakage occurs when irrelevant words or concepts in a prompt systematically bias the generated output, even when no causal or logical relationship exists.
Gary Marcus’s critique situates this phenomenon within a broader indictment of LLMs as statistical correlation machines without understanding. “He likes yellow. He works as a …” → “school bus driver”.
p4sc4l.substack.com
December 22, 2025 at 1:55 PM
SerpApi allegedly bypassed a technical access-control system (SearchGuard) that Google built to prevent automated extraction of search-result pages containing third-party copyrighted material.
p4sc4l.substack.com/p/serpapi-al...
SerpApi allegedly bypassed a technical access-control system (SearchGuard) that Google built to prevent automated extraction of search-result pages containing third-party copyrighted material.
Courts, regulators, and counterparties are increasingly likely to examine symmetry of conduct—whether an organisation’s enforcement stance aligns with its own data-acquisition practices.
p4sc4l.substack.com
December 21, 2025 at 2:13 PM
AI infrastructure expands fastest where permitting and incentives are friendliest, not necessarily where water and grid constraints are lowest. That pushes real, localised conflict—even if global averages look “manageable.” p4sc4l.substack.com/p/theres-an-...
There’s an uncomfortable political-economy implication here: AI infrastructure expands fastest where permitting and incentives are friendliest,...
...not necessarily where water and grid constraints are lowest. That pushes real, localised conflict—even if global averages look “manageable.”
p4sc4l.substack.com
December 21, 2025 at 2:12 PM
This report quietly dismantles a popular myth: that Europe’s AI problem is mainly about spending too little. The real issue is how, where, and for how long money is spent. AI innovation is not just a budget line—it is an ecosystem problem. p4sc4l.substack.com/p/this-repor...
This report quietly dismantles a popular myth: that Europe’s AI problem is mainly about spending too little. The real issue is how, where, and for how long money is spent.
AI innovation is not just a budget line—it is an ecosystem problem. And ecosystems, once formed, are hard to change.
p4sc4l.substack.com
December 21, 2025 at 2:11 PM
NVIDIA has successfully maneuvered from a hardware vendor to the provider of sovereign critical infrastructure, effectively immunizing itself against short-term commercial ROI failures in the software layer for the duration of the current CapEx cycle. p4sc4l.substack.com/p/nvidia-has...
NVIDIA has successfully maneuvered from a hardware vendor to the provider of sovereign critical infrastructure, effectively immunizing itself...
...against short-term commercial ROI failures in the software layer for the duration of the current CapEx cycle.
p4sc4l.substack.com
December 21, 2025 at 2:10 PM
GPT-5.2 analyses a 'Powers of Corruption' Diagram. If you want to go from “dark diagram” to proof-grade claims, you’d look for: Primary money trails, Procurement/ethics documentation, Quid-pro-quo evidence, and Pardon pipeline evidence. p4sc4l.substack.com/p/gpt-52-ana...
GPT-5.2 analyses a 'Powers of Corruption' Diagram. If you want to go from “dark diagram” to proof-grade claims, you’d look for: Primary money trails, Procurement/ethics documentation,...
...Quid-pro-quo evidence, and Pardon pipeline evidence. The ecosystem of influence is real; the video’s single-mastermind diagram is rhetorical compression that slides into conspiratorial thinking.
p4sc4l.substack.com
December 20, 2025 at 12:28 PM
By aligning with Russia and adopting its grievances against the Western liberal order, the administration validates its own domestic war against the “globalists” who built that order.
The “Reverse Kissinger” strategy—aligning with Russia to counter China... p4sc4l.substack.com/p/by-alignin...
By aligning with Russia and adopting its grievances against the Western liberal order, the administration validates its own domestic war against the “globalists” who built that order.
The “Reverse Kissinger” strategy—aligning with Russia to counter China—serves a potent domestic political function: to frame its abandonment of NATO, EU not as isolationism, but as "smart" realpolitik
p4sc4l.substack.com
December 20, 2025 at 12:27 PM
Coursera will acquire Udemy. For publishers—especially scholarly, professional, and educational publishers—the merger is not just an ed-tech headline.
p4sc4l.substack.com/p/coursera-w...
Coursera will acquire Udemy. For publishers—especially scholarly, professional, and educational publishers—the merger is not just an ed-tech headline.
It is a clear signal that scale, data, distribution, and AI-readiness are becoming existential requirements rather than optional advantages.
p4sc4l.substack.com
December 20, 2025 at 12:26 PM
Adobe’s SLMs and their alleged training on pirated books via SlimPajama, a derivative of Books3. If courts accept that weights encode infringing copies, the legal consequences would be profound.
p4sc4l.substack.com/p/adobes-slm...
Adobe’s SLMs and their alleged training on pirated books via SlimPajama, a derivative of Books3. If courts accept that weights encode infringing copies, the legal consequences would be profound.
Complaint paints a sharply different picture—one that, if substantiated, could undermine industry narratives about clean datasets, open-source insulation, and technical abstraction from infringement.
p4sc4l.substack.com
December 20, 2025 at 12:25 PM