Shows
AI Stress TestWhy the FDA's food label revolution predicts AI's transparency futureOver 35 years ago, the US FDA began transforming fragmented nutrition disclosure into a single mandatory standard - a regulatory evolution that moved most FDA‑regulated packaged foods onto a single standardised Nutrition Facts label over the course of a few years.
In leading edge research that parallels this labelling precedent, the AI Transparency Atlas has unmasked a critical AI transparency gap: even though many AI providers sit below 60% safety documentation compliance, most users remain blind to model behaviors, hallucinations and deception risks. This lack of AI labelling coincides with a focus on more transparent practice from a range of frontier la...
2026-01-2943 min
AI Stress TestWhy corporate AI adoption in 2026 mirrors the PC's inevitable ascentOver 44 years ago, IBM did something remarkable - it turned legitimacy into a market strategy. A $4,000 computer package that had seemed absurd to most people became essential once IBM made it respectable. Five years later, the 'unnecessary' PC dominated boardrooms and homes alike.
In leading-edge research, IEEE's Global Survey results capture an identical moment for agentic AI with 96% of global technologists agreeing that agentic AI innovation, exploration and adoption will continue at lightning speed and 59% of enterprises accelerating investment. The institutional credibility phase is underway. What happened next in the 1980s was inevitable adoption—and the adoption curve for AI is...
2026-01-2237 min
AI Stress TestBuild vs. buy, the cycle continuesThe 1960s-70s mainframe era established a pattern; enterprises rejected commercial software offerings, choosing instead to build custom applications in-house. The willingness to accept substantially higher costs and longer development timelines reflected a single calculus - strategic control over technology tethered to competitive advantage outweighed efficiency gains from standardised platforms.
Fast forward to today. A systematic study of production AI agents (engaging 306 practitioners and 20 detailed case studies) documents that 85% of case study teams proceeded without third-party agent frameworks, building custom implementations from scratch. Human evaluation was relied on in 74% of cases and agent autonomy was constrained to fewer than 10 steps...
2026-01-1641 min
AI Stress TestFrom fragmentation to dangerous consensusBetween 1830 and 1886, American railways faced a coordination crisis: 23 independent gauge decisions created a fragmented network where the Southern Railway & Steamship Association's coordinated conversion of approximately 11,500 miles in May-June 1886 solved the integration problem through institutional coordination.
In leading edge research from the University of Washington, frontier LLMs now exhibit 71-82% output homogeneity, potentially linked to RLHF (Reinforcement Learning from Human Feedback) alignment, suggesting that enterprises relying on multi-model decision-making inherit a coordination solution that has accidentally reversed itself - diversity in form, convergence in substance.
Both episodes reveal how systems driven by local optimisation and local switching costs create paradoxical fragility...
2026-01-0827 min
AI Stress TestFrom fragmentation to dangerous consensusBetween 1830 and 1886, American railways faced a coordination crisis: 23 independent gauge decisions created a fragmented network where the Southern Railway & Steamship Association's coordinated conversion of approximately 11,500 miles in May-June 1886 solved the integration problem through institutional coordination.
In leading edge research from the University of Washington, frontier LLMs now exhibit 71-82% output homogeneity, potentially linked to RLHF (Reinforcement Learning from Human Feedback) alignment, suggesting that enterprises relying on multi-model decision-making inherit a coordination solution that has accidentally reversed itself - diversity in form, convergence in substance.
These inversed scenarios reveal how systems driven by local optimisation and local switching costs create paradoxical...
2026-01-0827 min
AI Stress TestNew classes, new skillsBetween the 1850s and the 1930s, capitalism experienced a permanent separation of ownership from control as professional managers assumed operational authority - and nearly simultaneously, enterprises invented systematic governance frameworks (standardised managerial accounting, formalised audits, standardised business education) to monitor and constrain these newly powerful delegates.
The emerging agentic enterprise represents an analogous inflection point: MIT research identifies 2025 as a critical moment where AI orchestration transitions from siloed experimentation to integrated infrastructure, triggering parallel governance imperatives - policy-as-code frameworks, real-time monitoring systems, and new credentialisation pathways to manage AI agents as semi-autonomous coordinators operating within bounded governance parameters.
What differs fu...
2025-12-1130 min
AI Stress TestWhen ‘safe until proven otherwise’ becomes dangerousOver a century ago, in 1906, the first documented death from asbestos exposure was recorded in testimony. Yet it would take until the 1920s for widespread medical evidence to emerge, and until 2006 - 100 years later - for meaningful regulation to gain momentum with the Rotterdam Convention. The culprit? A confidence gap: industry and institutions trusted the material's benefits whilst deprioritising the evidence of its harms.
New enterprise AI research has identified a strikingly similar phenomenon: 78% of organisations claim full trust in AI systems, despite only 40% having implemented governance frameworks or ethical safeguards. Yet data shows that organisations prioritising trustworthy AI see...
2025-12-0428 min
AI Stress TestRenaissance blueprints, AI control systemsOver 540 years ago, following a plague that decimated up to half of Milan's population, Leonardo da Vinci sketched designs for an integrated ideal city where water management, sanitation, layered circulation and human flow were unified into a single coherent system - conceptualising the city itself as a technological solution to disease, congestion, and human suffering rather than a collection of isolated structures.
Singapore's national digital twin - a federated capability integrating real-time sensor networks from 10,000+ distributed points, AI-powered anomaly detection, and cross-agency governance coordination - exemplifies how frontier AI transforms infrastructure from a static, episodically-maintained liability into a continuously intelligent sy...
2025-11-2734 min
AI Stress TestThe economics of asymmetric advantage have fundamentally changedOver 37 years ago, Robert Morris released a self-propagating worm that infected roughly 10% of the entire internet within hours - awakening the world to the reality that autonomous code could outpace human response.
Today, new AI safety research has documented the first large-scale cyberattack where artificial intelligence orchestrated reconnaissance, exploitation, and data exfiltration across 30 global targets with 80–90% autonomy - rendering human-speed threat detection asymmetrically disadvantaged.
Whilst Morris's 1988 accident revealed systemic fragility, the November 2025 Claude manipulation reveals something far more consequential: a deliberate manipulation of agentic AI capabilities that defence teams are challenged to contain.
When defence requires matching human expertise against ma...
2025-11-2031 min
AI Stress TestWhen evolution stops being randomOver 80 years ago, Alexander Fleming witnessed what he feared most - bacteria evolving resistance to penicillin within a decade of its mass introduction; yet, despite his public warnings, the cascade of resistance emerged precisely as foreseen - triggering an evolutionary arms race where successive antibiotic deployments accelerated rather than slowed resistance emergence, collapsing therapeutic horizons across bacterial species.
In leading edge research released this month, Google's Threat Intelligence Group identified regenerative AI-native malware families employing just-in-time code regeneration, via LLMs, to rewrite their entire source code hourly to evade detection - rendering traditional incident responses fundamentally misaligned with threat velocity.
...
2025-11-1336 min
AI Stress TestInvisible sources of contaminationOver 171 years ago, a contaminated water pump in Victorian London killed 616 people in a month because the poison was invisible to inspection, undetectable by the science of the era and absolutely trusted by those who consumed it.
New leading edge research from the UK AISI, Anthropic and the Alan Turing Institute demonstrates that language models remain vulnerable to persistent backdoors inserted via minimal poisoned data, challenging the assumption that larger training datasets dilute poisoning effects.
In addressing Cholera, the rapid emergence of symptoms provided an advantage, enabling epidemiological responses - in contrast, data poisoning can remain dormant, undetectable and potentially...
2025-11-0631 min
AI Stress TestFrom nuclear testing moratoriums to AI safety thresholdsOver 33 years ago, the US conducted its final nuclear test - transitioning to a science-based Stockpile Stewardship Program that maintains civilization-ending arsenals through simulations alone, never testing them again.
New AI safety research has found frontier models achieving 70% performance on complex software engineering tasks and demonstrating potential for weaponised capabilities - prompting developers to deploy unprecedented safety controls not because they've definitively crossed danger thresholds, but because they cannot rule out crossing them.
Whilst nuclear warheads and AI systems are very different - one frozen-in-time physics, the other evolving from under 10% to 70% capability in under two years - we can draw int...
2025-10-3034 min
Creek Street Church SermonsSunday Morning Service 10th of March 2024MC - Mel Ferres, Offering - Snr Min Geoff Kean, Communion - Caleb Maxwell
2024-03-1251 min
Creek Street Church SermonsSunday Morning Service 10th of December 2023MC - Snr Ps Tim Maxwell, Offering - Snr Ps Tim Maxwell, Preaching - Snr Min Geoff Kean, Communion - Mel Ferres
2023-12-1150 min
Creek Street Church SermonsSunday Morning Service 12th of November 2023MC - Caleb Maxwell, Offering - Snr Ps Tim Maxwell, Preaching - Snr Min Geoff Kean, Communion - Mel Ferres
2023-11-141h 07