Showing posts with label #AI. Show all posts
Showing posts with label #AI. Show all posts

Thursday, 12 March 2026

HOW WAR HINDERS AMERICA IN THE AI RACE WITH CHINA

12 March 2026

1. The Pentagon, Strategy And Unexpected Consequences

The Pentagon has floor after floor of offices full of strategists and planners. One assumes they analyse first, second and third order effects of any conflict, prepare Plan Bs, timings and so on. They surely knew that this war was not going to work out the way Trump and Netanyahu thought it would work out they thought so and they said, three-star general Kaine said so, payborn Trump the military may not be able to complete the mission. 

And indeed, the world’s strongest military by far finds itself twelve days into a conflict and the Strait of Hormuz remains closed. Iran is pushing out drone boats laiden with explosives, so no tanker will risk that a nor will any insurer. 

Most commentary focuses on the obvious consequences: higher oil prices and disruption to global shipping. But there is an overlooked question. What thought has gone into the effect on America’s technological competition with China?

After all, American global leadership depends in part on winning the technology race, and today that race centres on Artificial Intelligence.

---

2. AI Runs On Electricity

Artificial Intelligence is often presented as a triumph of software and algorithms. In reality it is also a massive industrial system that runs on electricity.

Training large AI models requires enormous data centres packed with specialised processors. These installations consume extraordinary amounts of power.

Here the comparison with China is uncomfortable. Over the past two decades the United States has added relatively little to its electricity generation and transmission capacity. China, by contrast, has expanded its grid at breathtaking speed, reportedly adding the equivalent of the entire American electricity grid in roughly four years.

If the future of AI depends on access to abundant electricity, then the underlying energy infrastructure matters as much as the technology itself.


---

3. The Hyperscalers And The Infrastructure Race

At the centre of this system sit the hyperscalers.

These are the giant technology companies that operate vast global cloud computing networks. Their infrastructure forms the backbone of the modern internet, supporting Artificial Intelligence, cloud services, streaming platforms and corporate computing systems.

Companies such as Amazon, Microsoft, Google and Oracle are building ever larger data centres across the world.

These projects require two things above all: power and capital.

Much of the investment comes from the companies’ own balance sheets. But a substantial portion is financed through borrowing. That means the economics of AI infrastructure depend heavily on stable financial conditions and relatively low interest rates.

Wars complicate this equation. Military spending increases government borrowing, which pushes bond yields higher. Higher yields raise the cost of financing the massive infrastructure that AI development requires.

Amazon – through Amazon Web Services (AWS)
Microsoft – through Azure
Google – through Google Cloud
Meta – operating huge internal data-centre networks
Alibaba – dominant hyperscaler in China
Tencent – another large Chinese cloud provider

---

4. Energy Prices And The Geography Of Data Centres

Energy prices also matter.

Many hyperscale companies have been exploring locations in the Gulf region precisely because of abundant and relatively cheap energy. Large data centres require reliable electricity at competitive prices in order to remain viable.

A prolonged disruption in the Strait of Hormuz pushes oil prices upward, increases energy costs globally and introduces uncertainty into energy markets.

This is hardly the environment that technology companies prefer when planning multi-billion-dollar infrastructure projects.


---

5. The Strategy

The ingredients required for success in the AI race are surprisingly mundane.

A modern and expanding electricity grid.
Low and stable interest rates.
Reliable energy supplies.
Oil prices somewhere in the $60 range and certainly well below $100.

In short, the geopolitical conditions that allow hyperscalers to build the digital infrastructure of the future.

American hegemony ultimately depends on technological leadership, so these under appreciated economic conditions matter as much as aircraft carriers or missile systems.


---

6. Glossary

Artificial Intelligence (AI) – computer systems capable of learning from data and performing tasks that normally require human reasoning or pattern recognition.

Hyperscaler – very large technology companies that operate massive global cloud computing infrastructure supporting AI, internet services and corporate computing.

Data Centre – a facility housing thousands of servers and specialised processors used to store and process digital information. Cf data centre v. AI data centre. 

Electricity Grid – the network of power generation, transmission and distribution systems that deliver electricity to industry and households.

Strait of Hormuz – the narrow maritime passage between Iran and Oman through which roughly one fifth of global oil trade normally passes.


---

References


International Energy Agency – Electricity 2024
https://www.iea.org/reports/electricity-2024

U.S. Energy Information Administration – Electric Power Data
https://www.eia.gov/electricity

Bloomberg – AI Data Centre Power Demand
https://www.bloomberg.com/news/articles/2024-05-09/ai-power-demand-data-centers

McKinsey – The Rise of Hyperscale Data Centres
https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-coming-hyperscale-data-center-boom

Monday, 9 March 2026

A SPECIALIST HIGH TECH AGENCY FOR THE BUILT ENVIRONMENT

OVERVIEW

A small architectural agency has just made a rather remarkable move. Instead of competing in the crowded world of traditional design studios, it has begun repositioning itself as a specialist consultancy at the intersection of AI Artificial Intelligence, software and architecture. Equipped with advanced software engineering tools and a private AI infrastructure running LLM local language models, this agency can analyse planning rules, automate design workflows and simulate development scenarios. 

The story reflects a broader shift taking place across the construction and property development sectors: in response to the increasing complexity of projects, planning regulations, cost pressures and environmental requirements, there is now a demand for far more sophisticated analysis than traditional design workflows can provide.

This is precisely where small, technically specialised and agile consultancies are beginning to play a role and architects are moving from drawing buildings to designing intelligent systems.

Just as FinTech transformed banking fifteen years ago, a new generation of small groups of engineers are creating ConTech tools that will now begin reshaping how buildings are designed and delivered.

Local language model – an AI model running on private hardware rather than external cloud services.

Design automation – the use of algorithms to generate or optimise design solutions.

1. AI And The Future Of Architecture

There is currently a great deal of noise about Artificial Intelligence replacing jobs. The debate began in software development but it is now spreading into many other professions, including architecture. AI tools can already generate code, analyse data, produce reports and create design imagery. It is therefore understandable that architects are asking whether their profession may face the same disruption.

Yet history suggests something more subtle usually occurs. New technologies rarely eliminate complex professions. Instead they remove repetitive work inside those professions. When spreadsheets appeared they did not eliminate accountants, but they removed a vast amount of manual calculation. The role of the accountant shifted toward interpretation and advice.

Architecture contains many similar forms of repetitive labour. Drawings, compliance checks, schedules and cost calculations all follow predictable patterns that software can increasingly assist with. The profession therefore evolves rather than disappears. The architect gradually becomes less of a draughtsman and more of a designer of systems.

A useful historical parallel comes from finance. Around 2008–2012, small groups of engineers began building software platforms that challenged the traditional banking system. Companies such as Stripe, Square and TransferWise (now Wise) demonstrated that a handful of technically skilled founders could create tools that were faster, cheaper and easier to use than the systems maintained by large financial institutions. What later became known as the FinTech revolution began not inside the banks, but in small technology-driven teams experimenting with software.

Something similar may now be emerging in the built environment. As digital modelling, data systems and artificial intelligence become more central to how buildings are conceived and delivered, small specialist agencies are beginning to develop software tools that complement or sometimes bypass traditional architectural workflows. If the analogy holds, the ConTech wave of the 2020s may look surprisingly similar to the FinTech wave of the early 2010s: small technical teams building new digital tools around industries that have historically been slow to changes. 

  • Artificial Intelligence - computer systems capable of performing tasks that normally require human reasoning, pattern recognition and learning.

  • Automation - the use of machines or software to perform tasks with minimal human intervention.

  • FinTech – digital technologies and software platforms designed to automate or improve financial services such as payments, banking and lending.

  • PropTech (property technology) – technology platforms transforming property development, real estate investment and building management.

  • ConTech (construction technology) – software systems and digital tools used to improve how buildings and infrastructure are designed, constructed and managed.


2. What AI Can Already Do Inside Architectural Practice

AI assisted design tools are already capable of performing several technical tasks that traditionally consumed large amounts of time in architectural studios. These tools are developing rapidly and are beginning to alter the internal structure of many design offices.

Typical examples include:

• generating multiple design variations automatically through generative design systems
• analysing planning regulations and compliance requirements
• producing early stage spatial layouts and building massing studies
• estimating material quantities and construction costs
• assisting with documentation and technical reports

The result is not the disappearance of architects but a shift in where their expertise is applied. The architect becomes the person who defines the design problem, sets the constraints, selects the appropriate computational tools and evaluates the results.

The professional role therefore becomes closer to system design than manual drawing.

  • Generative Design - computational design techniques that automatically produce many design alternatives based on defined constraints such as cost, spatial requirements or energy use.

  • BIM (Building Information Modelling) - a digital model of a building containing geometric information, materials, engineering systems and construction data.

  • Parametric Design - a design method in which relationships between elements are defined mathematically so that designs update automatically when parameters change.


3. A Quiet Revolution: The Home AI Server

A particularly interesting development is the possibility of running powerful AI systems locally rather than relying entirely on large cloud platforms. Small agencies can now operate advanced language models on their own machines.

A modest home server equipped with modern graphics processing units can run open source models such as Llama or Mistral. When configured properly these systems can function as internal research assistants and automation tools.

For a small architectural consultancy, such a system can assist with tasks such as:

• analysing planning regulations and zoning rules
• generating scripts that automate BIM or CAD workflows
• assisting in writing technical reports and proposals
• analysing property datasets and development feasibility studies
• summarising technical research and engineering standards

Because the system operates locally, sensitive project information never leaves the organisation’s own infrastructure. For consultants working with commercial clients this can be extremely valuable.

What once required a large research department can increasingly be achieved by a technically capable small agency equipped with the right tools.

  • Large Language Model (LLM) - an AI system trained on extremely large collections of text that can analyse and generate human language.

  • Open Source Software - software whose source code is publicly available and can be modified or distributed freely.

  • GPU (Graphics Processing Unit) - specialised computer hardware designed for parallel processing and commonly used to run AI models efficiently.


4. The Rise Of The Architect–Technologist

These developments are producing a new hybrid professional profile: the architect who also understands software systems.

Construction remains one of the least digitised sectors of the global economy. Research by McKinsey has repeatedly shown that productivity growth in construction has lagged far behind most other industries. At the same time, global spending on buildings and infrastructure continues to expand.

This gap creates opportunity.

An architect who can write code, automate processes and analyse data can operate at a valuable intersection between design and technology. Instead of merely producing drawings, such professionals can design the digital systems that support the entire development process.

This work might include:

• building algorithmic design tools for architects and developers
• creating automated planning analysis systems
• modelling development feasibility using property data
• integrating AI assistants into architectural workflows
• developing digital twins for buildings or infrastructure

The profession therefore expands into a new territory combining architecture, data and computation.

  • Digital Twin - a dynamic digital model of a physical building or infrastructure system that updates using real world data.

  • Algorithmic Design - the generation of design solutions using computational rules or algorithms.

  • Construction Technology (ConTech) - digital tools and software systems designed to improve the construction and property industries.


5. The Importance Of A Small Independent Agency

One aspect of this story deserves particular emphasis. The creation of a small independent agency combining architecture and software development is itself a significant entrepreneurial achievement.

Most professionals remain employees throughout their careers. Establishing an agency requires technical competence, commercial initiative and a willingness to accept uncertainty. A small agency also provides something extremely valuable: strategic freedom.

An independent structure offers its customers something unique: experimentation with new technologies, development of proprietary tools and the ability to pursue specialised consulting work. In periods of technological change this flexibility becomes extremely important.

Large firms often struggle to adapt quickly because they are tied to established processes and organisational structures. Small agencies can explore new directions with far greater agility.

  • Entrepreneurship - the process of creating and managing a business venture that involves financial risk in pursuit of profit.

  • Consultancy - a professional service in which specialised expertise is provided to organisations on a project basis.


6. A One Year Strategic Path For Repositioning

A sensible strategy for the coming year is not rapid expansion but careful repositioning. The goal is to move the agency toward the intersection of architecture and technology.

During the first phase the focus should be technical capability. The agency can configure its home AI server, integrate local language models into research workflows and develop a library of small automation tools linked to BIM or design software.

The second phase should focus on intellectual visibility. Publishing articles that explain how AI and automation can reshape architectural practice helps establish credibility. Demonstrating working prototypes is far more persuasive than theoretical commentary.

The third phase involves engagement with the market. Small developers, design studios and property investors frequently lack digital expertise. A consultancy capable of automating planning analysis, modelling development scenarios or building property analytics tools can provide specialised services that traditional firms cannot.

By the end of the year the agency can present a clear identity. Rather than appearing as a small architectural practice it becomes a technology consultancy for the built environment.

  • Strategic Positioning - defining how an organisation differentiates itself within a competitive market.

  • Built Environment - the human made surroundings in which people live and work, including buildings, infrastructure and urban spaces.

  • BIM (Building Information Modelling) - the next step beyond a digital mockup, a digital system used in architecture and construction that creates a detailed three-dimensional model of a building containing not only geometry but also data about materials, structure, costs, and construction processes, allowing architects, engineers and contractors to collaborate using a shared information model.


7. The Real Opportunity

Artificial Intelligence will undoubtedly change architectural practice. It will reduce the need for some forms of routine drafting and documentation. 

Architecture sits at the intersection of engineering, economics, regulation, aesthetics and human needs. These domains require judgement, negotiation and responsibility. Machines can assist with analysis, but they do not replace human decision making.

What AI will do is increase leverage for those who understand how to use it.

AI engineers who learn to work with computational design systems may gain enormous productivity advantages for their clients. Those who ignore these tools may find that parts of the profession move beyond them.

Seen from this perspective, a small agency experimenting with software development and locally-hosted AI models can represent something more than an interesting opportunity for experimentation and hi-tech innovation. It may actually be an early mover into an architectural practice of the future.


The agency works with developers, architects and engineering firms to design digital tools and AI systems that improve how buildings and projects are conceived and delivered.

If the ideas in this article resonate with your own work in development, design or construction technology, do feel free to get in touch.


References

McKinsey Global Institute – Reinventing Construction: A Route to Higher Productivity
https://www.mckinsey.com/industries/capital-projects-and-infrastructure/our-insights/reinventing-construction

Autodesk – Generative Design Overview
https://www.autodesk.com/solutions/generative-design

Stanford Human Centered AI – AI Index Report
https://aiindex.stanford.edu

Meta AI – Llama Large Language Models
https://ai.meta.com/llama/

Friday, 19 September 2025

WHAT SHOULD DEVELOPERS BE THINKING ABOUT IN THE AGE OF AI

19 September 2025

What Should Developers Be Thinking About in the Age of AI?

As tools like Replit, GitHub Copilot, and Cursor advance rapidly, writing code is becoming increasingly automated. These AI platforms can generate boilerplate, suggest completions, and scaffold entire applications. But if machines can write the code, what remains for human developers?

The answer: everything that matters most

Beyond Code: The New Developer Skillset

The developers who will thrive bring architectural thinking - making clean, scalable design decisions that serve long-term project health. They weigh tradeoffs between monolithic and microservice architectures, plan for extensibility, and spot performance issues before they emerge.

They also possess product sense. Great developers don't just deliver what's in the ticket - they challenge it. They understand users and the business needs, about UX, and help shape better features through collaboration with designers and stakeholders. AI can mock up a UI, but it can't feel user friction or spot dead ends in a user journey.

Where AI Falls Short

What strikes me most about AI's limitations is its inability to handle ongoing real-world complexity. Most systems aren't clean slates - they involve legacy APIs, undocumented logic, and client-specific workarounds. These messy, ambiguous environments are where experienced developers excel, making sense of contradictions and bringing order out of chaos.

AI also can't coordinate people and teams. It's an algorithm trained on past experiences, not a person who can prioritise, understand real-world needs, or manage collaboration across teams. There's always a veil of reality between you and The Computer.

The Human Advantage

A good development team isn't just typing code - they're translating messy, evolving business needs into elegant, scalable, secure, trustworthy systems. They build for failure with proper logging, tracing, and testing. They review code, mentor juniors, and make tough decisions calmly, often under pressure.

AI can assist with all this, but it can't replace the depth of thinking needed when quality directly affects user trust and long-term success - like scaling to a million users or iterating / continuous improvement based on real customer feedback.

The New Reality

If you picture software development as a high-rise building, AI has essentially emptied the first half-dozen floors of workers and given you a fast lift to the higher levels of thinking.

The future belongs to developers who embrace this shift - those who focus on architecture, product understanding, team leadership and personnel coordination, and navigating real-world complexity. Coding was always very intellectually demanding, stressful even. Deciding what to build, how to build it, getting everyone on board, and ensuring what's being built serves real users - that's the new frontier.

Monday, 9 June 2025

AI CAREERS

9 June 2025

 AI Careers

(Salary ranges, career paths - to be added)


1.    Job Title: AI Prompt Engineer

Overview:
As an AI Prompt Engineer, you will design, test, and refine prompts to improve the performance of AI language models. This role sits at the intersection of software engineering, linguistics, and user experience.

Key Responsibilities:

  • Craft and iterate on high-quality prompts for AI models.
  • Analyse model outputs to refine instructions.
  • Collaborate with developers and product teams to implement prompt-based solutions.

Required Skills:

  • Strong communication and analytical thinking.
  • Familiarity with LLM behaviour (e.g., GPT-4, Claude).
  • Basic programming knowledge (e.g., Python, APIs).

 

2.    Job Title: AI Developer Advocate

Overview:
The AI Developer Advocate bridges the gap between AI tool creators and the developer community. You’ll help others build smarter tools with AI by writing content, hosting tutorials, and gathering feedback.

Key Responsibilities:

  • Educate developers on how to integrate AI tools.
  • Create demos, sample code, and blog posts.
  • Represent the company at meetups and conferences.

Required Skills:

  • Strong public speaking and writing skills.
  • Deep understanding of developer workflows.
  • Experience with modern AI libraries and APIs.


 

3.    Job Title: AI Integration Engineer

Overview:
This role focuses on embedding AI tools into existing business systems. As an AI Integration Engineer, you’ll ensure smooth, secure, and scalable implementation of AI functionalities in real-world applications.

Key Responsibilities:

  • Integrate APIs from AI providers into enterprise apps.
  • Monitor performance and error handling.
  • Collaborate with cross-functional teams to improve system architecture.

Required Skills:

  • Backend software development (Python, Java, Node.js).
  • Understanding of API protocols and data handling.
  • Cloud deployment and version control systems.


 

4.    Job Title: AI Solutions Designer

Overview:
As an AI Solutions Designer, you’ll map business problems to AI use cases and design human-in-the-loop workflows. You combine technical acumen with strategic thinking.

Key Responsibilities:

  • Work with clients to scope AI projects.
  • Design system architecture and user journeys.
  • Define evaluation metrics and testing scenarios.

Required Skills:

  • UX/UI awareness.
  • Strategic consulting or product design experience.
  • High-level understanding of AI capabilities.


 

5.    Job Title: AI Test & QA Analyst

Overview:
This role ensures AI tools work reliably and ethically. You'll test systems for accuracy, fairness, and performance across different scenarios.

Key Responsibilities:

  • Test prompt reliability and LLM outputs.
  • Conduct edge-case scenario analysis.
  • Build test suites and simulate user inputs.

Required Skills:

  • QA methodologies.
  • Familiarity with prompt tuning and LLM limits.
  • Documentation and reporting skills.


 

6.    Job Title: AI Technical Writer

Overview:
You translate complex AI systems into clear, useful documentation. You’ll work alongside developers and product teams to produce user guides, API docs, and onboarding material.

Key Responsibilities:

  • Write and maintain technical documentation.
  • Organise help centres and chatbot documentation.
  • Ensure clarity, consistency, and accuracy.

Required Skills:

  • Strong writing and editing skills.
  • Technical background or ability to grasp complex systems.
  • Familiarity with markdown, API tooling, and diagrams.
 

Update

Updated Job Descriptions for Emerging AI Roles in 2025

(Post-Prompt Engineering Era - prompt engineering was a stepping Stone in to AI. A couple of years ago, but now anyone who can type is expected to be able to write prompts, it's been operationalized, has become just a part of our daily lives, whether you are an office worker or an individual)


7. Job Title: AI Trainer

Overview:

An AI Trainer develops, tests and refines AI behaviour, ensuring natural, useful, and contextually appropriate responses in conversation-based systems. You’ll play a critical role in improving chatbot understanding, tone, and alignment with user intent.

Key Responsibilities:

Design realistic and varied user interaction scenarios.

Fine-tune chatbot behaviour based on user feedback and performance logs.

Create annotated datasets to improve model alignment.

Collaborate with data scientists and NLP engineers to deploy updated models.


Required Skills:

Linguistic awareness and UX sensitivity.

Familiarity with AI training pipelines and prompt tuning.

Ability to detect bias, hallucinations, or misalignment in model outputs.


Ideal Background:

Former content creators, UX writers, linguists, or prompt engineers transitioning to model behaviour design.



---

8. Job Title: AI Data Specialist

Overview:

The AI Data Specialist ensures AI models are trained on clean, relevant, and well-structured data. You will be responsible for preparing datasets, enforcing data governance, and continuously auditing data pipelines for quality.

Key Responsibilities:

Clean and structure large datasets for model consumption.

Detect anomalies, duplicates, or corrupted entries in training sets.

Collaborate with AI trainers and engineers to ensure data quality and relevance.

Maintain documentation and lineage tracking for data assets.

Required Skills:

Experience with SQL, Python (Pandas), and data labelling tools.

Understanding of machine learning model data needs.

Strong attention to detail, with an eye for statistical anomalies.


Ideal Background:

Data analysts or engineers pivoting into AI infrastructure support.


9. Job Title: AI Security Specialist

Overview:

As an AI Security Specialist, you’ll safeguard AI systems from evolving threats such as prompt injection, data poisoning, and adversarial attacks. You will work at the cutting edge of cybersecurity and machine learning.

Key Responsibilities:

Conduct vulnerability assessments on AI systems.

Design defences against prompt manipulation and misuse.

Ensure safe deployment of models within enterprise environments.

Monitor for suspicious access, misuse patterns, and insider threats.


Required Skills:

Strong grounding in cybersecurity principles and threat modelling.

Familiarity with LLM architecture, sandboxing, and red teaming techniques.

Knowledge of privacy-preserving AI techniques (e.g. differential privacy).


Ideal Background:

Cybersecurity professionals with interest in emerging AI threats; or ML engineers upskilling in security.