How AI Uncovered a Critical PHP Security Flaw Threatening Millions of Websites

25/10/26 10:00 Filed in: Technology | Business

Advanced AI models helped cybersecurity researchers identify and investigate complex weaknesses that had remained hidden within widely used software for years.

And that raises a much bigger question.

If AI can help security researchers discover vulnerabilities faster than ever before, could it also give cybercriminals the ability to exploit weaknesses that were previously beyond their capabilities?

In this episode of the Inspiring Tech Leaders podcast, I explore what happened with PHP2Shell, how AI contributed to its discovery, and why this incident could mark an important turning point in cybersecurity. But I also explore something that I believe is even more significant for technology leaders.

Many organisations still struggle to understand exactly what technology they're running, where their vulnerabilities exist and how quickly they can respond when a critical threat emerges.

And in a world where AI is accelerating both security research and potentially cyberattacks, that lack of visibility could become an increasingly serious business risk.

The biggest cybersecurity challenge may no longer be discovering vulnerabilities. It may be responding quickly enough to protect ourselves from them.

Will AI ultimately give cybersecurity defenders the advantage, or are we entering an era where attackers will be able to move faster than organisations can protect themselves?

Available on: Apple Podcasts | Spotify | YouTube | All major podcast platforms

Apple and Meta’s AI Alliance - The Rise of the Desktop AI Supercomputer

18/10/26 10:00 Filed in: Technology | Business

What if the future of artificial intelligence isn't sitting inside a billion-dollar data centre, but on the desk right in front of you?

For years, the AI revolution has been dominated by enormous investments in cloud infrastructure, specialist processors and increasingly powerful data centres.

But an unexpected development involving two of technology's biggest rivals, Apple and Meta, could challenge some of our assumptions about where AI is heading.

In this episode of the Inspiring Tech Leaders podcast, I explore how Meta's new Muse Spark Open AI model, with an extraordinary 816 billion parameters, can run locally across a collection of Apple's Mac Studio computers.

This is more than an impressive technical achievement. It raises some fascinating questions about the future of enterprise AI.

Could businesses begin running sophisticated AI systems on infrastructure they own and control, rather than relying entirely on cloud providers?

Could local AI offer greater control over sensitive business information, reduce certain operating costs and create new opportunities for deploying autonomous AI agents? And perhaps most importantly, could this development begin to challenge the business models of the major cloud AI providers?

For Technology Leaders, that introduces some important strategic considerations around cost, data sovereignty, infrastructure investment and long-term dependence on technology providers.

It also demonstrates something else about the rapidly evolving AI industry. Companies that have historically been fierce competitors can find common ground when their interests align around emerging technologies.

In this episode, I explore what Apple and Meta's unexpected collaboration means for the future of AI, why local AI could become increasingly important for businesses, and what technology leaders should be thinking about today.

The future of AI might not just be about who builds the most powerful model. It could be about who gives businesses the greatest control over how they use it.

Available on: Apple Podcasts | Spotify | YouTube | All major podcast platforms

Reflection AI's Beam - The $25 Billion Threat to OpenAI and Anthropic

11/10/26 10:00 Filed in: Technology | Business

Could a $25 billion AI challenger disrupt OpenAI, Anthropic and China's growing dominance in artificial intelligence?

A relatively young American company called Reflection AI, backed by Nvidia, could become an important player in the AI market.

Its newly announced AI model, Beam, isn't simply another attempt to compete with ChatGPT or Claude. Reflection has a much bigger ambition, which is to change how businesses build, deploy and control artificial intelligence.

For years, organisations have largely depended on a handful of technology giants to provide access to increasingly powerful AI models. But the rapid advancement of Chinese open-weight models from companies including DeepSeek, Alibaba and Z.ai is challenging that approach.

These models are becoming increasingly capable, often at significantly lower operating costs, raising important questions about whether businesses really need to pay premium prices for every AI workload.

Reflection AI believes there is an opportunity to offer Western businesses and governments a compelling alternative, combining powerful AI capabilities with greater control over data, infrastructure and costs.

But can Beam genuinely compete with China's leading open-weight models, let alone established providers such as OpenAI and Anthropic?

In this episode of the Inspiring Tech Leaders podcast, I explore Reflection AI's ambitions, Nvidia's strategic involvement and why the emergence of Beam could have significant implications for enterprise AI.

I also examine the growing importance of sovereign AI, the potential for organisations to operate their own AI infrastructure, and why the economics of artificial intelligence may become just as important as the capabilities of the models themselves.

Of course, there are challenges. Beam has yet to demonstrate that it can outperform the industry's leading models, and operating AI independently introduces additional responsibilities around security, governance and technical expertise.

For CIOs, CTOs and business leaders, I believe the bigger question is no longer simply which AI model is the most powerful, but which combination of technology, data and infrastructure delivers the greatest business value.

We could be moving towards an AI market where organisations have considerably more choice over the technology they use, how they deploy it and who ultimately controls it.

I'd be interested to hear your thoughts. Will open-weight AI eventually challenge the dominance of closed models, or will OpenAI, Anthropic and Google continue to lead the enterprise AI market?

Available on: Apple Podcasts | Spotify | YouTube | All major podcast platform

Gemini 4 Argon vs Claude Opus 5.5 vs GPT 6 - Which AI Wins for Business?

04/10/26 10:00 Filed in: Technology | Business

The AI race has just become much more interesting.

Google Gemini 4 Argon, Anthropic Claude Opus 5.5 and OpenAI’s GPT 6 family represent the latest generation of frontier AI models.

But if you’re making technology decisions for a business, there’s a problem with simply asking which one is the most intelligent.

Businesses don’t buy benchmark scores. They buy outcomes.

In this episode of the Inspiring Tech Leaders podcast, I compare these three AI platforms from a business perspective rather than disappearing into the technical jargon.

I look at where Gemini 4 Argon could fit into Google’s wider enterprise ecosystem, why Claude Opus 5.5’s combination of capability and efficiency is significant, and how OpenAI’s GPT 6 Astra and GPT 6.1 Sol are creating a choice between maximum capability and more economical AI at scale.

There is also a much bigger shift happening.

These systems are increasingly moving beyond simply answering questions and generating content. They are being designed to undertake longer, more complex pieces of work, interact with business systems and potentially complete entire workflows.

That changes the conversation.

The question is no longer simply, which AI model is smartest? It’s which AI can deliver the most value for your organisation.

That means looking beyond benchmarks towards the things that actually matter in business: productivity, integration, reliability, governance, security, human oversight and the total cost of completing a task.

And as the performance gap between frontier models becomes increasingly difficult to separate, those factors could ultimately become far more important than who happens to be at the top of a benchmark this week.

In this episode, I explore what technology leaders should be considering before choosing between Google, Anthropic and OpenAI.

If you were choosing an AI platform for your organisation today, what would matter most: capability, cost, integration, governance, or something else?

Available on: Apple Podcasts | Spotify | YouTube | All major podcast platforms