1. Start With Respect: Understand the Work That’s Come Before

Legacy doesn’t mean outdated. It means earned.

Start by listening. Learn the story behind the processes, the technology, the culture. Deeply understand what’s been built, who built it, and why.

This isn’t about sentimentality, either. It’s about trust. Businesses rooted in decades of experience often carry customer loyalty, institutional knowledge, and operational resilience.

There’s untapped gold right there.

Leaders who honour the past aren’t looking back, they’re building forward with everyone on board.

Empathy isn’t a soft skill; it’s a power move.

“Transformation does not necessarily mean that one has to turn one’s back on values and practices developed by an organisation over several decades. Embracing change while reverently carrying the legacy … is a fine balancing act, but one that can mean the difference between sustained growth and innovation.”3

2. Map the Future: Quarterly Vision, Quarterly Action

Don’t show up with a five-year plan wrapped in buzzwords. Show up with a clear vision broken into tangible, quarterly progressions.

Use quarterly reviews not just for performance tracking, but for storytelling. Where did we start? What did we try? What did we learn? This rhythm of reflection and recalibration turns big changes into achievable journeys. It creates space for people to grow into the change, not resist it.

“Quarterly business reviews are vital for driving business growth because they provide a structured way to track performance, tackle challenges, and uncover opportunities … These reviews help you make informed decisions that drive progress and improve results.”4

Legacy transformation isn’t about pace. It’s about direction and discipline.

3. Communicate Like a Product Owner

“Empathetic leadership understands and addresses the fears, anxieties, and uncertainties that change can bring. By putting people first, businesses can foster a culture of trust and openness, paving the way for smoother transitions.”2

If you’re not talking, people are assuming and assumptions kill momentum.

Over-communicate. Break silos. Treat every stakeholder like a collaborator. Share your roadmap, your risks, your wins, and your rethinks.

And do it often.

Product ownership isn’t just for tech, it’s a mindset. Be transparent. Be obsessed with feedback loops. And make sure everyone knows where they stand in the journey.

When people feel seen and heard, they show up for the mission.

4. Make It Simple, Make It Beautiful

Legacy environments tend to grow in layers: one workaround stacked on another. Your job is to cut through that clutter.

Simplicity isn’t easy. It’s an act of design.

Whether it’s migrating to modern systems, integrating AI, or improving the way teams interact with tools, clarity wins. Great design isn’t just about look and feel. It’s about user trust, reduced friction, and signal over noise.

When you deliver something that just works, people remember how it made them feel: confident, empowered, capable.

“The greatest challenge for legacy migration is understanding and reconciling the current-state to the future-state … Legacy systems often have dated and undesirable architecture, a cumbersome user experience … The future-state should also consider long-term technical trends and feasibility as well as business considerations for digital product offerings and revenue streams.”6

5. Build With Diversity and Bring HR and Governance In Early

Speed and innovation thrive in diverse, cross-functional teams. Different disciplines, different backgrounds, different ideas: this is how new possibilities get unlocked.

“Engaging [employees] in the process of change ensures not only that the legacy is secured but also makes them a part of the organisation’s future … Innovating for a purpose, and building from past successes will help an organization negotiate many complexities of change.”3

But in legacy environments, moving fast doesn’t mean moving recklessly.

Engage HR early to align on people’s needs. Bring in governance to navigate access, compliance, and policy without roadblocks. These aren’t hoops, they’re accelerators if brought in early.

Diversity needs space to move. Your job is to clear the runway.

This Is the Work of a Rebel Technologist

It’s not about disrupting for disruption’s sake. It’s about evolving something important— – carefully, creatively, and courageously.

You’re not here to replace the old guard. You’re here to build on what’s been earned, with a vision that respects the past and leans hard into the future.

Respect deeply. Design simply. Communicate obsessively. Move purposefully.

That’s how you make a legacy business future-ready – one bold, empathetic step at a time.

But what exactly does democratisation of AI mean? At its core, it’s about accessibility.

Indiscriminate access to near limitless intelligence.

Making powerful AI tools affordable, usable, and adaptable for businesses of all sizes, from startups in garages to SMEs in office parks, and even curious individuals tinkering on their laptops is the name of the game.

AI’s Cost Revolution: Power to the People

Consider this insane fact: just a year and a half ago, leveraging AI at the level of GPT-3.5 cost roughly $20 per million tokens. Fast forward to late 2024, and that cost was cut down to just $0.07 – a 280-fold reduction.

This isn’t just cost-cutting; this is a revolution.

Companies like OpenAI have played a crucial role by consistently lowering prices, reducing barriers, and enabling anyone, anywhere to harness generative AI’s potential.

The implication?

Startups no longer have to choose between investing in AI or keeping the lights on. Entrepreneurs can leverage AI-powered insights without burning through their runway.

VentureBeat just posted about how American Express is capitalising on Generative AI to achieve 40% fewer IT escalations and an 85% boost in travel assistance.

“… Interacting with IT – particularly chatbots – can be a frustrating experience. Automated tools can offer vague, non-specific responses or walls of links that employees have to click through until they get to the one that actually solves their problem,” says the author, Taryn Plumb.

By infusing generative AI into its internal IT support chatbot, Amex finally put this problem to rest.

“The chatbot now interacts more intuitively, adapts to feedback and walks users through problems step-by-step. Amex has significantly decreased the number of employee IT tickets that need to be escalated to a live engineer.”

Open-Source: Leveling the Playing Field

Closed AI models have traditionally held the upper hand. Until now.

Open-weight models like Meta’s Llama are rapidly closing the performance gap.

In January 2024, open-weight models lagged behind their closed-weight counterparts by 8%. Just one year later, the difference shrunk to a mere 1.7%.

Open-Weight Models are also better for the environment:

• Increased Transparency and Accountability: Public scrutiny can help identify and address potential biases or ethical issues in the model.
• Facilitates Research and Innovation: Allows researchers to build upon existing work and develop new applications for the model.
• Promotes Collaboration: Encourages community-driven development and improvement of the model.
• Lower Barrier to Entry: Enables users to leverage advanced AI capabilities without the need for extensive development resources or specialized knowledge.
• Flexibility and Customisation: Users can adapt the model to their specific needs by fine-tuning the weights.

This convergence means innovation is no longer restricted to the select few with hefty budgets and exclusive access.

Today, innovation belongs to everyone willing to explore, experiment, and evolve.

Decentralisation: The Rise of Autonomous Networks

Blockchain isn’t just about cryptocurrencies, it’s also reshaping AI.

Decentralised AI networks, such as the Thames Network, put power back into the hands of individuals. These blockchain-driven systems emphasise privacy, autonomy, and censorship resistance, encouraging a future where AI isn’t dominated by a handful of tech giants, but instead driven collaboratively by an inclusive community.

A Shift from Training to Leveraging

Interestingly, Stanford’s influential AI Index in 2024 spotlighted rising training costs for models like GPT-4 (approximately $78 million).

But, what it missed was the rapidly declining cost of AI inference and the explosive rise of open-weight models. The focus is shifting from training prohibitively expensive frontier models to leveraging widely accessible and affordable AI technologies.

This shift is monumental, emphasizing practical innovation over mere bragging rights of size and scale.

What Does This Mean for Businesses?

The democratisation of AI opens a floodgate of opportunities for businesses:

• Innovation Acceleration: Smaller businesses can now experiment rapidly, iterate frequently, and innovate boldly, unencumbered by prohibitive costs.
• Tailored Solutions: Businesses can leverage customisable open-source models, allowing for solutions precisely tailored to their unique challenges.
• Competitive Advantage: Democratisation levels the playing field, enabling SMEs to compete against industry giants by smartly integrating AI into their operations.

Societal Impact: Inclusive Innovation

The broader societal implications are profound. Lower barriersmean increased innovation across sectors like healthcare, education, and the creative arts.

Individuals previously priced out of AI innovation can now become active contributors, innovators, and disruptors.

Yet, democratisation isn’t without its challenges:

• Privacy: Decentralised systems must balance openness with robust privacy controls.
• Standardisation: The lack of unified metrics can make evaluating different AI solutions complex.
• Ethical Considerations: Responsible use of democratised AI must be a priority to avoid perpetuating biases or misinformation.

So, What’s Next?

If you’re in leadership or innovation, it’s crucial to reflect:

• Are you leveraging the democratised AI tools now available?
• Are you building a team of adaptable, AI-savvy generalists capable of thriving in this accessible AI landscape?
• Are you fostering a culture of experimentation, agility, and ethical responsibility?

We’re past the age of exclusivity, and now fully in the imagination age. In this era, the currency of success is adaptability and democratised AI is its catalyst.

Final Thoughts from a Rebel

The future isn’t built by gatekeepers, it’s forged by rebels.

The democratisation of AI isn’t just technological advancement; it’s a clarion call for a new generation of rebel generalists and adaptable thinkers.

So, let’s lean into this messy, dynamic new world.

Let’s embrace AI’s democratisation to fuel creativity, innovation, and meaningful disruption. Because the future, more than ever, belongs to the curious, the adaptable, and the rebel technologists.

The AI Coding Revolution Is Already Here

If you haven’t paired with an AI coding assistant recently, you’re missing out on a paradigm shift.

Just yesterday, I watched as Copilot generated the core components of an authentication system from a few commented requirements. Three years ago, that would have taken hours—now, it’s done in minutes.

What’s happening isn’t just incremental, it’s exponential. These tools now:

• Complete complex functions with impressive accuracy (though still require review)
• Assist in refactoring legacy code that would make most developers cringe
• Translate well-defined business requirements into functional code (with some gaps)
• Help debug issues quickly, sometimes faster than senior engineers
• Generate test suites that catch common edge cases (but don’t replace human intuition)The game has fundamentally changed. Full stop.

Two Competing Futures: The Great Developer Divide

The developer community is split on where this leads us. Check out these two perspectives that capture the tension:

 

Laura Wendel’s visualization stopped me in my tracks when I first saw it. That S-curve perfectly captures what many of us are experiencing: simple apps have become trivially easy to build, while complex systems might actually become harder to develop as our understanding of foundational code erodes.

I call this the “knowledge atrophy risk”—when you outsource your thinking to AI, those muscles weaken. How many phone numbers do you remember now that they’re all stored in your smartphone?

On the flip side:

 

 

Vijay’s historical perspective resonates deeply with me. Every major abstraction in tech history( from punch cards to assembly to high-level languages to cloud computing) didn’t eliminate programmers. It elevated what we could build.

His comparison to MidJourney prompts is spot-on.

I’ve watched designers transition from fighting against AI to embracing it as a creative partner by learning to “speak its language.” The same evolution is happening with developers right now.

The Acceleration Is Mind-Blowing

Want to understand how fast this is moving? Listen to OpenAI’s Chief Product Officer, Kevin Weil:

“I mean, at the rate that we’re going, I would be surprised if it’s 2027. I think it’s going to be sooner. O1, when it launched, our first like real launch of a coding model was something like the thousandth best engineer in the world. Add competitive coding contests. O3, which is coming soon, according to the same benchmarks, is the 175th best competitive coder in the world. And as we are starting to train, you know, the successor models, they’re already better.

And so I think this is the year that at least by the competitive coding benchmark, this is the year that that AI becomes better than humans at competitive code forever.”

Let that sink in. We’re months, not years, away from AI outperforming the world’s best human competitive coders. Forever.

My Journey: From Skeptic to Strategic Adopter

I’ll be honest, I was initially defensive when AI coding tools emerged. “They’ll never understand the nuance of what we build,” I thought.

Then I watch these videos of developers use AI to accomplish in a day what would have taken me a week as a junior dev.

That’s when it clicked: fighting this wave is like arguing with electricity. The question isn’t if AI will transform programming, it’s how we’ll transform ourselves alongside it.

The True Value of Human Developers in an AI World

Despite these advances, I remain bullish on the future of human developers for three crucial reasons:

1. Problem Discovery Trumps Problem Solving

In my fifteen years building products, I’ve learned that identifying the right problem is infinitely harder than implementing a solution. AI excels at answering questions but struggles with asking the right ones.

When I sat with healthcare clients last month, understanding their workflow challenges required empathy, domain knowledge, and interpersonal intelligence that no AI possesses.

The coding was the manageable part.

2. System Thinking Can’t Be Prompted

Enterprise architecture requires understanding how complex systems interact: technical, human, regulatory, and business.

I recently watched a team spend weeks trying to use AI to generate a complex financial system, only to scrap it all when they realised the AI couldn’t grasp the regulatory nuances that made their business viable.

Some knowledge can’t be easily encoded in a prompt.

3. Innovation Requires Human Leaps

Every framework that AI can now generate was originally created by human developers making conceptual leaps. React, TensorFlow, Kubernetes – these weren’t inevitable evolutions; they were creative disruptions.

AI models are trained on existing patterns. They excel at interpolation, but struggle with extrapolation. The next paradigm shift will still come from human creativity.

How I’m Adapting My Team for the AI+Human Future

At our core, Rebel Technologists embrace change rather than fear it. Here’s how we’re evolving:

1. Reframing developer interviews: Instead of whiteboard coding exercises, we now focus on prompt engineering, system design, and problem framing. Can candidates direct AI effectively? That’s more valuable than memorising algorithms.
2. Investing in domain expertise: Our most valuable developers aren’t the fastest coders, they’re those who deeply understand our customers’ industries. Technical skills with domain knowledge is the unbeatable combination.
3. Building AI-native workflows: Rather than bolting AI onto existing processes, we’re designing workflows where humans and AI each play to their strengths. Our documentation, code review, and even architecture discussions now assume AI participation.
4. Creating intentional learning paths: With simple tasks automated, we’re crafting deliberate learning experiences to ensure developers still understand fundamentals. Like pilots who train without autopilot, developers need to know what’s happening under the AI’s hood.

Thriving in the New Reality

For individual developers navigating this shift, my advice is straightforward:

• Become the bridge between business problems and technical solutions
• Master prompt engineering as a core skill
• Build your domain expertise in at least one vertical
• Develop your system thinking abilities
• Lean into what makes you uniquely human: creativity, empathy, and ethical reasoning

The most dangerous place to be is in the middle: better than a beginner, but not as good as an expert. AI will squeeze that middle first, driving developers either upmarket (strategic thinking) or downmarket (implementation specialists guiding AI).

The Future I See Coming

I believe we’re entering the most exciting era of software development in history. The code we’re writing today will seem as primitive as assembly language in a few years.

As barriers to entry fall, we’ll see an explosion of software creation from previously excluded populations. The next great developer might be a nurse who uses AI to build healthcare tools no traditional programmer would think to create.

The partnership between humans and AI will unlock applications we can barely imagine—systems that adapt in real-time to changing conditions, interface with humans through natural language, and solve problems too complex for either humans or AI alone.

This isn’t the death of programming. It’s programming’s renaissance.

Will AI replace programmers? No. But programmers who use AI will absolutely replace those who don’t. The future belongs to rebels who embrace this transformation and ride the wave rather than being swept away by it.

What’s your experience building with AI so far? Has it changed how you approach development?

Drop your thoughts in the comments! I’m fascinated to hear how other technologists are navigating this shift.

*Check out the original post: https://www.teraflow.ai/will-ai-replace-programming-the-future-of-developers/