AI is evolving at a rapid pace, and the conversation surrounding its potential impact on jobs is growing louder. While recent developments in artificial intelligence suggest significant advancements, their real-world impact on employment remains uncertain. Are we on the cusp of widespread automation, or will AI generate new opportunities? Here's what the data shows and what experts are saying.

AI’s exponential progress in software development

A chart produced by a nonprofit organization called Meta recently set the tech world buzzing. It tracks the improvement of AI systems in performing software development tasks. For years, progress was steady but underwhelming. However, by late 2025, the field began to see significant advancements. Since then, AI models like Gemini 3 Pro, Claude Opus 4.5, and GPT-5.2 have shown remarkable progress in their ability to handle coding and other technical tasks.

Interestingly, what feels like a sudden leap in capability is actually the result of exponential growth. As Mark Warner, a prominent AI researcher, and entrepreneur based in London explains, “The lived experience of exponential growth looks like nothing, nothing, nothing, and then everything.” While consumer-facing AI tools like chatbots have improved steadily, it’s among software developers using AI to generate code where these leaps have been most noticeable.

On a logarithmic scale, which more accurately depicts exponential trends, the evolution of AI has been steady rather than erratic. It is this ongoing development — not any single breakthrough — that underscores how transformative AI could be in accelerating technical work.

Is AI improving itself?

One of the most alarming concerns surrounding AI is whether systems are now capable of writing their own code — potentially leading to runaway self-improvement. While this scenario feels like science fiction, industry experts believe this isn’t happening yet.

Joel Becker from Meta notes that while current AI models are not directly improving themselves, they are dramatically speeding up the pace at which researchers and developers can refine and enhance these systems. This accelerated progress suggests that improvements are likely to continue at a consistent, if not increasing, pace, fundamentally altering how new technologies are developed.

Jobs at risk, or new demand?

The most pressing question remains: what does AI’s evolution mean for jobs? Historically, technological advancements have often paved the way for new industries and roles, and AI could follow a similar trajectory. Current employment data paints a surprisingly ambiguous picture — contrary to fears of AI-induced mass job losses, software development job postings in the U.S. are actually on the rise.

According to Indeed, the job-listing site, the recent spike in tech hiring had more to do with the COVID-era boom in remote work rather than AI. Over the last year, job postings in software development have stabilized, with no clear indication that AI is leading to automation-driven layoffs in the sector. Instead, AI-focused organizations like Anthropic, Meta, and other leading tech firms are actively hiring software developers.

Past lessons: Are we the horses?

To better understand AI’s trajectory, it’s worth revisiting a historical analogy. A chart comparing the efficiency of early steam engines to the population of horses in the U.S. demonstrates that the steady improvement in engine efficiency didn’t affect the reliance on horses for some time. However, once engines hit a tipping point, the use of horses in transportation plummeted.

This “horse analogy” raises unsettling questions: Are humans today, particularly in certain roles, comparable to horses during the industrial revolution? Will AI reach a tipping point where its impact on jobs becomes irreversible?

The short answer is unclear, as there are critical caveats. AI’s capabilities, measured in simulated environments, are not always indicative of real-world performance. Tasks AI can complete with 50% reliability may not translate well into economically valuable labor. Real-world work often involves a level of messiness and complexity that AI struggles to replicate.

The Jevons paradox and new growth

Another perspective comes from historical economics, specifically the Jevons paradox. It explains how increases in efficiency often lead to greater demand for a resource, not less. In the case of the steam engine, as its efficiency improved, coal consumption skyrocketed rather than decreased, as more industries adopted it.

Applying this logic to AI, the technology could increase demand for human-led innovation and application rather than eliminate jobs. Companies leveraging AI might require even more software developers, engineers, and other specialists who can implement these tools effectively. Career growth in fields directly tied to AI is already evident, as major firms continue expanding their recruitment efforts.

Beyond tech: The broader job market

While software development may be somewhat insulated for now, other industries face greater uncertainty. CEOs have floated the idea of automating fields like accounting, law, and consulting within the next 18 months. Yet, the complexity and variability of these professions make immediate replacement unlikely. Real-world jobs are far less standardized and predictable than the controlled environments in which AI systems are tested.

AI might excel at completing specific tasks, but handling the ambiguities and nuances of human-centric professions remains a significant hurdle. Real-life work often involves interpersonal interactions, ethical considerations, and unstructured problem-solving — areas where AI still lags.

Preparing for an AI-driven future

The general consensus among experts is that artificial intelligence is coming down the tracks fast, and businesses, governments, and individuals need to adapt. Whether that means reskilling for roles in AI-related fields, rethinking education to emphasize creativity and human-centric skills, or crafting policies to ensure equitable transitions, preparation will be key. The degree to which AI disrupts or complements human labor depends on how quickly organizations and workers adapt to these changes.

Key takeaways:

• Exponential growth may feel sudden but is steady: AI’s rapid evolution is rooted in consistent, exponential development over time.
• AI hasn’t hurt software development jobs — yet: Employment data shows stability in tech hiring, even as AI tools improve.
• Economic shifts may follow historical patterns: Lessons from the industrial revolution suggest new growth and opportunities often accompany labor transitions.
• Soft skills are harder to replicate: Complex roles like consulting and law may require more time to automate fully.

Conclusion

AI is undoubtedly accelerating, with major implications for work and society. Its impact on jobs is already sparking debate, but as recent history shows, the outcomes will likely be more nuanced than a simple narrative of replacement. Whether AI leads to massive disruption or creates new opportunities depends largely on how industries and individuals respond to its growing capabilities. One thing is certain: the AI revolution is here, and understanding its trajectory has never been more critical.