Why the AI Future Isn't as Dark as Portrayed

Optimism Amid the Revolution

In a recent article on X by Matt Shumer, the CEO of HyperWriteAI, he paints a stark picture of AI's rapid advancement leading to massive disruptions, including widespread unemployment and societal upheaval. While his piece is thought-provoking and urges preparation for big changes, I believe the reality won't be as bleak as portrayed. Like every major technological revolution in history, AI will create far more opportunities than it destroys. The path to fully autonomous robots and true Artificial General Intelligence (AGI) remains long and fraught with challenges, and even the development of Large Language Models (LLMs) appears to be approaching a plateau, where simply throwing more parameters at the problem doesn't yield proportional gains in "intelligence." In this post, I'll elaborate on these positive sentiments, backed by evidence from historical precedents and current trends.

Lessons from History: Revolutions Create Jobs, Not Just Destroy Them

Technological revolutions have always sparked fears of mass job loss, but history shows they ultimately expand the economy and generate new employment. Take the Industrial Revolution, starting in the late 18th century. The shift from manual labor to mechanized production displaced many artisans and farm workers, but it also birthed entirely new industries. For instance, the steam engine revolutionized transportation and manufacturing, leading to a boom in urban jobs. In England and Wales, census data from 1871 onward reveals that technology didn't obliterate work; instead, it quadrupled the number of bar staff since the 1950s and surged hairdresser roles in recent decades, as increased productivity boosted spending power and created demand for new services.

Similarly, the Second Industrial Revolution in the late 19th and early 20th centuries introduced electricity, the combustion engine, and mass production. While it automated tasks like welding in car manufacturing, it spurred massive growth in related fields. The Ford Model T assembly line is a classic example: it reduced the need for skilled craftsmen but created thousands of jobs in factories, supply chains, and even leisure industries like skiing, which exploded post-World War II thanks to better transportation and higher incomes. Overall, these shifts didn't shrink the workforce; they transformed it, with net job gains as economies grew.

Fast-forward to the digital revolution of the late 20th century. The rise of computers and the internet displaced typists and elevator operators, but it created roles in software development, web design, and data analysis. From 1940 to 1980, automation eliminated some routine jobs, yet new tasks emerged, leading to overall employment growth. This pattern holds: technology displaces, but it also augments human capabilities, fostering innovation and economic expansion.

The AI Boom: New Jobs on the Horizon

Applying this to AI, we're already seeing a similar dynamic. While AI automates repetitive tasks like data entry or basic customer service, it's creating a surge in demand for new skills. According to a McKinsey report, AI could generate 20-50 million new jobs globally by 2030 in sectors like healthcare, manufacturing, and finance. Roles such as AI ethicists, machine learning engineers, and data curators are exploding—think prompt engineers who fine-tune AI outputs or specialists in AI-driven analytics.

Evidence from recent studies supports this optimism. Firms adopting AI report not just task replacement but overall growth: a large increase in AI use correlates with 6% higher employment growth and 9.5% more sales over five years. PwC's 2025 Global AI Jobs Barometer analyzed nearly a billion job ads and found that industries most exposed to AI experience twice the wage growth and three times the revenue per employee compared to less-exposed ones. Even in automatable roles, wages rise as workers become more valuable alongside AI.

Sure, there will be transitions—younger workers in high-AI-exposure fields like software might face short-term dips. But overall, the data points to augmentation over destruction. The World Economic Forum predicts AI will displace 75 million jobs by 2025 but create 133 million new ones, netting 58 million gains. This isn't doom; it's evolution.

The Long Road to Fully Autonomous Robots and AGI

One reason I'm optimistic is that the hype around imminent AGI and robot takeovers overlooks massive hurdles. AGI—AI that matches human intelligence across all domains—requires breakthroughs in common sense, intuition, and adaptability that current systems lack. For example, today's AI excels at narrow tasks but struggles with context, causality, and real-world reasoning. Experts highlight five tough challenges: building common sense, enabling transferable learning (applying knowledge from one domain to another), bridging the "phygital" divide (integrating digital AI with physical worlds), scalability issues, and trust in opaque systems.

Computational limits are a big barrier. AGI demands vast resources beyond current GPUs, plus advances in neuroscience and cognitive science to replicate human-like cognition. Visual and auditory processing remain weak; models falter on tasks like mental rotations or understanding sequences over time. Even continual learning—updating knowledge without forgetting old info—is unsolved.

For robots, full autonomy is even farther off. Challenges include dexterity for handling unpredictable environments, energy efficiency, and safe human interaction. We're talking decades, not years, giving society time to adapt.

LLM Development: Hitting a Plateau?

Finally, the core of AI progress—LLMs like those powering ChatGPT—shows signs of diminishing returns. Early scaling laws suggested bigger models (more parameters) equaled smarter AI, but recent evidence indicates that's tapering off. Simply adding parameters doesn't linearly boost "intelligence"; instead, we're seeing smaller gains per increase.

A PNAS study on political persuasion found that scaling model size yields diminishing returns—even orders-of-magnitude jumps don't significantly improve single-message effectiveness. TechCrunch reports that AI scaling laws are showing fatigue, forcing labs to pivot from brute-force growth. Issues like "model collapse," where training on AI-generated data degrades quality, compound this. As Gary Marcus notes, LLMs have reached diminishing returns, with grim economics for endless scaling.

This plateau means progress will slow, allowing humans to integrate AI as a tool rather than a replacement. It's a positive sign: AI enhances us, but it won't outpace us overnight.

Conclusion: Embracing the Positive Side of Change

Matt Shumer's article is a wake-up call, but I see a brighter path. History proves revolutions create abundance, AI is already birthing new jobs, and technical barriers to AGI/robots ensure a gradual transition. With LLMs plateauing, we have breathing room to upskill and innovate. Let's focus on the opportunities—invest in education, ethics, and equitable access—to make this revolution one of prosperity for all.

What do you think? Share your thoughts in the comments or on X @ralf100.

Links: https://www.theguardian.com/business/2015/aug/17/technology-created-more-jobs-than-destroyed-140-years-data-census https://www.mckinsey.com/featured-insights/future-of-work/what-can-history-teach-us-about-technology-and-jobs https://news.mit.edu/2024/does-technology-help-or-hurt-employment-0401 https://www.innopharmaeducation.com/blog/the-impact-of-ai-on-job-roles-workforce-and-employment-what-you-need-to-know https://mitsloan.mit.edu/ideas-made-to-matter/how-artificial-intelligence-impacts-us-labor-market https://www.pwc.com/gx/en/services/ai/ai-jobs-barometer.html https://www.adpresearch.com/yes-ai-is-affecting-employment-heres-the-data/ https://www.innopharmaeducation.com/blog/the-impact-of-ai-on-job-roles-workforce-and-employment-what-you-need-to-know https://www.forbes.com/sites/bernardmarr/2025/03/13/beyond-chatgpt-the-5-toughest-challenges-on-the-path-to-agi/