Is AI growing faster than we can keep up?
The truth is, yes, artificial intelligence is growing faster than we can keep up. On nearly every measurable dimension, the data suggests this.
AI models are improving rapidly at handling real-world tasks.
METR, which monitors AI model progress, found that the duration of tasks AI agents can complete autonomously had roughly doubled every seven months over the prior six years.
More recently, that doubling time shortened to approximately four months in 2024–2025. By late 2025, the most capable models were reliably completing tasks that would take a skilled professional five hours to do.
In terms of cost efficiency, report shows that the amount of model capability available per dollar has been doubling roughly every 3–4 months for Google's flagship models, and every 6–8 months for OpenAI's. Between early 2023 and late 2025, cost-adjusted performance of frontier models increased by more than an order of magnitude.
This means that for every dollar spent, users can do much much more than they could in 2023.
This development reprices labour and accelerates time-to-completion of tasks.
AI tools now reach approximately 1.5 to 2 billion people and investments into AI products and solutions are expected to reach over $3.3 trillion by 2029.
This trend is outpacing workforce readiness.
Here's where the "can't keep up" factor becomes evident in this story. Organizations are racing to deploy AI, but people are simply not ready for the impact, not on jobs, not even on a personal level and this is both the positives and the potential negatives.
Over 90% of global enterprises are projected to face critical AI skills shortages by 2026. IDC estimates that skills gaps may cost the global economy up to $5.5 trillion in product delays, quality issues, missed revenue, and lost competitiveness. While 94% of CEOs identify AI as their top in-demand skill, only 35% of leaders feel they have prepared employees effectively for AI roles.
Claude Fable 5
Beyond all that's already been said in the past months and years, Anthropic recently launched its most capable model yet, just after being in the news for asking for a "pause" on AI developments from global labs.
Fable 5 is it's most recent and most powerful model, a so-called Mythos-class model with exceptional performance in software engineering, knowledge work, vision, scientific research, and other areas.
So far, there's a strong sentiment of this being the closest thing to AGI given it's performance across public testing since launch. Right now, a more powerful version called simply Mythos 5 is accessible to trusted partners, particularly those that make up Project Glasswing, an initiative from Anthropic that gives Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks access to its most capable models in an effort to secure critical software infrastructure.
What to expect?
Innovations cannot simply pause for people to catch on, so it's only inevitable that development will outpace adjustments necessary across industries and on a personal level.
So we can expect pain in certain parts of world economies but provided that AI growth doesn't conveniently and at a low cost, hurt critical systems and economies, on an exploit level, then long-term benefits of this developments should exceed temporary pain impacts.
Posted Using INLEO