Tobi Lütke, CEO of Shopify, recently sparked widespread discussion with a tweet declaring, "I shipped more code in the last 3 weeks than the decade before. The top AI models / agentic systems right now are an entirely different thing to what people used until the beginning of December."
This statement, posted on January 8, 2026, encapsulates a seismic shift in software development and beyond, driven by advancements in AI agents like those from Anthropic's Claude and similar systems.
Over the past few weeks, as AI tools have evolved rapidly, several truths have crystallized across the tech community. This article explores these insights, supplemented with recent data on AI's transformative impact, to illustrate how coding — and knowledge work at large — is being revolutionized.
The End of the "Slop-Code" Era
Gone are the days when AI-generated code was dismissed as unreliable "slop" — low-quality output riddled with errors. Today, if subpar code emerges, it's often due to user laziness or inexperience rather than inherent system flaws.
McKinsey's 2025 State of AI report highlights this progress: 64% of organizations report AI enabling innovation, with agentic use cases in IT and knowledge management leading the charge.
Tools like Claude Code have advanced to handle complex tasks with high fidelity, reducing technical debt through iterative refinements. As one X user noted in response to similar discussions, "AI didn't lower the bar — it removed the hiding spots," exposing those who rely on superficial knowledge while empowering true system thinkers to ship faster.
This shift is evident in productivity gains. Shopify itself has integrated AI deeply since Lütke's April 2025 memo mandating AI use as a "baseline expectation," leading to employees tackling "implausible tasks" with 100x efficiency.
Gartner echoes this: By 2026, 40% of enterprise applications will embed task-specific AI agents, up from less than 5% in 2025, signaling a move toward high-quality, automated outputs.
Democratization of Software Creation
Anyone with desire and focus can now build software — no advanced degree required. AI removes technical barriers, but it doesn't solve the "why?" — motivation and persistence remain human domains. As developer Omar Khattab put it on X, the goal isn't "vibe coding" but higher abstractions that make creation more accessible without sacrificing quality. This has led to a boom in hobbyist and personal apps, with beginners using tools like Claude to prototype ideas in hours.
Data supports this: McKinsey notes that 88% of organizations use AI in at least one function, but only 1% achieve maturity, highlighting the gap between access and mastery. Yet, for individuals, the barrier drop is profound — enabling solo builders to create what once needed teams.
100x Advantages for AI-Native Organizations
Top organizations like Shopify are gaining massive edges through AI-native processes, while laggards risk irrelevance.
Lütke's experience underscores this: His productivity surge reflects Shopify's culture of AI integration, where tools like the open-source Roast framework orchestrate feedback loops for code quality. Gartner predicts that enterprises adopting AI will outperform others by at least 25% by 2026, particularly in knowledge management.
Conversely, 95% of AI pilots yield zero ROI due to poor integration, per MIT's 2025 findings. Success hinges on redesigning workflows, as seen in Shopify's hiring policy: Prove AI can't do the job before adding humans.
The Ripple to All Knowledge Work
What started with code will permeate sales, law, finance, content, logistics, management, and science within a year. McKinsey forecasts agentic organizations with 50-100 agents per person in mature setups, transforming functions like deep research and service-desk management.
Gartner adds that 60% of knowledge workers will use generative AI daily by 2027, with neutral job impact through 2026 as AI augments rather than replaces.
In legal fields, tools like Thomson Reuters' CoCounsel cut review times by 6-8 hours daily; in marketing, AI frameworks automate positioning in minutes.
Individuals as AI-Powered Corporations
Every person is now a "corporation," as explored in a viral YouTube video: You're the CEO, with agents handling finances, emails, and more. Governance is key — define roles, permissions, and culture to avoid "collective stupidity." This mirrors historical collectives like guilds, blurring lines between individuals and companies.
Companies as Cybernetic Entities
Firms evolve into cybernetic organizations, with agents outnumbering humans. Management shifts from hierarchies to networks, more disruptive than traditional IT transformations.
McKinsey identifies new archetypes: M-shaped supervisors for orchestration, T-shaped experts for exceptions. Deloitte projects the agent market at $45 billion by 2030.
Essential Skills: Agent Systems and Personal Agency
Building personal agent systems is crucial; agency—drive to act—remains paramount amid abundant opportunities. As one X commenter noted, "AI just made limitations visible," rewarding those who debug and understand systems.
Software Commoditization
Software loses uniqueness; margins drop as users build alternatives at cost. Gartner forecasts 55% of new ML models via automated pipelines by 2026.
Hardware Through an Information Lens
Hardware is informational processes: Hiring, procurement, logistics—all agent-orchestrable. Robots are software incarnate, shifting focus from atoms to workflows.
Also read:
- Claude Cowork: Anthropic's AI Agent for Streamlining Non-Technical Workflows
- Netflix Turns Tarot into a Streaming Prophecy: Grand Central’s Mystical Reveal of 2026’s Biggest Shows and Films
- Google Vids' AI Avatars: Powered by Veo 3.1 for Surprisingly Polished Talking Heads
Conclusion: Seize the AI Edge
Lütke's tweet signals a new era where AI amplifies human potential. With 92% of companies boosting AI investments, the divide grows between adapters and laggards.
Lütke's tweet signals a new era where AI amplifies human potential. With 92% of companies boosting AI investments, the divide grows between adapters and laggards.
