Mark Zuckerberg predicts that AI will write most of Meta’s code in 12 to 18 moths. Microsoft’s CEO Satya Nadella said that AI is already writing 20%-30% of Microsoft’s code… but it’s not just coding that is getting automated. According to Goldman Sachs, 300 million jobs are at risk of being automated by AI, a United Nations report estimates that AI could disrupt over 40% of roles globally and McKinsey estimates that 30 percent of hours currently worked across the US economy could be automated.
A clear trend is emerging among major tech companies: a move towards becoming “AI-first.” This shift, heralded by CEOs like those at Duolingo, Shopify, and Fiverr, isn’t just about integrating AI tools; it represents a fundamental rethinking of how businesses operate, with significant consequences for the workforce and potentially the investment landscape.
High Profile CEOs are publicly stating their companies are ‘AI-First’
Duolingo is explicitly betting on AI as a platform shift, similar to the move to mobile in 2012. CEO Luis von Ahn announced in a companywide email that Duolingo would become “AI-first”. This means rethinking much of how they work, as minor tweaks to systems designed for humans won’t be sufficient. As part of this strategy, Duolingo plans to gradually stop using contractors for work that AI can handle.
Here’s the memo he shared with employees:
AI use will also become a factor in hiring and performance reviews, and headcount will only be approved if a team cannot automate more of its work. While von Ahn states the move is not about replacing employees (“Duos”) but about removing bottlenecks to allow them to focus on creative work and real problems, it undeniably impacts the future of contracted work and the criteria for internal roles. Duolingo sees AI as crucial for scaling content creation and building new features like Video Call, which were previously impossible, bringing them closer to teaching as well as human tutors.
Shopify CEO Tobi Lütke has implemented a new hiring policy that prioritises AI over human hires. In a memo shared publicly, Lütke stated that teams must demonstrate why their needs cannot be fulfilled using AI before requesting additional resources. This approach is part of Shopify’s commitment to integrating AI into its operations. This shift follows significant layoffs at Shopify in recent years, including 20% of staff the year prior and quiet layoffs in customer service. While some see this strategy as potentially extreme or a PR stunt, it reflects the growing adoption of AI for efficiency.
Micha Kaufman, founder and CEO of Fiverr, issued a stark warning about the “coming AI jobocalypse”. His message to staff and freelancers is clear: “AI is coming for your job“. He explicitly states this applies to a wide range of professions, including programmers, designers, product managers, data scientists, lawyers, customer support representatives, salespeople, and finance personnel. Kaufman emphasizes that tasks previously considered ‘easy’ will disappear, ‘hard tasks’ will become the new easy, and ‘impossible tasks’ will become the new hard.
You must understand that what was once considered an ‘easy task’ will no longer exist; what were considered ‘hard tasks’ will be the new easy, and what were considered ‘impossible tasks’ will be the new hard.
Micha Kaufman, Fiverr CEO
To survive this “AI onslaught,” he suggests becoming exceptional and staying afloat by mastering the latest AI in your domain, studying prompt engineering, proactively pitching ideas, and creating opportunities rather than waiting for them.
The Proof is not in the pudding… yet
While companies are racing to adopt AI and restructure operations around it, the actual economic impact of AI remains limited. In a recent experiment by Carnegie Mellon called TheAgentCompany, advanced AI agents placed in a virtual startup environment were tested for task performance. The result? The top agent (based on Claude’s 3.5 model) completed only 24% of its assigned tasks. Researchers concluded there’s still a large gap between AI’s benchmark performance and its ability to carry out real-world jobs reliably.
Much of the current push feels driven more by pressure to appear innovative and cut costs than by proven AI productivity gains. For many businesses, sweeping AI integration without clear ROI could do more harm than good. Another approach that is rising is the hybrid approach: rather than AI replacing people (and jobs) it’s seen as a power multiplier.
Aaron Levie, CEO at Box, sees a middle ground, where AI is turbocharging employees with AI Agents. He shared an internal memo publicly which describing how enterprises going “AI-first” will see AI Agents working 24/7 across the business. This means knowledge work, historically confined to specific hours, will get done continuously, accelerating business decisions and enabling more effective outcomes. He notes that he personally uses AI agents to kick off research tasks in the evening and wakes up to the work completed, compressing hours of work. This concept implies that “knowledge work stops sleeping”.
Uncharted territory ahead
There’s no doubt in my mind that implications of rising automation for jobs are profound. The warnings from CEOs like Kaufman and the policies at Duolingo and Shopify underscore the potential for AI to replace tasks and even roles, particularly those that are repetitive or considered ‘easy’.
For Venture Capital, this pervasive shift implies changes in the landscape for investment. The need for new infrastructure and tools for AI agents, and the potential for complex infrastructure around distilling large AI models into more efficient forms all point to potential areas of significant growth and investment.
In conclusion, the “AI-first” announcements from companies like Duolingo, Shopify, and Fiverr are not isolated events but reflections of a broader, accelerating trend. Companies are fundamentally restructuring operations around AI capabilities, leading to significant implications for the workforce, demanding adaptation and upskilling. While this presents challenges regarding job displacement, it also creates opportunities for those who can effectively work with and build AI technologies. For investors, this transformation signals a need to understand how AI is redefining industries and identify the technologies and business models that will thrive in this new “AI-first” era.
