A well-known researcher stirred quite a debate on X after unveiling a new startup with an ambitious and highly controversial goal: replacing all human workers with AI systems. Tamay Besiroglu has been an influential figure in major AI advancements over the past decade. He claims that Digital Labor is ready to take over nearly every job currently done by people, from customer service and logistics to medicine and education – the startup aims to disrupt most industries with AI automation.

In a post on X, Besiroglu laid out an eye-popping total addressable market by aggregating all current human wages – approximately $18 trillion in the U.S. and over $60 trillion worldwide – positioning Mechanize to capture that entire sum through digital labor replacement

The announcement immediately provoked a firestorm of criticism. Critics argue that the startup, named Mechanize, raises serious ethical, economic, and social concerns. Labor advocates warn of mass unemployment and a widening gap between tech developers and the general workforce. Others worry about the long-term consequences of relying on machines to perform roles that require judgment, empathy, or human connection.

Key Takeaways

• On April 19, 2025, AI researcher Tamay Besiroglu launched Mechanize with the stated goal of “full automation of all work” and the complete replacement of human labor by AI agents across every industry.
• Besiroglu estimates the startup’s total addressable market at roughly $18 trillion in U.S. wages and over $60 trillion globally. Besiroglu also clarified that Mechanize will focus on automating white-collar roles in its initial phase rather than manual labor requiring robotics.
• Just after the announcement, it provoked a massive debate – X users warned of widespread job displacement and reignited scrutiny of Besiroglu’s dual role at nonprofit research institute Epoch after its December admission of undisclosed OpenAI support for one of its AI benchmarks.
• Besiroglu argues that AI-driven automation could deliver “explosive economic growth,” raise living standards, and diversify human income through higher complementary-role wages, rents, dividends, and welfare; however, critics question whether an economy devoid of traditional wages can sustain consumer purchasing power.

Mechanize to Redefine Work Through Full-Scale Digital Labor Automation

In April 2025, Tamay Besiroglu, a renowned AI researcher and co-founder of the non-profit research organization Epoch, announced the launch of Mechanize. This bold new Silicon Valley startup is established with the target of achieving “the full automation of all work” and eventually “full automation of the economy.” The announcement, made via a widely shared post on X, instantly reverberated through the AI community and mainstream media, causing both intrigue and alarm over the practical and ethical dimensions of replacing every human worker with autonomous software agents.

A year into the age of AI agents, their shortcomings – such as unreliable performance, inadequate memory retention, and difficulty with long-horizon planning – have become apparent, highlighting the formidable technical hurdles Mechanize must overcome to realize its ambitions. Mechanize enters the AI agent automation scene at a time when tech giants and startups are in cutthroat competition to deploy autonomous systems for tasks such as updating CRMs, handling outbound sales, and conducting financial analysis.

Approximately 92% of enterprises plan to increase AI investments over the next few years, and Mechanize is buckling up to meet the growing demand. Mechanize distinguishes itself by not merely automating components of work but by aiming to subsume all human labor under a unified digital framework.

Besiroglu’s reputation was forged through his work at Epoch, a non-profit AI research institute he founded in 2021 to develop standardized testing platforms for evaluating model capabilities, earning a reputation for objectivity among frontier researchers. Epoch’s performance evaluations became a go-to reference for labs like OpenAI, whose open support for one of Epoch’s late-2024 benchmarks triggered a debate over transparency and impartiality in AI assessment. The launch of Mechanize has intensified scrutiny on the relationship between academic research and commercial ventures, as critics question whether Besiroglu’s dual roles might compromise the independence of future benchmarkings.

According to its founding announcement, Mechanize will construct virtual work environments, evaluation benchmarks, and comprehensive training datasets designed to capture the full spectrum of human job functions—from routine data entry and email composition to complex project coordination and decision-making under uncertainty. The startup categorizes human work into discrete task archetypes and designs corresponding simulation modules to train agents capable of reprioritizing in the face of obstacles and collaborating with human coworkers or other agents in multi-agent settings. With the help of this, Mechanize will accelerate the model training, evaluation, and deployment cycle, lowering the time and cost barriers that currently constrain enterprise automation projects.

While Mechanize’s long-term vision is to target the full economy, its starting point for the roadmap concentrates on automating white-collar occupations where AI can interact with digital interfaces without requiring advances in physical robotics. It allows the company to create high-fidelity virtual workstations—complete with mock applications, databases, and communication channels—effectively emulating office workflows without significant hardware investment. Besiroglu contends that by perfecting office automation, Mechanize can establish a blueprint for subsequent phases that integrate physical robotic actuators to tackle manual labor tasks once hardware matures.

Through a tweet, the startup’s founders calculated Mechanize’s total addressable market by aggregating global wage expenditures, estimating roughly $18 trillion per year in the United States and over $60 trillion worldwide – an enormous scale of potential return on investment for comprehensive automation. The startup frames human labor as an addressable technology market, the team challenges traditional economic assumptions about labor scarcity, and proposes treating digital workers as an investable asset class. They argue that enterprise clients will swiftly adopt AI agents if cost-per-task metrics outperform human wages. It unlocks new efficiency frontiers and reshapes corporate budget allocations.

Mechanize’s seed round attracted prominent backers, including Nat Friedman, former GitHub CEO; Daniel Gross, serial AI entrepreneur; Patrick Collison, Stripe co-founder; Dwarkesh Patel; Jeff Dean, Google AI pioneer; Sholto Douglas; and Marcus Abramovitch, managing partner at AltX. Abramovitch praised the founding team’s depth of expertise, describing them as “exceptional across many dimensions” and commending their “deeper thinking” on AI systems than any peers he knew.

Public Reaction to Mechanization Highlights Deepening Fears Over Job Displacement and AI’s Societal Impact

The announcement made by Mechanize on X sparked fierce reactions, where users expressed respect for Besiroglu’s prior work yet lamented the prospect of mass job displacement. Prominent X user Anthony Aguirre wrote, “Huge respect for the founders’ work at Epoch, but sad to see this… a giant loss for most humans.” Meanwhile, on LinkedIn, commenters questioned how society would allocate power and resources once human effort was deemed obsolete, with some denouncing the startup’s vision as dehumanizing.

Concerns about job loss and economic inequality were echoed in broader public opinion polls. Recent data indicates that 54% of respondents believe AI poses a ‘significant risk’ of large-scale unemployment, with technology professionals (58%) expressing the most concern, followed by educators (52%) and healthcare workers (48%). The survey sample included participants from over 20 countries and spanned multiple demographic cohorts. These results highlight the tension between AI’s promise of efficiency gains and the very real fear of obsolescence felt by workers across industries.

While many social media users decried Mechanize’s ambition as a threat to livelihoods, others argued that focusing on specialized AI capabilities, rather than pursuing unfettered general superintelligence, could drive safer, more tractable innovations.

Economic historians and labor economists caution that AI-driven automation typically reconfigures task portfolios rather than wholesale eliminating occupations. MIT economist David Autor notes that over the past two centuries, mechanization propelled labor out of agriculture and manufacturing into higher-value service and creative work, all while enhancing overall productivity and incomes. Nonetheless, Autor warns that AI’s diffuse impact could exacerbate inequality if left unchecked, given its capacity to replace both routine and non-routine cognitive tasks. He advocates for intentional policy design to empower non-elite workers and integrate AI advances in education and healthcare applications.

Performance Limitations Persist as AI Agents Expand into Routine Enterprise Tasks

Despite these aspirations, current AI agents exhibit notable performance gaps. “For instance, Simular AI’s S2 agent—combining a large language model with specialized task-specific modules—achieved only a 34.5% completion rate on complex computer tasks and 50% on smartphone workflows. These benchmarks fall well below human proficiency; it means that even ‘state-of-the-art’ hybrid architectures struggle with contextual understanding and still require extensive human oversight for reliable execution.

Industry leaders are racing to overcome these challenges. Platforms like Salesforce’s Einstein GPT and Microsoft’s Power Automate are integrating LLM-based planning with enterprise API workflows to offer turnkey agentic automation solutions. Smaller startups are complementing these efforts with specialized tools for synthetic data generation, anonymized model training, and secure deployment environments to meet stringent compliance requirements.

Sectors like customer service, appointment scheduling, document review, and routine financial reporting—where tasks are largely rule-based and interface-driven—stand to be among the first disrupted by Mechanize’s approach. In theory, these replacements could liberate employees to focus on higher-order strategic and creative activities, but the speed and scale of this transition may strain existing workforce retraining programs. Besiroglu argues that displaced labor could pivot to roles collaborating with AI, deriving income from complementary human-agent partnerships.

This vision echoes familiar science fiction ideas about a future where automation eliminates the need for most labor, allowing people to focus on knowledge, creativity, and leisure. However, critics argue that this view overlooks the deeper social and psychological functions of work, including its role in shaping identity, building community, and providing purpose. Some have compared Mechanize’s messaging to the post-scarcity economy depicted in Star Trek, but note that such a transition would require major changes to property rights and the structure of modern economies.

Meeting these challenges will take more than new technology. Economists point to the need for strong social policies, such as universal basic income or job guarantees, and tax reforms to ensure that the benefits of automation are broadly shared. There is also growing pressure to set ethical standards and prioritize human-centered design in AI development to avoid outcomes that deepen inequality.

Mechanize’s ties to Epoch raise concerns about independence and potential conflicts of interest, especially as benchmarking standards play a growing role in shaping investment and regulatory choices. This makes transparency in governance an important issue.

In the short term, Mechanize is actively hiring (as noted in the subsequent tweet) in areas like engineering, AI research, and virtual simulation. The company is moving quickly to build and test its digital workspace and benchmarking tools. Its public updates highlight a push to recruit individuals who want to work on redesigning how labor is structured.

Whether Mechanize succeeds or overreaches, its actions mark a key point in the development of AI-driven automation. The discussions it has triggered—around technical limits, fair economic outcomes, and accountability—highlight the need for input from multiple sectors. As the company shifts from concept to implementation, it will face scrutiny from industry, policymakers, and academics, all watching to see whether full automation of labor is realistic or not.

Conclusion

The launch of Mechanize has reignited one of the most debated topics in the world of technology today: not just the potential capabilities of AI, but the ethical and practical limits of what it should be allowed to do. With its ambitious aim to automate every conceivable form of labor, the company has raised challenging questions about long-established beliefs surrounding work, value, and economic participation. While some view this as a promising step towards enhanced productivity and operational efficiency, the mixed responses—from enthusiasm to serious concerns—highlight the prevailing uncertainty regarding AI’s evolving role in shaping society.

As Mechanize moves forward, all eyes will be on the company, not only for the technical execution of its plans but also for how it addresses the broader human issues tied to the future of automation. The path it chooses could have far-reaching implications for the way we think about work and human involvement in the overall economy.