Tech Entrepreneurs Doubt AI Agents Will Replace Human Workers Due to High Costs

The rising expenses associated with using AI agents in business may hinder their ability to replace human workers who can perform equivalent tasks at a lower price, according to insights from notable tech investors.

In a recent episode of the All-In podcast, Jason Calacanis disclosed that he allocates approximately $300 daily on an AI agent from Anthropic, despite it functioning at merely 10% to 20% of its potential. “When do tokens outpace the salary of the employee?” he pondered, referencing the tokens required to access most AI systems.

Chamath Palihapitiya, CEO of Social Capital, echoed this sentiment, asserting that AI models must demonstrate at least double the productivity of their human counterparts to justify their costs, suggesting a possible budget for AI usage.

Paraphrased Quotes

Jason Calacanis: “We incur $300 per day for an agent utilizing the Claude API, which achieves around 10%-20% productivity, totaling about $100,000 annually per agent.”
“¿Cuándo superan los tokens el salario del empleado?”
(When do tokens outpace the salary of the employee?)

On another note, Mark Cuban remarked that the financial burden of AI integration emphasized by Calacanis and Palihapitiya constitutes a compelling argument against AI substitution for jobs. He further noted that sustaining eight Claude AI agents could require double the cost of maintaining a single employee’s daily output, leading him to question if these bots offered the necessary productivity advantage.

The apprehension regarding AI potentially displacing a significant portion of the labor force has stirred uncertainty in recent times, with various companies attributing layoffs to the introduction of AI in their operations, thereby rendering certain positions redundant.

A July paper from Microsoft highlighted that roles in knowledge work, customer service, and sales remained particularly vulnerable to AI displacement.

David Sacks, a senior figure in AI and crypto at the White House, is among those who assert that these fears might be exaggerated, claiming AI still depends on human guidance for “driving business value.” In contrast, firms like McKinsey & Co suggest that the intention behind AI agents is to fully automate tasks, requiring minimal human intervention.

AI Agents Likely to Use Stablecoins as Their Currency

The advent of AI agents has seen a surge in adoption among cryptocurrency enthusiasts. Jeremy Allaire, CEO of Circle, forecasted that in the coming five years, billions of these AI agents will rely on stablecoins for various transactions.

Changpeng Zhao, co-founder of Binance, stated that cryptocurrency will likely become the primary currency for these AI agents, further asserting that blockchain represents the optimal technological interface.

AI agents are currently active on various blockchain networks, including Ethereum Layer 2 Base, where the Virtuals Protocol leverages micropayments and automates trading for users, while ASI Alliance on Fetch.ai manages assets and executes various economic responsibilities.

Recently, OpenAI announced a new benchmark aimed at evaluating the efficacy of diverse AI models in identifying and remedying security flaws in smart contracts. OpenAI recognized that their findings were crucial for assessing AI performance in financially significant contexts, noting that smart contracts safeguard substantial assets, and AI agents are poised to significantly impact both attackers and defenders alike.