Superhuman's Rows Buy: The 4-Hour AI 'Tax' Nobody Is Talking About

AI-driven productivity gains often mask a hidden “tax” of rework and organizational redesign.

• Superhuman’s acquisition of Rows is intended to enhance its AI productivity platform, but a “4-hour tax” of rework for every 10 hours of AI-driven productivity gain may negate many of the promised benefits.
• A National Bureau of Economic Research (NBER) survey reveals that 80% of companies using AI report no measurable impact on productivity or employment.
• Businesses must prioritize organizational redesign and skills training to fully realize the productivity gains from AI, or face increasing technical debt and wasted investment.

Superhuman’s $825 Million Gamble: Will Rows Solve the AI Productivity Paradox?

Superhuman, with its last valuation at $825 million in 2021, is betting big on AI-powered productivity. The acquisition of Rows, a company focused on simplifying data work through integrations and no-code automation, signals a push towards an all-in-one AI workspace. But will this acquisition truly enhance productivity, or is it just another attempt to solve the elusive “AI productivity paradox,” where promised gains often fail to materialize.

The core issue is that while AI can offer task-level efficiency improvements, these often don’t translate to measurable gains at the enterprise level. A National Bureau of Economic Research (NBER) survey revealed that a staggering 80% of companies using AI report no measurable impact on productivity or employment over the past three years. This suggests that simply внедрять AI tools isn’t enough; organizations need to fundamentally rethink their workflows and processes.

The Dark Side of AI-Powered Emails: Why Lina M. Khan’s FTC is Cracking Down, according to TechCrunch

AI is being touted as a way to streamline email workflows and boost productivity. However, Lina M. Khan’s Federal Trade Commission (FTC) is cracking down on “AI-washing,” where companies exaggerate the capabilities of their AI offerings. The FTC’s “Operation AI Comply” is a clear message: there’s no “AI exemption” from existing laws against deception and fraud.

The potential for misuse is significant. AI-generated phishing emails, for example, are achieving dramatically higher click-through rates compared to traditional phishing attempts. This highlights the darker side of AI, where it can be used to exploit vulnerabilities and deceive users. The FTC is therefore right to be vigilant in ensuring that AI is not used to mislead consumers or engage in unfair business practices.

The “Commit Count” Trap: Why Bill Harding Fears Runaway AI Technical Debt

Many companies still measure developer productivity by metrics like lines of code written or the number of commits made. Bill Harding, CEO of Amplenote and GitClear, warns that this can lead to a dangerous situation where AI tools accelerate the accumulation of “technical debt.” Focusing solely on output can mask underlying problems with code quality, maintainability, and security.

METR (Model Evaluation & Threat Research) found that developers using AI tools actually took 19% longer to complete tasks, despite believing that AI had sped them up by 20%. This suggests that AI can create the illusion of productivity while simultaneously introducing hidden inefficiencies and complexities. Unless companies adopt more sophisticated ways of measuring and managing technical debt, they risk creating systems that are difficult to maintain, debug, and scale.

The BCG Paradox: When Consultants Benefit While Open-Source Developers Suffer

While some studies, such as one conducted at BCG, show significant productivity boosts from AI, the benefits aren’t universal. BCG consultants, for example, reportedly finish work 25% quicker with 40% higher quality when using AI. But a contrasting study revealed that experienced open-source developers took 19% longer to complete tasks when using AI tools.

This highlights the importance of context and skill level. AI may be more effective for certain types of tasks or for individuals with specific skill sets. The fact that open-source developers, who are often highly skilled and experienced, struggled with AI tools suggests that these tools may not always be a good fit for complex or nuanced tasks. It also underscores the need for careful training and adaptation to ensure that AI tools are used effectively.

Beyond the Hype: The Modest AI Productivity Forecast From Wall Street

Despite the widespread excitement around AI, Wall Street’s forecasts for its impact on productivity are surprisingly modest. AI is projected to increase productivity and GDP by just 1.5% by 2035. This suggests that the real benefits of AI will only accrue over the long term and that the immediate impact may be less dramatic than many expect.

The key takeaway is that AI is not a magic bullet. It’s a tool that, when used strategically and effectively, can improve productivity and drive economic growth. However, it’s important to have realistic expectations and to recognize that the full potential of AI will only be realized over time. We must also consider the broader societal and economic implications, including the potential for job displacement and the need for workforce retraining.

Alternative AI Workflows Are Emerging

Andreas Hassellöf, CEO of Ombori, believes that the AI productivity paradox is not a technology problem, but an organizational adoption issue. Investment in workforce training amplifies AI productivity benefits and skills gaps remain the largest barrier to effective AI integration. Rather than full automation, the focus should be on human-AI collaboration, treating AI as a teammate, not just a tool. Organizations can use workplace analytics platforms to track AI tool usage, employee engagement metrics, and productivity indicators.

Workday’s research indicates that for every 10 hours of productivity gained through AI adoption, organizations “pay back” four hours correcting and rewriting low-quality output. Nearly 40% of AI time savings are lost to rework. Measuring adoption requires frequent review and verification.

LinkedIn is facing a class-action lawsuit for allegedly harvesting private messages to train AI models without user consent. This highlights the critical area of privacy concerns in AI development. A class action lawsuit against Eightfold AI alleges violations of the Fair Credit Reporting Act (FCRA).

The Bottom Line

Superhuman’s acquisition of Rows is a risky bet on solving the AI productivity puzzle. Without addressing fundamental issues of organizational design, training, and workflow optimization, it may just create more efficient ways to waste time. It’s critical to invest in workforce training to amplify the productivity benefits of AI. Don’t automate broken processes; fix them first. This is how companies will harness the true potential of AI and avoid the “4-hour tax” that can negate its promised gains. Businesses must prioritize organizational redesign and skills training to fully realize the productivity gains from AI, or face increasing technical debt and wasted investment.

The productivity paradox is a training and organizational design failure, not a technology failure, as Andreas Hassellöf of Ombori aptly puts it. The core problem is not the AI tools themselves, but how they are integrated into existing workflows and how employees are trained to use them effectively.

A study found Copilot users averaged just 1.14 Copilot actions per day. This suggests that many users are not fully leveraging the capabilities of the tool. The focus must shift to training and support to ensure that employees can use AI tools effectively and avoid the pitfalls of the AI productivity paradox. This approach can help companies achieve real productivity gains and avoid the “AI-washing” trap that the FTC is cracking down on.

AI’s real value hinges on human ingenuity.

• Novum Tools