Opinions on artificial intelligence remain sharply divided, with the technology variously characterized as an essential collaborative partner, a glorified search engine, or a vastly overrated product. This fundamental disagreement persists as tech industry leaders continue to promote the narrative that AI will trigger a new industrial revolution by fundamentally altering the nature of employment. Tech executives championing AI have long spun the narrative that the tech will revolutionize jobs and bring about a new industrial revolution.
The debate intensified significantly this past week following a viral essay from investor and former AI startup CEO Matt Shumer. He claimed that AI is poised to disrupt any role involving computer-based tasks, citing his own experience using OpenAI’s GPT-5.3 Codex tool to develop a complex app. Shumer stated that the AI managed user flow, design, and tens of thousands of lines of code, effectively making decisions regarding judgment and taste. However, AI researchers previously criticized Shumer in 2024 for exaggerating the capabilities of his models. In response to that controversy, Shumer acknowledged the incident as the greatest mistake of his professional life. After a viral essay from an AI CEO and investor claimed the tech is coming for any job that involves sitting in front of a compu, the discrepancy in how people view AI has perhaps never been so apparent as this past week. He went on to claim that the AI was able to test the app and make decisions concerning taste and judgment. According to he apologized at the time and, It was the “biggest mistake” of his “professional life” and that he learned through the process). According to shumer, He primarily uses OpenAI’s GPT-5.3 Codex tool and that he was working on a “medium to high complexity app” for testing purposes. “I’ll tell the AI: ‘I want to build this app. It could begin improving upon itself, as well, and he surmised that if AI could write code that well. Correction: A previous version of this story misstated Shumer’s title.
Some experts argue that the confusion surrounding AI’s actual utility stems from a lack of standard terminology and varying levels of exposure. Matt Murphy, a partner at Menlo Ventures who has invested in companies like Anthropic, noted that there is a wide spectrum of experience with the technology, which continues to change rapidly. This lack of uniformity makes it difficult to assess whether current AI systems are truly capable of replacing professional workflows or if they are merely being marketed as such. Plus fewer limits on usage, but those who pay get access to another feature: Agents that can handle some work for you rather than just chatbots that craft responses. And no one seems to agree on which is right. Yet it’s all being referred to in the same way, but there may be a simpler explanation as to why people have taken such divergent stands: People use different types of AI in different ways. Here’s roughly what it should look like, here’s what it should do. Especially since he was vague about which model he used and what type of app the AI built for him, some experts are skeptical the use cases Shumer outlined are possible even with paid plans. Some studies have poured cold water over how capable AI truly is and how quickly it’s being adopted.
A critical factor in this perception gap is the difference between free and paid AI services. While free tiers can provide sophisticated dialogue and helpful summaries, experts like Carnegie Mellon University professor Emily DeJeu warn that it is misguided to judge the technology’s ultimate potential based solely on free tools. Oren Etzioni, professor emeritus at the University of Washington, compared the difference between these tiers to an eager, inexperienced intern versus a seasoned professional. He noted that while free AI is useful for basic tasks, it should not be relied upon for high-stakes roles like legal assistance. But users will typically have to pay to conduct deep research or draft sophisticated documents using AI, free AI tiers are good at writing summaries and generating content. AI is a helpful tool that programmers use to speed up development, but it’s still prone to mistakes, and AI models aren’t writing themselves.
Despite this, companies are beginning to move advanced features into free offerings to remain competitive. For instance, Anthropic recently launched its Sonnet 4.6 model, which the company claims performs closer to its elite, paid-only Opus models. James Landay, cofounder of the Stanford Institute for Human Centered AI, suggested that this trend is narrowing the performance gap between free and premium services. For example, is only available in the $20-per-month Pro plan and higher, anthropic’s Claude Cowork agent.
Current skepticism is also fueled by data suggesting that AI does not always increase productivity. A study by the organization Model Evaluation and Threat Research found in July that software developers actually took 19% longer to complete coding tasks when using AI tools, though this analysis was based on software from early 2025. Additionally, researchers from the Center for AI Safety and Scale AI reported that leading models frequently produced flawed results when assigned complex tasks like data visualization. While some researchers and executives are sounding the alarm about safety concerns on their way out the door, skeptics think it’s all marketing hype.
Landay argues that the narrative surrounding AI’s impact on jobs is often overstated, particularly regarding software engineering. While he acknowledges that coding is a logical structure well-suited for machine testing, he notes that many other professional roles do not share that same rigid framework. As AI companies like Anthropic continue to release industry-specific tools, the industry remains at a crossroads between genuine innovation and high-stakes marketing. That launch, followed by Shumer’s essay, stoked concerns that AI will eventually broadly automate knowledge work the way it’s starting to streamline software engineering jobs.
