The future of work feels increasingly uncertain as artificial intelligence rapidly advances. One voice cutting through the noise is Matt Shumer, a notable figure in the AI space and co-founder of OthersideAI. His viral essay, “Something Big Is Coming,” captivated millions. Shumer’s stark warning compares today’s AI landscape to the unsettling calm before the COVID-19 pandemic. He urges everyone to recognize the profound, imminent disruption that could reshape society on an unprecedented scale.
But is this a legitimate, urgent call to action, or a case of “weaponized hype” from an industry insider with commercial interests? This article dives deep into Matt Shumer’s compelling predictions, examines the critical counter-arguments, and offers balanced, actionable advice for navigating an AI-driven future.
Matt Shumer’s Vision: An Unstoppable AI Tsunami
Matt Shumer, also a General Partner at Shumer Capital, believes we are on the cusp of a technological revolution. He stresses that AI’s impact could dwarf events like the COVID-19 pandemic. His nearly 5,000-word essay, which garnered over 60 million views on X, aimed to alert a broad audience beyond the tech bubble. He wants ordinary people, like his nearing-retirement lawyer father, to grasp the impending changes.
Shumer’s personal “holy shit” moment came after interacting with OpenAI’s GPT-5.3-Codex. This model was famously described in its release notes as “instrumental in creating itself.” This experience solidified his conviction: AI is now capable of performing his own technical work. He describes this realization as “confusing and terrifying.” As a 26-year-old, he saw his career path potentially vanishing. He candidly admits using advanced AI, specifically Claude, to refine his viral essay. He argues this simply reinforced his core message about AI’s capabilities.
The Evidence Behind the Alarm Bells
Shumer highlights concrete examples to support his urgent message. He claims that since 2025, new AI techniques have dramatically automated his technical job functions. He now describes complex app requirements in plain English. The AI then autonomously writes tens of thousands of lines of code. It self-tests, iterates design, and delivers a finished, often superior, product in hours. Crucially, this often requires no human correction. He describes this as a “different thing entirely,” where AI “does my job better than I do.”
He forecasts this scenario, where AI surpasses human capability in specific tasks, for virtually all white-collar sectors. This includes law, finance, medicine, accounting, consulting, writing, and design. He predicts this transformation within one to five years, possibly sooner. Shumer points to a $1 trillion market value wipeout in software as a sign of the market’s dawning recognition.
His concerns are not isolated. Anthropic CEO Dario Amodei conservatively predicts AI could eliminate 50% of entry-level white-collar jobs within the next one to five years. Similarly, xAI CEO Elon Musk has characterized AI as a “supersonic tsunami.” Musk believes it will swiftly displace jobs not involving physical labor.
Why Current AI is Different
Shumer dismisses skepticism rooted in early AI versions. He argues that ChatGPT from 2023 or early 2024, prone to “hallucinations” and limitations, is irrelevant. Current models are “unrecognizable” from those of even six months ago. He notes that free AI tools typically lag over a year behind the paid, advanced models. He shares an anecdote of a managing partner at a major law firm. This partner uses advanced AI daily like an “instant team of associates.” The partner observes significant capability improvements every couple of months.
A Counter-Narrative: Is It Hype or Reality?
While Shumer’s warnings resonate with many, others express significant skepticism. Critics argue that the AI industry often suffers from a “Chicken Little problem.” Alarmist predictions, they say, frequently accompany commercial interests. This makes them difficult to take seriously.
Gary Marcus, a prominent AI researcher, critically dissects Shumer’s viral post. Marcus labels it “weaponized hype,” presenting an unrealistic and one-sided view of AI capabilities. He notes Shumer’s claims about complex applications generated without errors lack concrete data. Marcus highlights that AI benchmarks, like METR’s task-time, measure only 50% correctness for specific coding tasks. This is far from reliable, error-free performance on general tasks.
The Problem with Unchecked Optimism
Marcus also points out Shumer’s neglect of common AI issues. These include “hallucination and boneheaded errors” often found in real-world applications. He cites research from Caltech/Stanford and Apple that reviews AI reasoning errors, contradicting Shumer’s overly optimistic picture. Marcus even questions Shumer’s past, recalling “exaggerated claims” about an open-source AI model that underperformed.
Real-world evidence frequently challenges AI’s touted legal prowess. Lawyers across the country face censure for using AI, with one lawyer documenting 912 cases of AI hallucinations in the legal profession. Even OpenAI’s GPT-5.2 has a reported hallucination rate of 10.9%. This rate only drops to 5.8% with internet access. Such figures cast doubt on AI’s trustworthiness for critical tasks.
The Value of Human Input and Nuance
The Spyglass article, in particular, offers a nuanced critique. It agrees with Shumer that coding is ripe for immediate transformation. However, it argues that AI’s suitability for coding does not directly translate to seamless disruption across all white-collar jobs. The author emphasizes the intrinsic value of the process of creation, like writing. It’s the act of thinking and forming genuine opinions that matters, not just the output. This suggests humans will continue tasks for inherent value, even if AI can perform them.
An AI-generated counter-argument, crafted by Claude itself for Spyglass, further challenges Shumer’s stance. Claude firmly rejects the COVID-19 pandemic comparison. It notes that COVID’s impact was immediate and obvious, requiring no evangelists. AI, conversely, has improved for over a decade, with many “everything just changed” moments that proved to be mere “vibes.” Claude asserts a “vast chasm” between “tools are getting better” and predictions of 50% white-collar job elimination in 1-5 years. Those building technology, it suggests, often overestimate the speed of societal adoption.
Navigating the Nuance: What’s Truly Happening?
It’s clear that something significant is indeed changing. Newer AI systems do enable faster development, particularly in coding. Automating tedious tasks like email or legal document review is welcome and becoming increasingly feasible. However, the path to fully autonomous, error-free complex operations across all white-collar fields remains contentious. The trustworthiness and security of autogenerated code, for example, are major concerns. Marcus suggests that AI systems that are “increasingly at ease just trusting them” even when they shouldn’t, actually become “more dangerous.”
The debate isn’t whether AI will have an impact, but how fast and how universally this will unfold. Will it be a rapid, pandemic-like shock, or a more gradual, internet-like transformation over decades? The consensus among critics leans towards the latter. They warn against a “discount rate” for predictions from AI CEOs. Their company valuations are often tied to beliefs in rapid, powerful AI advancement.
Preparing for the AI-Driven Future: Actionable Steps
Despite the debate over timelines and specifics, a common thread emerges: individuals must engage with AI now. Preparation and adaptability are key. Here are comprehensive steps synthesized from expert advice:
Prioritize Early Engagement: The greatest advantage comes from understanding, using, and adapting to AI early.
Invest in Advanced AI Tools: Subscribe to the most capable models available (e.g., GPT-5.2 on ChatGPT or Claude Opus 4.6). Always ensure you select the highest-performing version.
Integrate AI into Core Workflows: Move beyond simple questions. Assign complex, real-world professional tasks. Lawyers might use AI for contract analysis and counter-proposal drafting. Financial professionals can build models from messy data. Managers can analyze quarterly team data. Experiment and iterate, as even partial successes today will likely be near-perfect soon.
Abandon Ego: Resist dismissing AI as a fad or believing your field is immune. Recognize that using AI doesn’t diminish expertise, it augments it.
Strengthen Financial Resilience: Build savings and exercise caution with new debt. Financial flexibility is crucial for potential industry disruption.
Leverage Temporary Defenses: Roles requiring physical presence, deep human relationships, licensed accountability (where a human must sign off), or operating in heavily regulated industries might offer more time. However, view these as temporary shields, not permanent solutions.
Rethink Education: For the next generation, emphasize learning to collaborate with AI. Cultivate curiosity, adaptability, and genuine passions. Don’t optimize for traditional career paths that may become obsolete.
Seize New Opportunities: AI significantly lowers barriers to creation (e.g., apps, books) and skill acquisition (providing a 24/7, infinitely patient tutor). Embrace these accessible avenues.
Cultivate Adaptability: The most crucial skill is not mastering one tool, but the habit of learning new ones quickly. Experiment constantly and become comfortable with continuous, rapid change.
Commit to Daily AI Experimentation: Dedicate an hour daily to actively use AI for novel or challenging tasks. This consistent engagement provides a critical understanding of its capabilities, offering a significant advantage.
- Maintain Critical Thinking: While experimenting, remain unpanicked. Discount extreme predictions, especially from those with vested interests. Focus on understanding real-world capabilities.
- www.businessinsider.com
- mashable.com
- garymarcus.substack.com
- fortune.com
- spyglass.org
Frequently Asked Questions
What are the main concerns Matt Shumer has about AI’s impact on the future of work?
Matt Shumer expresses deep concern that AI advancements will lead to widespread job displacement across white-collar sectors. He believes AI can already perform much of his own technical work, building complex applications faster and better than humans. He warns that AI is a “general substitute for cognitive work,” implying few jobs requiring reading, writing, analyzing, or communicating via computer are safe in the medium term. Shumer compares this disruption to the COVID-19 pandemic, emphasizing its speed and magnitude.
What are practical steps individuals can take to prepare for AI disruption?
Individuals can prepare by actively engaging with AI tools. Shumer advises investing in paid, advanced AI models and integrating them into core professional tasks. He recommends dedicating an hour daily to experimentation, abandoning ego, and strengthening financial resilience. Other advice includes cultivating adaptability, rethinking education for future generations, and leveraging temporary defenses in specific job roles. The goal is to understand AI’s capabilities and learn to collaborate with it effectively.
Should people trust all AI predictions from industry insiders?
No, it’s advisable to view AI predictions from industry insiders with a critical eye. While experts like Matt Shumer offer valuable insights from their firsthand experience, some critics argue that alarmist warnings can sometimes serve commercial interests or reflect an overestimation of societal adoption speed. Experts like Gary Marcus and even an AI-generated counterpoint from Claude suggest applying a “discount rate” to such predictions, especially those tied to company valuations, and to distinguish between “tools getting better” and widespread, immediate societal upheaval.
Conclusion: Adapt or Be Left Behind
The debate surrounding Matt Shumer’s “Something Big Is Happening” highlights a critical juncture for humanity. While there are valid points of skepticism regarding the pace and universality of AI’s disruptive power, there’s no denying that artificial intelligence is rapidly evolving. It’s profoundly impacting various sectors, especially coding and tasks requiring cognitive automation.
The most potent takeaway from this discussion is the urgent need for personal engagement and adaptability. Whether AI ushers in an immediate tsunami or a gradual, internet-like transformation, those who actively seek to understand, integrate, and collaborate with these powerful tools will be best positioned. The future of work won’t be about being replaced by AI, but about being replaced by people who use AI. Embrace curiosity, experiment daily, and prepare for a future where continuous learning and critical thinking are paramount.
