People are not taking AI agents seriously enough.
With the recent release of OpenAI’s consumer-targeted AI agent — called Operator — it’s become clear that the intent of agentic models is to wholesale replace human work done on computers.
First, they’re coming for online work.
Operator functions best when navigating an internet browser. OpenAI’s marketing shows casual use — having the agent order groceries or book a reservation for you.
But agents can go far beyond simple food orders. These models have been trained to replicate virtually anything a human can currently do in an internet browser. Communicate with your coworkers on Slack, develop and deploy a prototype, monitor and manage your stocks, the list goes on… and on… and on.
Goodbye, humans
Over 92% of jobs require digital skills. Whether they are in an office or sitting at home, a significant percentage of humans make a living exclusively from computer work. Developers, day traders, product managers, you name it — industries of all types are saturated with humans clicking around various websites and performing tasks.
AI agents were made to click around on various websites and perform tasks. And right now, today, they’re pretty damn good at it. Take a look (thanks Yana Welinder):
In this example, Operator does the following within 5 minutes:
Sets up an integration between Slack and a product management app (Kraftful)
Prioritizes feedback and added to roadmap
Writes a PRD based on user feedback, starts making user stories
Syncs a user story (with additional info) to an issue tracking platform (Linear)
Assigns the Linear ticket to an engineer
This is a product manager’s full flow for ticket creation, automated entirely. The possibilities for automation work with agents is near endless. It can:
Write detailed notes from a meeting transcript, email out personalized action items to each attendee
Create new recipes, write human-sounding recipe blogposts, and continually publish them to a food blog
Analyze a competitor’s website, develop a version of the site, and deploy the site to Cloudflare or other hosting platform
Monitor incoming messages on a customer service platform and reach out to users as needed
Research a competitor’s marketing material and plan a week’s worth of marketing content
Search through thousands of documents and databases on stock market data, identify the optimal moves, and make transactions every minute
Initially, the models powering agentic tools will be imperfect. Operator is slow and reaches out to its human handler often for assistance. But this will change rapidly as both the frontend agent product and the model powering it inevitably get more and more powerful.
OpenAI’s Operator agent currently runs on a model based on their GPT-4o model, a model that will be two generations behind with the release of the o3 model in a month or two. The leap in logic from GPT-4o to o3 is significant, as you can see in the ARC-AGI benchmarks:
This means that by the time Operator is released to the public (it’s in a research preview for ChatGPT Pro users at the moment), the model that is powering it could be one 16x more powerful than the one powering that demo shown earlier.
Within the year, AI agents will be performing online tasks for humans on a scale that’s hard to comprehend. And then what?
Goodbye, internet
The internet as we know it (a human-centric ecosystem of ideas, commerce, and culture) is going to die, and sooner than you might expect. Today’s agents like Operator are poised to inhabit platforms built for human quirks and reshape them for machine efficiency.
First, the working internet collapses. Tools like Slack and GitHub, designed for human collaboration, will strip away UX niceties (loading animations, “read time” estimates) in favor of API-first, agent-optimized interfaces. Productivity platforms will resemble Bloomberg Terminals: stripped-down dashboards where AI agents trade tasks at inhuman speeds.
The economic foundations of the human web crumble next. Advertising, which fuels $740B in annual global spending, implodes when bots dominate traffic. Why pay for Google Ads targeting agents that don’t consume products?
The human internet shrinks smaller and smaller, with people primarily logging onto to consume content (YouTube, Netflix, blogs) or communicate with friends and family (while hopefully avoiding AI impersonation scams). The Dead Internet Theory, a popular conspiracy that claims the internet is already dominated by robots, comes reality.
Beyond the browser
Ultimately the impending societal crisis isn’t job loss — it’s value collapse. Within the next few years, millions of AI agents will automate workflows faster than humans invent new tasks.
Consider: one human directing 10,000 agent “product managers” could replicate entire corporate divisions overnight. Situations like these, duplicated across industries, will result in a digital labor glut where even AI-powered services commoditize into near-free utilities.
Human workplace survivors will cluster into roles like:
Agent wranglers. Prompt engineers who troubleshoot workflows (e.g., “Why did our agent CFO invest $2M in pickleball NFTs?”).
Automation ethicists. Lawyers and philosophers auditing agent decisions, akin to today’s AI bias researchers.
Physical-world hybrids. Roles merging agent oversight with hands-on labor, like nurses managing surgical AI while handling patient care.
But the real disruption begins when agents escape the browser. Projects like OpenAI’s partnership with Figure Robotics hint at this future: the same logic automating Slack workflows today could pilot humanoid robots tomorrow.
Climate science. Ruggedized agents in the Arctic analyzing ice cores via offline, onboard vision models, debating findings with scientists (or other bots) via satellite link.
Manufacturing. AI agents trained on CAD software commandeer 3D printers, adjusting designs in real-time to bypass supply chain gaps
Infrastructure. AI agents autonomously managing urban systems; optimizing traffic flows through real-time sensor networks, directing repair crews to fix potholes or power grids, and simulating disaster responses to reroute emergency services. These systems would self-adjust to environmental stressors (e.g., flooding, heatwaves) while human engineers audit safety protocols and ethical implications.
The internet is the first domino. As agents master digital work, their move into the physical world (via robots, drones, and IoT networks) will blur the line between “automation” and “agency.”
Reclaiming human agency
The Dead Internet Theory serves as both warning and blueprint. To navigate this transition, we must:
Invest in detection. Develop tools to distinguish human/AI content, as urged in recent publications like Building Human Systems of Trust.
Redefine work. Embrace roles like “automation ethicists” to govern AI agents’ societal impact.
Preserve human spaces. Legislate digital reserves for organic interaction, free from agent invasion.
With these efforts in mind, its time to think seriously about building a society where humans are no longer required to work. What does this look like and who is responsible for making the changes necessary to make it a reality?
When an AI can rebuild a website and a bridge, human economics as we know it becomes obsolete.
To an everyday consumer this is scary stuff. I can’t begin to understand the impact. I hope it’s monitored and reined in. It appears this can be a good thing if we keep it in check. This is just one woman’s opinion.
the wild wild west, again.