Andrej Karpathy released a job market visualizer for the US, and the data it surfaces is more nuanced than the "AI is taking all the jobs" narrative suggests. I spent some time exploring it, and a few patterns jumped out.
First, the tool itself. It aggregates job postings across industries and geographies, visualizes trends over time, and lets you slice the data by role type, location, salary range, and required skills. It's the kind of tool that Karpathy is uniquely positioned to build - he understands both the AI space and the data visualization craft.
Now, the patterns.
AI-related job postings are up, but not in the way you'd expect. The biggest growth isn't in "AI Engineer" or "ML Researcher" roles. It's in traditional roles that now require AI skills. Marketing managers who can use AI tools. Financial analysts who can build AI-powered models. Operations managers who can deploy automation. The AI skill premium is spreading horizontally across every function, not concentrating vertically in AI-specific roles.
This is actually what I've been predicting. The real impact of AI on the job market isn't creating a new category of "AI jobs." It's raising the bar for every existing category. The marketing manager who can't use AI tools is at a disadvantage against one who can. That dynamic is showing up clearly in the data.
Software engineering postings are restructuring, not declining. The doom narrative says AI will replace software engineers. The data shows something more complex. Junior engineering postings are down. Senior and staff-level postings are stable or up. The interpretation that makes sense to me: companies are using AI tools to boost the productivity of experienced engineers rather than hiring more junior engineers. The floor for entry is rising, not the ceiling.
This has equity implications that deserve serious discussion. If the entry-level ramp into software engineering gets steeper because AI handles the tasks that juniors used to do, how do people gain the experience needed for senior roles? This is a pipeline problem that the industry needs to address proactively.
The geographic distribution is shifting. Remote work + AI tools is accelerating the decentralization of tech employment. The visualizer shows growing job postings in mid-tier cities and a slight decrease in traditional tech hubs. If an engineer in Austin can use AI tools to be as productive as a team of three in San Francisco, the economics of hiring in expensive cities change.
"AI-proof" roles are a myth. Some commentators have been categorizing jobs as "AI-proof" or "AI-vulnerable." The data doesn't support clean categories. Almost every role is being reshaped. The degree varies, but the direction is consistent.
What I take away from the visualizer is that the AI job market impact is primarily about augmentation and skill evolution, not replacement. The people who adapt - who learn to use AI tools, who focus on the parts of their job that AI can't do well, who develop judgment and taste that machines lack - are seeing their value increase. The people who don't adapt are seeing their options narrow.
Karpathy's choice to build this as a public tool is characteristically thoughtful. He's giving people data instead of opinions. You can look at the same charts and draw different conclusions. But the data is there, and it's more informative than the hot takes on either side of the "AI and jobs" debate.
I'd recommend spending some time with the visualizer regardless of your industry. Understanding how AI is reshaping the labor market isn't just an academic exercise - it's career-critical information. The patterns are moving fast, and the people who see them early have an advantage.
The future of work isn't AI versus humans. It's humans with AI versus humans without it. The data is clear on that.