The people building AI are now warning that AI is making them worse at thinking
Addy Osmani, an engineer at Anthropic, the company that makes Claude, published a warning about what he calls 'cognitive surrender': the condition of having AI handle enough of your work that critical thinking stops being exercised. That warning comes from inside the industry most invested in accelerating AI adoption. Two center-left outlets picked it up. The rest of the spectrum hasn't touched it.
AI agents raising concerns about 'cognitive surrender' among tech workers
2 sources · hover a dot to see coverage
What happened
Anthropic engineer Addy Osmani published a warning about 'cognitive surrender,' his term for the erosion of critical thinking that occurs when AI agents take over enough cognitive work that the human stops practicing it. His concern: nothing in his day-to-day forces him to think hard anymore. Separately, Business Insider reported that managers are now pushing back against AI-generated work product, with one CEO, Michelle Volberg of a tech recruiting firm, sending a staff-wide email asking employees to 'please stop talking to me like a robot.' Some companies have moved to explicit policies against AI slop; others have reached the point of terminations. The Verge, meanwhile, reported that Big Tech's AI infrastructure spending shows no signs of stopping after Q3 earnings calls from Amazon, Microsoft, Alphabet, and Meta. Two outlets covered this. The gap in the rest of the spectrum is the story.
Business Insider ran both the warning and the workplace backlash as separate pieces
Business Insider published two distinct stories: one centered on Osmani's 'cognitive surrender' framing, and one on the managerial backlash to AI-generated work product. The pairing is editorially significant because it captures both ends of the same problem. Workers are outsourcing cognition; bosses are receiving the output and finding it hollow. Volberg's email to staff is a specific, dateable event — roughly six months ago — that grounds the broader pattern in something checkable. The two pieces together make an argument that neither makes alone: the AI productivity promise is generating its own resistance, from the top down and the inside out.
What one side told you that the other didn't
The person sounding the alarm works at the company selling the tool.
Osmani isn't a critic, a regulator, or an academic. He's an engineer at Anthropic, one of the leading AI labs, warning that the agents his employer builds are degrading his own capacity for independent thought. That's not a fringe concern from outside the industry. It's an insider's account of a structural problem with the product. Only one outlet named him specifically and quoted his reasoning. The rest of the coverage landscape, including every right-leaning outlet, has not picked this up at all.
Right-leaning outlets are entirely absent from this story.
Outlets that have covered AI productivity gains, AI job displacement fears, and tech sector layoffs extensively have published nothing on 'cognitive surrender' or the managerial backlash to AI-generated work. That's a gap worth noting because the story has obvious resonance across the political spectrum: it touches on worker deskilling, corporate overreach, and the gap between what technology promises and what it delivers. The absence isn't ideological in any obvious way. It's just a hole.
Spending is accelerating while the people doing the work are raising flags.
The Verge's earnings-call piece and Business Insider's Osmani piece sit in direct tension, and no single outlet holds both facts at once. Big Tech committed to record AI infrastructure spending in Q3, with no end signaled. At the same time, one of the engineers inside that ecosystem is publicly describing a loss of cognitive agency. The investment thesis and the human cost are being covered in separate rooms by separate reporters, and the reader has to assemble the connection themselves.
What to watch
If Osmani's 'cognitive surrender' post circulates further or draws a response from Anthropic leadership, right-leaning outlets will face a choice: cover it as a tech-industry self-criticism story or ignore it. Watch for coverage in the next two weeks. If it stays confined to center-left tech media after going wider on social, that asymmetry becomes the more durable finding.
This kind of gap — where one side reports a fact and the other doesn’t mention it — shows up in about 40% of major stories.