We've arrived at the moment every dystopian sci-fi writer has been waiting for: Goldman Sachs just hired its first non-human employee. Not a particularly charismatic one, mind you—just an AI software engineer named Devin that can code, debug, and apparently won't ask for a raise. While Goldman's tech chief Marco Argenti waxes poetic about "hybrid workforces" and humans working "side by side" with AI, the real question isn't whether this technology works. It's whether we're being honest about what comes next.
MIT research shows AI will replace 2 million manufacturing workers by 2025. But finance jobs might disappear even faster because everything is data-based. And Goldman isn't exactly being subtle about the implications. Argenti admits this more powerful AI "has the potential to boost worker productivity by up to three or four times the rate of previous AI tools." When productivity increases by 300-400% per worker, basic math suggests you don't need as many workers. But let's pretend otherwise for a moment.
The euphemisms are flying fast and thick. Bloomberg Intelligence report released Thursday found that global banks are expected to cut as many as 200,000 jobs in the next three to five years as AI takes on more tasks. Yet Argenti frames this as augmentation, not replacement. "We're going to start augmenting our workforce with Devin," he told CNBC, "which is going to be like our new employee." New employee. Not replacement. Got it.
This semantic gymnastics would be almost admirable if it weren't so transparently disingenuous. So far in 2025, there have been 342 layoffs at tech companies with 77,999 people impacted. That's 491 people losing their jobs to AI every single day. Meanwhile, Goldman plans to deploy "hundreds of Devins" initially, potentially scaling to "thousands." Simple question: if one AI can do the work of multiple humans—and do it faster, cheaper, and without sick days—why exactly would you keep all those humans around?
The "hybrid workforce" narrative becomes even more questionable when you examine who's actually driving these decisions. Companies will reskill human managers to oversee a hybrid workforce. The role of human resources will evolve into a department for human and machine resources. Notice how quickly "human resources" becomes "human and machine resources"? That's not evolution—that's a fundamental shift in how organizations view their workforce composition.
But here's where the Goldman experiment gets genuinely interesting, and concerning. 28 percent of executives in software engineering expect generative AI to decrease their organizations' workforces in the next three years, while 32 percent expect the workforce to increase. That's essentially a coin flip on whether AI creates or destroys jobs. And we're basing trillion-dollar economic decisions on coin flips?
The real tell isn't in what Goldman is saying—it's in what they're not saying. When Argenti describes the future where "engineers are going to be expected to have the ability to really describe problems in a coherent way and turn it into prompts," he's describing a world where engineering becomes prompt engineering. That's not the same job. That's a fundamentally different skill set, likely requiring fewer people.
70% of employees believe generative AI (GenAI) would change 30% or more of their work. Note: change, not eliminate. But when 30% of your work disappears, what happens to the other 70%? Does it magically become more valuable, or does it get redistributed among fewer people?
The most honest assessment comes from Goldman's own research: artificial intelligence has the potential to bring about sweeping changes to the global economy... they could drive a 7% (or almost $7 trillion) increase in global GDP and lift productivity growth by 1.5 percentage points over a 10-year period. Massive productivity gains typically don't come from having the same number of people doing slightly better work. They come from having fewer people doing dramatically more work.
We're not against AI. We're against the mythology being built around it. The "hybrid workforce" isn't some harmonious collaboration between humans and machines—it's a transitional phase toward automation. Goldman's Devin pilot isn't the beginning of human-AI collaboration; it's the beta test for human-AI replacement.
The uncomfortable truth is that most white-collar work is more routine than we'd like to admit. These repetitive task jobs are AI's easiest targets. And once AI can handle the routine stuff, it won't stop there. It'll move upstream to more complex tasks, then to strategic thinking, then to decision-making. The question isn't whether AI will replace human workers—it's how quickly and how we'll adapt.
Maybe we need fewer euphemisms and more honesty. The future of work isn't about humans and AI working "side by side." It's about figuring out what humans can do that AI can't, and being brutally realistic about how small that list is becoming.
Ready to navigate the AI transformation without the corporate spin? Contact Winsome Marketing's growth experts to develop strategies that position your business for the AI future—whatever that actually looks like.