AI is being implementd offshore by teams that dont have the necessary AI expertise and struggle with English communication and comprehension. What could possibly go wrong? It would be funny if it weren’t so sad. This company is in serious trouble. They keep laying off experienced, highly skilled employees in the US, and at some point you have to ask: who is left with the knowledge and experience to actually take care of the customer?
10 replies (most recent on top)
Get ready for everyone's data to be sold to scammers.
Are there any Indian AI models? Surely they have the best models with all their engineering knowledge.
I swear to god if another VP buys in to another AI grifter vendor because they’re selling them the moon and stars (it’s not live yet, but it’s on the roadmap!) I’m going to scream so hard it will shatter the void.
A few years before I parted ways with Optum, my team lead arbitrarily chose to eliminate the Systems Analyst (SA) position, arguing that our developers and QA people could handle the analysis, documentation and communication. Unfortunately, the developers and QA folks didn't know our Business clients as well, hated performing documentation and didn't have nearly the same proficiency with English & writing.
The documentation artifacts became less accurate, the IT deliverables had more defects and the Business clients were less happy. I know for a fact that some mistakes were visible on UnitedHealth member ID cards and Explanation of Benefits (EOB) statements.
The recent push to implement AI reminds me a lot of that earlier effort to reduce costs and streamline processes by eliminating what had been an important role.
AI has caused our team to have to rework thousands of issues off shore did. No repercussions because they're cheap labor but we better hit those metrics otherwise we're in trouble.
Most of the emails I get from sales or LOB leaders proposing new concepts are full on GPT. Including the emoji X and checkmark. But you know when the only thing you have to look forward to is a healthy monthly layoff and a 1% raise with 5% bonus, I can see why people just phone it in.
I agree that offshoring is a major problem. But AI is a problem even for onshore developers: I recently watched an email back-and-forth between senior devs who were literally parroting Copilot. Like one sends over Copilot output, the other feeds it into Copilot and responds with the output, and so on. Neither of them actually understood what it was saying, maybe because it vomits out a small novel. Like they presented on it and had to embarrassingly pause partway through to admit they didn't understand it. What a bleak future for devs who actually care and then have to try to pick of the pieces when stuff breaks or gets unmaintainable. And that's without considering all the worldly reasons to hate hyperscale data centers.
Crazy that we're all of a sudden fine with the idea of communicating entirely through an AI middleman instead of employing the bare minimum of cognitive function. The generated response of "hey ChatGPT, write a response to this" are not your thoughts.
Do you guys actually read these AI generated slop posts that say so much to say so little and think "wow, so insightful!"? Are you the same people spamming my corporate inbox with podcasts of two AIs talking to each other with mindless bullsh-t, or my LinkedIn feed with AI generated grindset posts?
@ay Shut the fu-k up ChatGPT.
The biggest problem isn't that AI is being implemented offshore. It's that AI is being implemented without the right combination of technical expertise, institutional knowledge, business context, and communication skills.
AI doesn't eliminate the need for experienced engineers—it arguably makes that experience more important. Someone still has to understand the architecture, validate what the AI produces, recognize when it's wrong, make the right tradeoffs, and understand what the customer actually needs.
If a company simultaneously lays off the people who have spent years building that knowledge and then expects less-experienced teams to replace them with AI, the risk isn't simply lower productivity. It's loss of institutional knowledge, weaker technical oversight, poor requirements comprehension, security and quality problems, and ultimately customer dissatisfaction.
The dangerous part is that these problems may not show up immediately. The code may still get produced. Tickets may still get closed. Metrics may still look acceptable. But over time, technical debt and knowledge gaps accumulate until the organization discovers that nobody actually understands why the system works the way it does—or how to fix it when something goes wrong.
AI can absolutely make engineering organizations more productive. But AI is an amplifier, not a substitute for engineering judgment.
Cutting experienced engineers while simultaneously increasing dependence on AI may turn out to be one of those decisions that looks brilliant on a spreadsheet right up until the customers start experiencing the consequences.