A software engineer's guide to not getting replaced by AI
Survive AI. Don't get replaced by a GPU.
Layoffs are back with a vengeance. This go around, rising interest rates aren’t the only problem. The exploding investment in AI seems to be a central theme.
Here’s a few recent layoffs:
UPS: 48,000 employees
Amazon: Up to 30,000 employees
Intel: 24,000 employees
Nestle: 16,000 employees
Accenture: 11,000 employees
Ford: 11,000 employees
Novo Nordisk: 9,000 employees
Microsoft: 7,000 employees
PwC: 5,600 employees
Salesforce: 4,000 employees
Paramount: 2,000 employees
Target: 1,800 employees
Kroger: 1,000 employees
Applied Materials: 1,444 employees
Meta: 600 employees
Amazon just announced plans to cut up to 30,000 corporate jobs, its largest layoff ever. Why such deep cuts when the business is still growing? In a word: GPUs. CEO Andy Jassy has warned employees that advances in AI (like autonomous “AI agents” and chatbots) mean fewer people will be needed in certain roles. Amazon is desperately redirecting budget from salaries to capital expenditures for AI infrastructure – essentially trading headcount for GPUs.
The money saved by letting people go is being used to buy the compute power (GPUs) that might replace them.
This trend isn’t unique to Amazon. Across the industry, a new boss has arrived: GPU capex. Cloud providers and tech giants are pouring billions into AI data centers and chips to meet surging demand, while trimming payroll to make room for these costs. Amazon, for instance, committed an eye-popping $100 billion to cloud infrastructure this year, spending over $31 billion in Q2 alone on chips, data centers, and power.
“We have more demand than we have capacity,” Amazon’s finance chief admitted
It may feel dystopian, but it’s our new reality. In late 2025, being a software engineer at a big internet company means living under the specter of the GPU. Meta, for example, just slashed 600 jobs from its AI division in an efficiency push, even while aggressively ramping up its AI capacity elsewhere. Meta’s internal memo admitted its AI unit had become “overly bureaucratic,” and by cutting staff they hope to speed up output and make each remaining person “more load-bearing”.
Notably, Meta is still hiring for its most critical AI teams even after these cuts, which should give you some hints about how to survive.
Your skill at writing code alone is no longer a guarantee of job security. So how can you, as an individual, avoid getting “replaced by a GPU”? I’ll give you my take.
Use AI better than your peers
The harsh truth is that using AI effectively is quickly becoming a non-optional skill for developers. Internal initiatives at multiple companies now encourage (or even require) engineers to incorporate AI in their workflows. This is exactly the kind of boost companies need to justify keeping you on board when others are being let go.
Early adopters of AI are routinely seeing double-digit productivity bumps in software development. Those who ignore these tools risk falling behind their peers. If you aren’t using AI in your day-to-day, you’re making life harder than it needs to be.
What does “using AI better” look like in practice? It means going beyond occasional toy experiments and deeply integrating AI into your problem-solving process.
Adopt AI-augmented development tools. A new breed of IDEs and assistants are emerging to turbocharge coding. Take Cursor, for instance, which is an AI-powered code editor that can autocomplete entire functions and help refactor code intelligently. Or GitHub Copilot, which integrates into your editor to suggest code as you type. Tools like these can handle the grunt work and repetitive patterns, letting you focus on the logic and design. Likewise, keep an eye on NotebookLM (Google’s AI notebook) or other AI-enhanced documentation tools. Being the person on your team who knows how to harness these cutting-edge tools is a huge competitive edge.
Stay current with AI advancements in your field. The landscape is evolving monthly. New libraries, APIs, and models (like OpenAI’s latest, or domain-specific AI tools) will keep emerging. Make it a habit to try out relevant AI tools and read about how other engineers are using them. Showing that you can quickly adapt to and exploit new AI capabilities signals to employers that you’re not going to be left behind.
The bottom line is to make AI your sidekick. If you use AI better than your peers, you become more valuable and perhaps less likely to be seen as “excess human overhead.”
This is the reason a few thousand of you have subscribed to the newsletter. I spend hours staying on top of this so you only have to spend minutes. Join us if you haven’t already!
Be a revenue generator
Not all engineering roles are created equal. In an age of belt-tightening, companies will protect the teams that directly drive profit and growth. This has always been true to some extent, but the AI era is amplifying the divide.
What’s the difference? A profit center is a team or project that clearly generates revenue or business value. A cost center is more about internal support or nice-to-have capabilities that don’t show up directly in the company’s bottom line. For example, in a consumer product company, the engineering squad building new customer-facing features is a profit center; the team maintaining an internal tool or doing research might be viewed as a cost center. In Amazon’s case, AWS (cloud) is a profit center whereas something like an experimental Devices project or an internal HR tool is more of a cost center. We saw Amazon target its People Experience (HR) and Devices divisions for major cuts in this layoff, while its core AWS and AGI (AI research) groups were somewhat more insulated.
So how do you not get replaced by a GPU here? Make sure you’re working on stuff that makes money (or that directly builds the product or AI capabilities that will lead to money). If your current role feels distant from the company’s main value stream, consider moving closer.
Being in a profit center also usually means your work is more visible to senior leadership (since it affects metrics they care about). Use that to your advantage. Make sure you understand how your project impacts revenue or growth, and communicate that impact. For example, if you improved a feature that increased conversion rate or reduced infrastructure cost, quantify it and make sure it’s known. (side note, AI can help with this)
Be really easy to work with
It might sound almost trivial compared to learning AI tools or choosing the right team, but being an easy-to-work-with, positive presence on your team is actually one of the most underrated forms of job security. Attitude and reputation matter.
Being easy to work with means a few concrete things for you as a software engineer:
Communicate and collaborate proactively.
Maintain a positive, adaptable attitude.
Be reliable and low-drama.
In short, make yourself someone that your coworkers and bosses want to keep around. AI can automate a lot of tasks, but it can’t replace the experience of working with a friendly, trustworthy colleague who makes the team more productive. Plus, life is better when you’re good to people.
The fact that you’re reading this means you’re proactive about staying ahead. So if you’re truly ready to take AI seriously and supercharge your engineering career, consider joining the community of like-minded engineers who are doing the same.
Dozens of your peers have already become paid members of our newsletter, unlocking the full archive of deep-dive guides on using the bleeding-edge tools that can make you a 10x engineer in the AI era. We cover practical, tactical tutorials on tools like Cursor, advanced code assistants like Claude Code and ChatGPT, and even NotebookLM for smart note-taking. These guides show you exactly how to incorporate these tools into your workflow and get that extra edge over the competition. By joining us, you’ll not only get access to all these resources, but you’ll also connect with a community that is committed to adapting and thriving in this new landscape.
Companies are investing in AI to stay competitive. You should invest in yourself for the same reason. Adapt, learn, and continually sharpen your skills with AI and beyond. Some members even expense their subscription to their company’s learing and education budget. Here’s an email you can steal!
I’m just not buying it. I know we keep hearing it. I know it’s all over the media. But only an estimated 3% of the world occupations are big tech. Everyone else is still hiring SW engineers. And those big tech companies, ALSO still hiring. The other 97% of the rest of the world occupations are just watching and going about their business. Fear sells.
Regarding this article, your previous analysis on skill adaptation proves particulary prescient as GPU capital expenditures reshape workforce demands.