Former Blizzard engineer's 6-month job search completely failed, bluntly saying the software industry is doomed

In 2026, the unemployment rate in the tech industry is already at the highest level since the dot-com bubble of 2001. After a software engineer who had worked at Blizzard for 7 years spent 6 months job hunting with no results, he described the absurd scene of AI proctoring and the reality that companies simultaneously require developers to list the number of tokens burned for Claude as a KPI.
(Background: Anthropic was “blocked” by the U.S. government and took down the Fable model; foreign media point to three major concerns: fear of helping China’s open-source AI)
(Additional context: Musk has transformed into a compute arms dealer! SpaceX signs a $6.3 billion deal to lease Nvidia GB300 to boost open-source AI)

An anonymous software engineer, in a recent entry on his personal blog, recorded an experience from the past half year in interviews that repeatedly hit dead ends: “The Work and Software Industry Are Done.” He said he has about 10 years of development experience, including 7 years working at the well-known gaming company Blizzard Entertainment, and that in June 2025 he was laid off along with the entire team.

During the 6-month job search, he summed it up with a few words: “Made it to the final round only to be rejected,” “Applications sent out to a black hole,” and “Resumes perfectly match, yet no responses at all.”

And what infuriated him most wasn’t the rejection letters—it was the screening mechanism itself. Coderpad, HackerRank, and various AI proctoring software: the design logic of these tools is full-screen lock, preventing candidates from checking API files, simulating an environment that simply doesn’t exist in real work. He said bluntly that the people who designed this process clearly have never actually written code.

Even more absurdly, in this “no AI allowed” testing environment, he tried to follow the rules, but other candidates used their phones along with AI tools to cheat effortlessly. In contrast, honest people who followed the rules end up at a disadvantage in the screening.

He didn’t provide evidence in the article to support this claim, but the logic itself is a structural paradox: a system that says it wants to test real ability creates incentives for cheaters to succeed.

AI simultaneously sits in two roles

Now, AI in the workplace plays two roles at the same time: gatekeeper (the interview side) and substitute (the production side).

Data supports this logic. Research shows that after developers use AI tools, productivity per sprint increases by about 40% to 55%. In simple terms, a 10-person team can produce the work that used to take 15 people. For companies, this gives a clear economic rationale to reduce headcount rather than expand scale.

This also explains why newly graduated job seekers are in an even worse situation. This former Blizzard engineer observed that many companies currently don’t even hire junior developers; they expect models from companies like Anthropic to directly replace the need to bring new hires up to speed and train them.

The data backs up this observation. The number of employed software engineers under age 25 has already fallen by nearly 20% from the 2022 peak. This isn’t just market fluctuation—it’s a structural contraction.

90% of managers say “no impact,” but the data says otherwise

In 2026 layoff data, 56% of the events—affecting about 156,270 employees—explicitly cite AI, automation, or machine learning as the main cause in announcements or in financial reports. This is the highest proportion so far. At the same time, however, another survey shows that 90% of corporate managers say AI has “no impact” on employment at their company.

The gap between these two sets of figures reveals a systemic fracture between corporate public communication and actual decision-making: layoff announcements say “AI-driven efficiency improvements,” but when managers face surveys, they claim there is no impact on employment.

This former Blizzard engineer offers no policy recommendations, and he calls for nothing—because there are no signs, at least for now, that the systemic shift brought by AI is going to reverse. He simply wrote down the experience of being “worn out to exhaustion” by it all.