{
  "id": 13314630,
  "title": "Same AI, Four Different Jobs: Startup, Mid-Size, Big Tech, IT Services",
  "url": "https://urgent.news/2026/10/10/same-ai-four-different-jobs-startup-mid-size-big-tech-it-services",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-10T03:35:00.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/debashish_ghosal/same-ai-four-different-jobs-startup-mid-size-big-tech-it-services-521k"
  },
  "original_language": "en",
  "account": "In March 2025, a quarter of the W25 batch at Y Combinator reportedly had codebases that were about 95% generated by AI, according to managing partner Jared Friedman. At the same time, Stack Overflow's survey found that 28% of developers believed their company's IT or Information Security teams had strict rules prohibiting the use of AI agent tools. As the source indicates, despite the use of the same models and technology, engineers in different company sizes experience vastly different working lives due to the impact of AI.\n\nIn a startup, the engineer's role changes significantly. With 95% of the code generated by AI, the engineer becomes highly technical and capable of building their product from scratch. They now spend most of their time reviewing code and finding bugs, as there is no existing infrastructure to maintain. This shift places a heavy emphasis on speed and ownership, but also exposes the engineer to potential risks such as being the only person aware of the system's inner workings.\n\nIn a mid-size company, AI is seen as a tool to increase impact by 2x within a year. Engineers are given more autonomy and are encouraged to self-report their progress, as the organization places less emphasis on formal measurement. This setting offers the engineer a voice in how AI is evaluated, but also carries the risk of losing trust if the company grows too large to maintain that level of personal accountability.\n\nAt large enterprises and big tech companies, AI becomes a metric used to measure performance. Engineers face pressure to work faster than sustainable rates and have to deal with the surveillance of their work. Their time is consumed by untracked, invisible tasks, such as reviewing AI output and fixing bugs. According to a 2026 survey, 54% of developers at large enterprises fear that their performance reviews will be based on AI data, while 46% report experiencing unsustainable workloads and privacy concerns.\n\nFinally, in IT services, AI is transforming the way services are priced. With persistent clients demanding 25% to 30% less work delivered faster, IT service companies are tying contracts to outcome measures. The Nifty IT index saw a decrease of 20% in 2026, reflecting the impact of AI on the industry. For engineers in this field, the challenge lies in adapting to the new pricing models and ensuring their skills remain relevant in an increasingly automated environment.",
  "summary": "In March 2025, Y Combinator managing partner Jared Friedman said that a quarter of the W25 batch had codebases about 95% generated by AI. That same year, in Stack Overflow's survey, 28% of developers agreed that their company's IT or InfoSec team has strict rules that don't allow them to use AI agent tools. Same year, same models, same Twitter timeline, and completely different working lives.…",
  "key_points": [
    "25% of Y Combinator's W25 batch had 95% AI-generated codebases in March 2025",
    "Startup engineers become highly technical, focusing on code review and bug fixing",
    "Large enterprises pressure engineers to work faster, face privacy concerns"
  ],
  "editors_take": "The same AI technology leads to vastly different working lives for engineers in startups, mid-size companies, big tech, and IT services, with varying impacts on their roles, autonomy, and job security.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}