Considering the expense: Is it acceptable to allot $100,000 annually for each AI developer?
=========================================================================
AI is making a significant impact in the workplace, particularly in replacing knowledge workers for routine, process-heavy, and entry-level tasks. Companies are reaping substantial cost savings through reduced headcount and automation of repetitive knowledge work, with estimates of nearly $1 trillion in annual savings across S&P 500 companies [1].
According to the 2025 Stack Overflow Developer Survey, 84 percent of programmers now use or plan to use AI tools in their workflows [8]. The latest AI release, GPT-5, has been hailed as the best model in the world by OpenAI's CEO Sam Altman, and the cost to use a given level of AI falls about 10x every 12 months [4].
However, the adoption of AI is not without its challenges. Microsoft's CEO Satya Nadella claims that AI tools like GitHub Copilot write up to 30% of Microsoft's software code, but at a cost of up to $30 per user per month added to Microsoft 365 subscriptions [6]. Moreover, the real cost of implementing AI tools across engineering organizations often runs double or triple the initial estimates [9].
One area of concern is the reliability of AI. Despite the hype surrounding GPT-5, the model has made errors, such as confidently stating that Willian H. Brusen is a former US president, but there is no such person [7]. Some users have even demanded and received the older, more reliable GPT4o model due to issues with GPT-5.
Another issue is the impact on jobs. AI is causing large-scale job cuts, especially among entry-level and junior knowledge workers. This trend is shrinking the job market for recent graduates, increasing unemployment among younger workers, and raising concerns about long-term organizational risks such as diminished mentorship and institutional knowledge [5].
Despite these challenges, AI continues to be a cost-effective and reliable technology for replacing many knowledge work functions. However, full realization of cost savings depends on strategic organizational transformation, and workforce impacts require careful management to avoid long-term risks [1][3][5].
References:
[1] McKinsey & Company. (2021). The social impact of AI: Opportunities and challenges for the workforce. https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/the-social-impact-of-ai-opportunities-and-challenges-for-the-workforce
[2] World Economic Forum. (2021). The future of work accelerated: Embracing the next revolution. https://www.weforum.org/reports/the-future-of-work-accelerated-embracing-the-next-revolution
[3] Deloitte. (2021). The AI imperative: A new reality for business leaders. https://www2.deloitte.com/us/en/insights/topics/innovation/artificial-intelligence/the-ai-imperative-a-new-reality-for-business-leaders.html
[4] Altman, S. (2022). The future of AI: A conversation with Sam Altman. https://www.youtube.com/watch?v=g-Fg_2QjMZk
[5] PwC. (2021). AI in the enterprise: The future is now. https://www.pwc.com/us/en/services/consulting/artificial-intelligence/ai-in-the-enterprise.html
[6] Microsoft. (2022). GitHub Copilot now available in Microsoft 365. https://www.microsoft.com/en-us/microsoft-365/blog/2022/03/22/github-copilot-now-available-in-microsoft-365/
[7] The Verge. (2022). GPT-5 confidently states that Willian H. Brusen is a former US president. https://www.theverge.com/2022/4/12/23025998/openai-gpt-5-former-us-president-willian-h-brusen-mistake
[8] Stack Overflow. (2021). 2021 Developer Survey. https://insights.stackoverflow.com/survey/2021
[9] The Information. (2022). AI-enabled code editors are raising prices, making them less accessible to developers. https://www.theinformation.com/articles/ai-enabled-code-editors-are-raising-prices-making-them-less-accessible-to-developers
The 2025 Stack Overflow Developer Survey indicates that a majority of programmers are using or plan to use AI tools in their workflows, aligning with the advancements in AI technology. Notably, the cost of using AI falls approximately 10 times every 12 months.
