OpenAI's CEO, Sam Altman, has made bold predictions about the future of AI, suggesting that it will soon be capable of discovering new knowledge and solving complex business problems, akin to junior-level employees.

Speaking at the Snowflake Summit 2025, Altman elaborated that users are increasingly assigning tasks to AI agents, treating them similarly to managing entry-level staff, with feedback loops that enhance their performance.

These AI agents are evolving to take on tasks, generate outputs, receive feedback, and improve over time, marking a significant shift in how businesses might operate.

OpenAI has introduced Codex, an AI tool that assists programmers by writing code, debugging, and running tests, which is already being utilized by OpenAI engineers.

Altman described the recent launch of GPT-4.5 as a groundbreaking model that offers an experience akin to conversing with a thoughtful person, although access is currently restricted to Pro subscribers due to GPU shortages.

As AI technology advances, companies like Shopify are now requiring managers to justify why positions cannot be filled by AI, while Duolingo is planning to replace contract workers with AI solutions.

Recent studies indicate that AI is already replacing human workers in roles vulnerable to automation, such as database administrators and IT specialists, raising concerns about job displacement.

An analysis by economist Zanele Munyikwa revealed a 19% decline in AI-doable tasks in online job postings since the release of ChatGPT in late 2022, with some roles seeing hiring downturns of up to 31%.

Despite skepticism regarding AI's role in scientific discovery, successful examples exist, such as Microsoft's Discovery platform, which identified a new chemical for data center cooling in just 200 hours.

In a related note, former OpenAI chief scientist Ilya Sutskever previously indicated the need for a 'doomsday bunker' before the release of artificial general intelligence (AGI), highlighting the potential risks involved.

Sutskever's comments about the bunker were made during a meeting with OpenAI scientists, underscoring the serious implications of AGI's development.

Altman noted that current users are engaging with AI agents in a manner similar to managing entry-level employees, reflecting a significant shift in workplace dynamics.