OpenAI's CEO Sam Altman has stated that the company will have artificial general intelligence (AGI) by the end of 2026, provided one accepts his definition of the term. According to a TIME report, Altman's confidence is rooted in the development of Astra, OpenAI's upcoming model family, which functions as an automated AI research intern. The model can write code, run experiments, and report results, significantly reducing the time required for tasks that would otherwise take a human researcher about a week. Altman believes Astra represents the first model capable of inventing new things in a meaningful way, a key characteristic of AGI.
Astra enables persistent agents, virtual coworkers that operate independently over extended periods. Altman highlighted the potential for recursive self-improvement, where AI could enhance its own capabilities. However, opinions remain divided on whether such a system can truly achieve the level of autonomy and generalization required for AGI. OpenAI defines AGI as highly autonomous systems that outperform humans at most economically valuable work, emphasizing an economic perspective. Critics argue that more is needed, including a robust understanding of the world, which current large language models (LLMs) may lack.
OpenAI is also developing a puck-shaped hardware device for early 2027, designed to sense its environment and communicate via voice. The company is focusing on tighter product development, combining Codex with ChatGPT in a project called 'The Merge,' resulting in ChatGPT Work. This aims to deliver agentic AI to a broader audience by enabling task execution rather than just answering questions. OpenAI is also testing new monetization strategies, including ads within ChatGPT, which have shown promise in early tests.
Source: thedecoder
