What you need to know
- In a recent interview, OpenAI CEO Sam Altman indicated that AGI might be here sooner than most people think, despite reports of stunted AI development due to scaling laws.
- The executive says the AGI moment won’t feature the safety concerns expressed, further indicating that it will have “surprisingly little” societal impact.
- Sam Altman says there’s a long continuation of advances between AGI and superintelligence, with great expectations of AI agents in 2025.
Despite claims that top AI labs, including OpenAI, Google, and Anthropic, are struggling to develop advanced AI models due to a lack of high-quality data for training, Sam Altman is seemingly certain OpenAI will hit the coveted AGI benchmark in 2025. The executive shared a cryptic message on social media indicating “there’s no wall,” potentially addressing the reports claiming scaling laws have begun to stunt AI advances.
The OpenAI CEO vaguely indicated that the AI firm could achieve the AGI benchmark in 2025. Interestingly, Altman claimed the benchmark’s societal impact would be “surprisingly little.” He further indicated that AGI is achievable with current hardware. However, this isn’t the case, he claims “you’ll be happy to have a new device.”
Sam Altman recently shared more AGI insights while speaking at The New York Times Deal Book Summit:
“Our intention is to treat AGI as a mile-maker along the way. We have left ourselves some flexibility because we don’t know what will happen but my guess is we will hit AGI sooner than most people in the world think and it won’t matter much less.”
As you may know, there are a lot of safety and privacy concerns around the rapid development of AI. Multiple OpenAI staffers, including the former lead of superalignment, Jan Leike, left the firm, citing safety concerns.
“Building smarter-than-human machines is an inherently dangerous endeavor,” Jan stated. Perhaps more concerning, the executive claimed OpenAI was seemingly more focused on shipping shiny products while safety procedures took a backseat.
More recently, a damning report indicated OpenAI sent invites to GPT-4o’s launch party even before testing began, ultimately mounting pressure on the safety team to rush through critical safety procedures in under a week. While OpenAI admitted that the product’s launch was stressful, it categorically indicated that it didn’t cut any corners.
And a lot of the safety concerns that we and others expressed actually don’t come at the AGI moment.
OpenAI CEO, Sam Altman
Altman previously indicated that AGI would whoosh by. “I think in 5 years, it looks like we have an unbelievably rapid rate of improvement in technology itself. People are like, man, the AGI moment came and went,” stated the CEO.
Interestingly, a former OpenAI researcher indicated that while the firm might be on the verge of hitting the AGI benchmark, it’s not prepared or equipped to handle all that it entails.
During the interview, Altman indicated that there will be a long continuation from the AGI benchmark to superintelligence, which he claimed was “a few thousand days away.”
OpenAI big plans for AI systems in 2025 and beyond
As you may know, major tech AI labs are venturing into the AI agents space, with Anthropic, Salesforce, and Microsoft chasing the crown. Salesforce CEO Marc Benioff has been on Microsoft’s case, claiming the tech giant has done a tremendous disservice to the AI industry while referring to Copilot as the new Microsoft Clippy.
OpenAI is also slated to join the agentic arm’s race in January with its just-announced AI agent dubbed Operator. Like most of the available AI agents, it will help control computers and perform tasks autonomously.
Sam Altman further disclosed that OpenAI plans to dabble in the AI agent space more in 2025. He claims there will be AI systems that people will look at and say, “Wow, I did not expect that.”
