Why Salesforce CEO Marc Benioff Isn’t Buying the AGI Buzz
Artificial intelligence continues to grow from strength to strength. It already plays a major role in our daily lives, powering everything from personal assistants to business-critical applications. Large language models (LLMs) in 2025 have advanced rapidly, as shown by the release of ChatGPT-5 and other cutting-edge AI models.
Within the Salesforce ecosystem, we’re also seeing AI take shape. Agentforce is slowly being adopted at the enterprise level, as Salesforce continues to enhance its flagship AI tool.
But the potential “next frontier” – artificial general intelligence (AGI) – is far more daunting. For those unfamiliar, AGI refers to AI systems capable of human-level cognitive abilities. Think Jarvis from Iron Man, Samantha from Her, or Data from Star Trek. It’s a level of technology that can reason, learn, and act on par with, or even beyond, humans.
Currently, forecasting for AGI varies widely – some industry experts suggest it could emerge within a few years, while others believe it’s still decades away. Salesforce CEO Marc Benioff recently weighed in, stressing that we are nowhere near achieving true AGI.
“We Have All Been Sold A Lot of Hypnosis”
Appearing on the VC20 podcast, Marc Benioff addressed the topic of AGIs in regards to Amazon’s AGI head. David Laun, the head of Amazon’s AGI Lab, has stated that fewer than 1,000 people worldwide qualify as elite AI talent, prompting the Salesforce CEO to call this “a complete oxymoron.”
“You’re talking to someone who is extremely suspect if anyone uses those initials (AGI). I think that we’ve all been sold a lot of ‘hypnosis’ around what’s about to happen with AI,” he said.
“[It’s] not that it couldn’t happen one day, but it’s not the state of technology today.
“When you look at large language models, they are two things: a finite set of algorithms, which have gotten a lot better, and two, a relatively finite set of data that has come off the internet. Those two things together really have provided a state-of-the-art LLMs today. When you work with these LLMs, it’s very cool. But it’s not a person, it’s not intelligent, it’s not conscious, it doesn’t have a childhood, it doesn’t have compassion. It’s not a being.”
He claimed that there is currently some level of hypnosis around AI and what is currently possible, as well as what is going to happen. At present, AI should be acknowledged for the uses it has, but it’s important not to glaze over its plethora of issues.
The important thing to realize here is that Marc Benioff evidently has the belief that AGI will be here at some point, largely tied to his enthusiasm for AI. However, there is no denying that AGI is somewhat of a futuristic concept. After all, do we even want AGI?
When Can We Expect AGI?
If you were to ask ten experts when AGI is expected to arrive, you’d likely receive ten different answers.
Prediction markets like Metaculus currently expect “weakly general AI” by the late 2020s, with stronger forms emerging by the early 2030s. On the other hand, large surveys of machine learning researchers are more cautious, giving a 50% chance of “human-level AI” by the 2040s.
Besides Benioff, a number of high-level CEOs have also shared their predictions over the last couple of years. Demis Hassabis (Google DeepMind) puts AGI at five to ten years away, while Dario Amodei (Anthropic) has suggested as early as 2026. Sam Altman (OpenAI) believes we’re “a few thousand days” away, and Jensen Huang (NVIDIA) points to around five years’ time.
Others, like Yann LeCun (Meta), are skeptical that today’s large language models will ever get there, while Satya Nadella (Microsoft) downplays the race altogether, stressing usefulness over labels.
It’s important to consider that although it is difficult to determine exactly when AGI comes into fruition, it’s very likely that it’s already in the works.
Salesforce, for one, had been working on its AI products and research long before any news of them became public or AI had even entered the mainstream market, and this practice certainly isn’t unique.
Not only that, but AGI is currently being tied to the age-old belief of the tech gap between consumers and government/leaders, especially as something as powerful as an AGI is likely to fall under the Invention Secrecy Act.
So, by the time anyone can see or get their hands on an AGI, it will be interesting to see how it is delivered.
Whether AGI is actually five or fifty years away, the current spread of predictions tells us more about human bias than actual machine capability. Despite some much-needed skepticism around AGI, there’s no arguing that many high-profile companies are working towards achieving it – it’s in a similar vein to the space race, if you will.
Some see it as an inevitable aspiration, while others are noticing the reckless and dangerous side of pushing for such technology so quickly.
Like the space race, AGI is much less about the finish line and more about the breakthroughs these companies are making along the way – and these breakthroughs are quickly shaping how businesses, including Salesforce, are deploying AI in the present.
Final Thoughts
For many, the AGI is the peak of AI technology. But for something as powerful as a human-like computer to be created, it will require considerable development, research, resources, and guardrails, meaning that we will likely not see one for a very long time.
Even CEO Marc Benioff, who is known for being extremely pro-AI, holds conservative views on this matter, indicating that we should perhaps be mindful of our expectations for AGI, helping to develop our existing AI into the best tool it can be.
The post Why Salesforce CEO Marc Benioff Isn’t Buying the AGI Buzz appeared first on Salesforce Ben.
