Marc Benioff, the billionaire co-founder of Salesforce, has strong opinions about AI. He’s focusing Salesforce on using AI agents for customer service, but he thinks a lot of what people say about AI is just hype. He compares this to how excited he was when Apple’s Steve Jobs sent him one of the first iPhones. Benioff talks to WSJ’s Christopher Mims and Tim Higgins in this episode of Bold Names.
So, what do you think? Is AI really all it’s cracked up to be, or is it just a fantasy? Can AI do what people say, or is it just hype?
Interesting interview! Benioff’s right that there’s a lot of hype about AI right now, but he misses some things too.
As someone working in AI, I can tell you that AI’s capabilities are real. The issue is that some companies are overhyping simple features or making unrealistic claims about artificial general intelligence (AGI).
Take Microsoft’s Copilot, for example. It’s basically just GPT-4 with a few Microsoft integrations—not that revolutionary. On the other hand, newer models like Claude 3.5 and Gemini 1.5 are much better at complex tasks and reasoning.
The real breakthroughs are happening in smaller teams that are pushing AI boundaries. We recently built a real-time web search system that beats others out there, and an intelligent model router that selects the best AI for different tasks.
Still, I agree with Benioff about the need for more realistic expectations. AI is amazing, but it’s not magic. It’s all about how we apply it in practical ways instead of just buying into the hype.
@Tan
The Copilot thing is wild. How can GPT-4 be one of the best but Copilot is so bad? The way it’s prompted in the background is just awful.
I disagree with you about Gemini though. It’s much harder to work with and takes a lot more effort to figure out. It’s not because it’s more powerful, but because it’s just not user-friendly. The reasoning is worse than GPT-4. Claude is better, but for longer sessions, I still prefer GPT-4. It’s just more reliable.
I totally agree that individuals pushing AI forward are the real innovators. Big companies aren’t doing anything special with AI—just trying to shoehorn it into products where it doesn’t really fit.
@Tan
You lost me here. Big tech companies trying to sell enterprise solutions and smaller teams pushing AI aren’t really different. Both can innovate in their own ways.
I agree that some people are too optimistic about AI, expecting things to keep improving as fast as they have been. But CEOs are in a tough spot—they have to keep hyping things to please investors. That’s not surprising.
I also agree that AI needs to be used in practical ways. A lot like the dotcom boom, where anyone could create a website and call it a business, now everyone’s using AI to claim they’ve built something magical. Most businesses setting up with AI don’t really understand the fundamentals. However, there’s still some real innovation happening. Some companies will thrive, but others will fall short when their unrealistic expectations aren’t met.