Hoffman, a partner at the venture capital firm Greylock, made the comments during a This Week in Startups podcast episode released on Wednesday.

The way he sees it, artificial intelligence is advancing at such a pace that we’ll “have a personal assistant for any professional informational task [within] 2 to 5 years.” The change will occur in any role where something is done with information: “I process information, you do something with it—make an investment decision, write a memo, write a prescription, something like that.”

Not that A.I. adoption will be even across sectors, he added: “Now, what the adoption will look like, will it be useful to essential, there’s a variety of things—but as an amplifier, as amplification intelligence vs. artificial intelligence, it is off-the-charts amazing.”

Hoffman, as an early investor in and advisor to OpenAI—the company behind A.I. chatbots ChatGPT and GPT-4—has thought more than most about the implications of the technology. He recently published a book—Impromptu: Amplifying Our Humanity Through AI—co-written by GPT-4.

Hoffman is optimistic about A.I.’s impact on humanity but isn’t oblivious to fears that it will be used to replace white-collar professionals.

Judging by the musings of IBM chief Arvind Krishna, using the “personal assistant” that Hoffman mentioned might become more mandatory than optional.

Krishna wrote in a recent Fortune op-ed, “We must start preparing the workforce for collaboration with A.I. tools.” The tools will “tackle the kind of tasks most people find repetitive, which frees up employees to take on higher-value work.” 

Whether the technology “frees up” workers or makes them redundant is up for debate. Krishna recently said IBM will slow or suspend hiring for back-office jobs. “I could easily see 30% of that getting replaced by A.I. and automation over a five-year period,” he said. (Bloomberg estimated that would translate to more than 7,000 lost jobs, though IBM later clarified to Fortune that instead of a blanket hiring pause the company would be “very selective” about hiring for jobs that aren’t client- or technology-focused.)

The employees who survive A.I.-fueled workforce reductions could be the ones who learn how to best work alongside technology. 

“I’m of the mind that A.I. isn’t going to replace people, but people who use A.I. are going to replace people,” Kara McWilliams, head of ETS Product Innovation Labs, which offers educators a tool that can identify A.I.-generated answers, told the Financial Times earlier this year.

“Unquestionably, many of the tasks in white-collar land will look very different in the next five to 10 years,” Mustafa Suleyman, who cofounded the A.I. lab DeepMind, told attendees of the GIC Bridge Forum in San Francisco this week. “There are going to be a serious number of losers [and they] will be very unhappy, very agitated.” 

In addition, there could be more pressure from bosses to do more in a day thanks to A.I. James Clarke, the CEO of digital marketing firm Clearlink, recently told his staff: “Many content writers today are now exclusively using A.I. to write. I can do that in about 30 minutes of an eight-hour workday. So what do we need to do? Let’s put out 30 to 50 times our normal production.” 

Asked on This Week in Startups how much of the work currently done by humans could be offloaded to tools like GPT-4, Hoffman said it depends on the job. But if you’re doing something like writing reports, for instance, or taking minutes, “then your answer is probably 50% to 80%…what used to take you three hours to do will now take you 15 minutes.” 

Hoffman added on a hopeful note that an employee might well use the extra time to make a report “a lot better within that time frame.” 

Whether the boss will need it to be any better is an open question.