RSS News Feed

Mark Zuckerberg: AI Means You Can Start a Company With a Small Team


Mark Zuckerberg said today’s founders and entrepreneurs should take advantage of the technology he couldn’t when he was building Facebook two decades ago.

“If you were starting whatever you’re starting 20 years ago, you would have had to have built up all these different competencies inside your company, and now there are just great platforms to do it,” the Meta CEO said said at the Stripe Sessions conference this week.

Zuckerberg says using technologies like AI can help today’s founders “focus on the core idea” of a company.

“I think that this is just going to lead to much better quality stuff that gets created around the world because now you’re just being able to have these, like, very small talent-dense teams that are, like, passionate about an idea,” Zuckerberg said.

Replacing the ‘midlevel engineer’

Zuckerberg has talked about the effect of AI on much larger companies, including Meta, on the Joe Rogan Experience podcast in January.

“Probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of midlevel engineer that you have at your company that can write code,” he said.

Of course, LLMs have been known to have troubling hallucinations at times, and companies could see negative repercussions from hollowing out their mid-level engineer ranks.

“Ease of use is a double-edged sword,” Harry Law, an AI researcher at the University of Cambridge, previously told BI. “Beginners can make fast progress, but it might prevent them from learning about system architecture or performance.”

Using AI too extensively in coding could also make scaling or debugging difficult, he warned.

“Security vulnerabilities may also slip through without proper code review,” he said.

Still, companies are finding ways to charge ahead with AI

Y Combinator CEO Garry Tan said in a CNBC interview in March that “vibe coding” will help startups stay leaner by allowing smaller teams of engineers to produce work that would otherwise take a team of 50 to 100 developers.

“I mean, the wild thing is people are getting to a million dollars to 10 million dollars a year revenue with under 10 people, and that’s really never happened before in early stage venture,” he said.

“You can just talk to the large language models and they will code entire apps,” Tan continued. “You don’t have to hire someone to do it.”

Vibe coding, the hot new buzzword in the valley, was coined by OpenAI cofounder Andrej Karpathy in a post on X in February.

“There’s a new kind of coding I call ‘vibe coding,’ where you fully give in to the vibes, embrace exponentials, and forget that the code even exists,” Karpathy wrote in his post. “I’m building a project or webapp, but it’s not really coding — I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.”

Shopify CEO Tobi Lütke recently told managers that before asking to hire someone new, they must prove AI couldn’t do the job better alone.

In March, Anthropic cofounder and CEO Dario Amodei said AI could be “writing essentially all of the code” in 12 months’ time.

Google CEO Sundar Pichai said in the company’s third-quarter earnings call in October that more than 25% of the new code created at the company is generated by AI and then checked by employees.

Pichai said using AI to code boosted the company’s “productivity and efficiency.”

“This helps our engineers do more and move faster,” Pichai said.

OpenAI CEO Sam Altman said in early February that he expected software engineering would look “very different” at the end of 2025.

The use of AI at large tech companies to aid, expedite, or outright do their employees’ work dovetails with the industry’s growing focus on efficiency in recent years.

Zuckerberg proclaimed 2023 a “year of efficiency” for the company, which has seen several rounds of mass layoffs in recent years. Several of its peers have also slashed thousands of jobs as they focus on flattening their organizational structures and pushing out their “lowest performers.”





Source link