When it comes to artificial intelligence, business advice that has proven reliable before often seems less important.
Consider OpenAI. Founded in 2015, the artificial intelligence project behind ChatGPT and GPT-4 is already appreciated About 30 billion dollars It is the talk of Silicon Valley. But its success was hardly inevitable. If its founders heeded some of the traditional startup rules, today’s OpenAI might be an obscure company.
OpenAI CEO Sam Altman, who previously led startup accelerator Y Combinator, discussed his company’s unusual rise during Fireside chat This week it was hosted by fintech company Stripe.
“OpenAI went against all of YC’s advice,” Altman told Stripe founder and fellow billionaire John Collison.
He was blown away by the methods: “It took us four and a half years to launch a product. We would be the most capital-intensive startup in the history of Silicon Valley. We were building technology without any idea who our customers would be or what they were going to use it for.”
On Saturday, Altman chirp: “chatgpt has no built-in social or sharing features, you have to register before you can use it, no ingrained viral loop, etc. Seriously questioning my years of advice to startups.”
chatgpt has no built-in social or sharing features, you must register before you can use it, there is no inherent viral loop, etc.
Seriously questioning the years of advice I’ve given startups
– Sam Altman (@sama) May 6, 2023
When asked if potential OpenAI investors shook their fingers and told him he was doing it wrong, Altman replied, “Yeah, I was just kind of like, I really don’t care. Don’t invest.”
Of course, knowing startup rules in and out allows Altman to break them with confidence. In addition to leading Y Combinator, whose success depends on evaluating startups, he also served as CEO of Reddit and is a prominent investor – he was one of the first investors in Stripe.
“Maybe you’re more self-aware, and you don’t have to care so much,” Collison noted, to which Altman replied, “Yeah.”
Greg Brockman, president and co-founder of OpenAI, also spoke this week about ways to break the company’s rules.
“You’re supposed to have a problem to solve, not a technology looking for the solution,” he said Tell the maybe podcast this week. He added that they spent “just a couple of months writing all the different ideas we could work on for both GPT-3 and GPT-4… maybe we could do something medical or something legal.”
Instead they decide to ignore the rule entirely – to great success.
AI is very different, Brockmann concluded: “Every company, every individual, every business is a language business. It has language that flows deeply. So if you can add a little bit of value in existing language workflows, it will be possible for widespread adoption.” wide. “