Want to work at OpenAI? Curiosity and grit matter more than a Ph.D, the head of ChatGPT says
Instead, they value something far more human: curiosity and initiative.
In a Tuesday episode of OpenAI 's podcast, Turley, the head of ChatGPT, said the single most important trait he looks for isn't technical prowess, but curiosity.
"It's the number one thing that I've looked for," Turley said. "And it's actually my advice to students when they ask me, 'What do I do in this world where everything's changing?'"
"There's so much we don't know," he said. "There's a certain amount of humility you have to have about building on this technology."
He added that working with AI is less about knowing all the answers and more about asking the right questions.
"When it comes to working with AI," he said, "It's asking the right questions that is the bottleneck, not necessarily getting the answer."
Mark Chen, OpenAI's chief research officer, echoed that sentiment.
He said he didn't have much formal AI training when he joined the company in 2018.
"I also came into the company as a resident without much formal AI training," he said. "Even on research, I think increasingly less, we index on you have to have a Ph.D. in AI."
Instead, Chen said OpenAI looks for candidates with "agency"—the ability to identify and solve problems independently.
"It's really about being driven to find, 'Hey, here's the problem. No one else is fixing it. I'm just going to go dive in and fix it,'" he said.
That kind of self-direction is echoed by other OpenAI veterans.
Peter Deng, who served as OpenAI's VP of consumer product and previously led product teams at Meta, put it bluntly on an episode of Lenny's Podcast last month: "In six months, if I'm telling you what to do, I've hired the wrong person."
That expectation aligns closely with OpenAI's own culture of hands-on ownership.
Turley recalled that ChatGPT itself was born out of a hackathon-style sprint, with people from across teams — including infrastructure and supercomputing — coming together to ship a product quickly.
"Fundamentally," Turley said, "we just have a lot of people with agency who can ship."
He added that this is what makes OpenAI unique and is a key focus of the company's hiring process.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
18 minutes ago
- Yahoo
Nvidia is on track to become the most valuable company in history
The AI chipmaker Nvidia's shares hit a new all-time high on Thursday, briefly giving the company a market capitalisation of $3.92 trillion (€3.33tn), the highest in history for any company. This surpassed Apple's record of $3.91tr set in December 2024, even though Nvidia's market capitalisation dipped once again below this level at market close. The chipmaker's shares traded as high as $160.98 at their peak on Thursday, before the price dipped below this level, placing the market capitalisation at around $3.89tr when daily trading wrapped up. Tech companies' shares benefitted from a better-than-expected nonfarm payrolls report in the US, an indicator of a resilient US economy. This optimism was boosted by forecasts that businesses would continue to spend on AI advances, boosting demand for AI chips. Related Nvidia shares surge as earnings beat despite chip export restrictions to China Nvidia expects $5.5bn revenue hit as US curbs chip exports to China Nvidia shares are up more than 50% in just less than two months. Analysts expect that the company will break the valuation record soon and retain its elevated share price by the close of the trading day. 'Chip giant Nvidia is on track to achieve a new closing high,' said Dan Coatsworth, investment analyst at AJ Bell, adding that the 'AI revolution is still intact'. AJ Bell head of financial analysis Danni Hewson added that, 'After all the gloomy predictions that this might be the year the AI bubble bursts, Nvidia's found another gear. The chipmaker is on track to smash a coveted record and become the world's most valuable company ever.' The value of Nvidia currently is more than three times the total market capitalisation of the stock market in Spain and more than four times that of the Italian stock exchange. Error while retrieving data Sign in to access your portfolio Error while retrieving data
Yahoo
21 minutes ago
- Yahoo
Artificial intelligence to help growers assess hurricane crop damage
The Brief University of Florida scientists are developing new technology using artificial intelligence (AI) to make assessing crop damage after hurricanes faster and easier. Farmers will be able to get answers about their fields and crops by interacting with an AI assistant through a chat-based interface that will eventually work with smartphones and computers. TAMPA - University of Florida scientists are developing new technology using artificial intelligence (AI) to make assessing crop damage after hurricanes faster and easier. Imagine you're a farmer, and, unfortunately, a hurricane has just come through: How do you assess crop damage? Right now, it might be through surveys, field visits, and drones. But wouldn't it be easier, and faster, if you could just ask your phone or computer: What percentage of my field is damaged? That's the kind of technology scientists at the University of Florida Institute of Food and Agricultural Sciences (UF/IFAS) are working on right now. "The system is like ChatGPT combined with some maps," Nikos Tziolas, a faculty member at the UF/IFAS Southwest Florida Research and Education Center, said. Dig deeper Known as the Gaia Bot, farmers interact with an AI assistant through a chat-based interface by typing in questions about their crops and fields. "We just ask if my crops look healthy," Tziolas explained while showing FOX 13 an example. "And in real time, (the AI system) generates this map." OTHER NEWS: Higher rain chances for Bay Area on Thursday, more heavy rain possible for July 4th Farmers type in simple questions and also respond when asked for more information about their fields and crops. Then the conversational AI system analyzes satellite imagery, weather information, and soil maps developed in UF labs before responding. "The system automatically retrieves this data without any experience from the user," Tziolas added. By quickly generating maps and data for specific fields, this technology – which is funded by a USDA-National Institute of Food and Agriculture grant -- can help tell farmers which areas saw the most flooding and how their crops are looking before and after a storm. What's next UF/IFAS scientists expect this technology to be available to farmers at no cost later this hurricane season in the fall. Agricultural losses can be costly after hurricanes. UF estimates the state saw up to $640 million in agricultural damage during Milton. The Source FOX 13's Ariel Plasencia gathered the information for this story.
Business Insider
an hour ago
- Business Insider
'I don't trust it, I don't like it': Lawmakers sound off on why they don't use AI
AI has become a big topic on Capitol Hill, and some lawmakers are finding ways to use it in their daily lives. But while some are using Grok to come up with one-liners, asking ChatGPT to punch up speeches, or deploying AI to scrutinize their own records, plenty of their colleagues still haven't gotten into it. "I haven't had the chance to really explore the tools," Democratic Rep. Gabe Vasquez of New Mexico told BI. "I'm a little behind the times in that sense." There's no single reason why. Some remain skeptical about the technology's ability to provide reliable information. Others fear that if they use it, they'll erode their own faculties. At least one has publicly fought with an AI chatbot on X. And often, it simply comes down to a lack of interest. For Sen. Tim Kaine, it's a combination of several of those things. "I feel like I'm so comfortable with my ability to access information that I don't really need it," the Virginia Democrat said, explaining that he's never used AI chatbots "voluntarily." He does, however, encounter artificial intelligence involuntarily: Recently, his email client began furnishing him with AI-generated summaries of his email. "It's like, first, I didn't ask for it. Second, it's got a high error rate," Kaine said. "Third, to do an AI summary of emails I can read myself is a use of energy, somewhere, that I have to feel like is kind of an excess." 'I still like to compose original thoughts' BI spoke to roughly a dozen lawmakers in June about their use, or lack thereof, of AI chatbots like Grok and ChatGPT. Polling has shown that a growing number of Americans are using AI in the workplace, with a recent Gallup poll finding that 40% of employees use it a few times per year and 27% use it frequently or daily. Republican Rep. Greg Murphy of North Carolina, who remains a practicing urologist in additional to his congressional duties, says he's "dabbled" with AI, primarily for medical questions. He says he doesn't use it for his official work, and he doesn't want his staff to do so either. "I don't want people writing speeches for me, or doing letters to constituents with that," Murphy said, adding that if his constituents "wanted something from ChatGPT, they could have Googled it themselves." Sen. Josh Hawley, a fellow Republican, takes an even harder line against AI — he doesn't use AI chatbots out of principle. "I don't trust it, I don't like it, I don't want it being trained on any of the information I might give it," the Missouri senator, a frequent critic of Big Tech, told BI. Some lawmakers' reservations are rooted in their hesitance to rely on technology for something they already do themselves. "I've just never felt the need, and I like being creative," Democratic Rep. Alexandria Ocasio-Cortez of New York told BI. "Writing is a use-it-or-lose-it skill, and I don't want to lose it." "I still like to compose original thoughts, and not rely on technology to help state what I feel," Democratic Rep. Jim McGovern of Massachusetts said. Large language models are also known to confidently deliver faulty information — known as "hallucination." For Sen. Elizabeth Warren, that's enough to discourage her from using it. "I might as well go out on the street and ask a random person a question and see what words they come up with," the Massachusetts Democrat told BI. "Maybe they're right, but then again, maybe they're not, and the only way to know that is to do the research yourself."
