Latest news with #ThePragmaticEngineer
Business Insider
a day ago
- Business
- Business Insider
'Keep AI on the leash' because it's far from perfect, says OpenAI's cofounder Andrej Karpathy
Andrej Karpathy thinks we're getting way too excited about AI, especially when it comes to deploying agents that act without supervision. In a keynote at an event hosted by Y Combinator earlier this week, the computer scientist said people need to "keep AI on the leash." The OpenAI cofounder said current large language models still make mistakes no human ever would. Karpathy likened LLMs to "people spirits" — uncanny simulations of human intelligence that hallucinate facts, lack self-knowledge, and suffer from "amnesia." "They will insist that 9.11 is greater than 9.9 or that there are two R's in 'strawberry,'" Karpathy said in a talk published on Y Combinator's YouTube channel on Thursday. "They're going to be superhuman in some problem-solving domains and then they're going to make mistakes that basically no human will make." Even though LLMs can churn out 10,000 lines of code in seconds, he said, that doesn't mean developers should sit back and let them run wild. "I'm still the bottleneck," he said. "I have to make sure this thing isn't introducing bugs." "It gets way too overreactive," he added. Karparthy urged developers to slow down and write more concrete prompts. "I always go in small incremental chunks. I want to make sure that everything is good," he said. "It makes a lot more sense to spend a bit more time to be more concrete in your prompts, which increases the probability of successful verification, and you can move forward," he added. Karparthy did not respond to a request for comment from Business Insider. The OpenAI cofounder coined the term "vibe coding" in February to describe the process of prompting AI to write code. The idea, he said, is that developers can "fully give in to the vibes" and "forget the code even exists." AI still needs supervision Karpathy isn't the only one urging caution. Bob McGrew, OpenAI's former head of research, said on an episode of Sequoia Capital's "Training Data" podcast earlier this week that human engineers are still essential — not just to guide AI, but to step in when things get messy. When something goes wrong or if a project "becomes too complicated for AI to understand," a human engineer can help break the problem down into parts for an AI to solve. AI agents are like "genies," said Kent Beck, one of the authors of the seminal "Agile Manifesto" — they'll often grant your wish, but not always in the way you'd like them to. "They will not do what you mean. They have their own agenda," Beck said on a recent episode of " The Pragmatic Engineer" podcast. "And the best analogy I could find is a genie. It grants you wishes, and then you wish for something, and then you get it, but it's not what you actually wanted." Beck also said results are so inconsistent that using AI to code can sometimes feel like gambling. Despite the nascent tech's limitations, even the biggest tech companies are betting on AI for the future of coding. AI writes more than 30% of Alphabet's new code, up from 25% last year, said CEO Sundar Pichai on the company's most recent earnings call.
Business Insider
4 days ago
- Business
- Business Insider
He helped write one of the seminal texts about software engineering. Here's what he thinks about AI agents.
AI agents are like "genies," said Kent Beck, one of the authors of the seminal "Agile Manifesto" — they'll often grant your wish, but not always in the way you'd like them to. "They will not do what you mean. They have their own agenda," Beck said on a recent episode of The Pragmatic Engineer podcast. "And the best analogy I could find is a genie. It grants you wishes, and then you wish for something, and then you get it, but it's not what you actually wanted." After 50 years in programming, including a stint at Facebook and a hand in the foundational text about agile software development, Beck said he's now having the most fun of his entire career — and it's partly thanks to AI agents, even despite their unpredictability. "I'm trying all of the tools," he said. "Because right now, nobody knows what process is going to work best. Nobody knows anything. We should all be trying all the things that we can imagine, and then the truth will emerge out of all of that. So that's what I'm doing." Beck was one of the 17 authors of the 2001 paper the " Agile Manifesto," which outlines four values and 12 principles for faster software development. The paper inspired the "agile method" of software development, which became commonplace in Silicon Valley because of its efficiency-boosting techniques. Today, AI is already being used to replace human labor, particularly in programming, where it's hitting early-career engineers the hardest while simultaneously speeding up software development. Given the current state of coding, Beck believes that fundamental organizational skills are increasingly important rather than particular technical specializations. "So, having a vision, being able to set milestones towards that vision, keeping track of a design to maintain the levels or control the levels of complexity as you go forward," Beck said. "Those are hugely leveraged skills now, compared to, 'I know where to put the ampersands, and the stars, and the brackets in Rust.'" Though Beck does incorporate AI into his process, he doesn't necessarily trust it, he said. The technology isn't consistent enough to be relied upon. "Sometimes it even seems like the agent kind of has it in for you," he said. "'If you're going to make me do all this work, I'm just going to delete all your tests and pretend I'm finished, ha, ha, ha.'" Results are so inconsistent, he added, that using AI to code can sometimes feel like gambling. "It feels like a rat and the pellet," he said. "It's like there's just a run button and I have to click it every time. And I click it and it is a dopamine rush because this is exactly like a slot machine. You've got intermittent reinforcement, you've got negative outcomes and positive outcomes." Once in a while, though, the output will be just right — and Beck will be tempted to spin the wheel all over again. "The distribution is fairly random, seemingly. So it's literally an addictive loop to have it. You say, 'Go do this thing.' And then sometimes it's just magic."
Indian Express
02-06-2025
- Business
- Indian Express
Should you intern at a big tech firm or startup? OpenAI engineer offers career advice
Many young college graduates around the world, particularly in India, aim to become software engineers. But they often face an important question: Should they kick off their careers at a big tech company such as Google or Microsoft? Or should they take a chance on a startup like OpenAI? An OpenAI engineer believes that interning at both a big tech company and a startup will help students get a clearer sense of the career path they want to take. Janvi Kalra, who is part of the technical staff at OpenAI, said that students should diversify their experiences after college. In an appearance on an episode of the podcast 'The Pragmatic Engineer', Kalra said that both tracks have positives and negatives. Kalra's comments come at a time when entry-level tech jobs are being directly impacted by AI. SignalFire, a VC firm that analyses job movements across 80 million companies on LinkedIn, reported that major tech companies—including Meta, Microsoft, and Google—recruited fewer recent graduates in 2024 compared to previous years. New graduates accounted for just 7 per cent of new hires in 2024, down 25 per cent from 2023 and over 50 per cent from pre-pandemic levels in 2019. At startups, the rate of new graduate hiring dropped from 30 per cent in 2019 to under 6 per cent in 2024. After earning a degree in computer science from Dartmouth College, New Hampshire, US, Kalra interned at Microsoft and Google. She then worked at a productivity startup called Coda before joining OpenAI, as per her verified LinkedIn profile. She said that given the opportunity, gaining experience with both startups and larger firms early on in one's career would be beneficial. 'Given that Big Tech and startups are such different experiences and you learn so much at each, it would be more educational to do one startup internship and one Big Tech internship to get a very robust overview of what both experiences are like very early,' she said. According to Kalra, some of the upsides of joining big tech companies include financial benefits, potential prestige, less pressure, more time to work on projects, and learning to build software at scale. 'It's very different to build something that works, versus build something that works when it's swarmed with millions of requests from around the world and Redis happens to be down at the same time. Very different skills,' she said. 'There are also practical, good reasons to go to Big Tech. I'd get my green card faster. I'd get paid more on average. And the unfortunate reality, I think, is that the role does hold more weight. People are more excited about hiring an L5 Google engineer versus an L5 from a startup, especially if that startup doesn't become very successful,' she added. Kalra also pointed out that there were several reasons to work at a startup, such as the experience of programming software from scratch as well as tackling challenges on your own. 'From a software engineering space, maybe one quarter you're working on a growth hacking front-end feature, and the next quarter you're writing Terraform. But even in terms of the non-technical skills, you get an insight into how the business works,' she said. 'Startups also afford you more responsibility, along with a better chance of materially affecting the company with your work. You just get more agency in what you work on. You get the opportunity to propose ideas that you think would be impactful for the business and go execute on it,' she added.
Business Insider
01-06-2025
- Business
- Business Insider
Should you join a startup or Big Tech out of college? An OpenAI engineer weighs in.
Janvi Kalra, an engineer at OpenAI, thinks students should diversify their experiences after college, with at least one internship at a Big Tech firm and another at a startup. That way, she said on an episode of The Pragmatic Engineer podcast, you have a better idea of what career path you should take. Kalra interned with Microsoft and Google. She then worked for productivity startup Coda before transitioning into her current role at OpenAI. She said both tracks have advantages and disadvantages. "The way I saw it, the upside of going to Big Tech was, first, you learn how to build reliable software for scale," Kalra said. "It's very different to build something that works, versus build something that works when it's swarmed with millions of requests from around the world and Redis happens to be down at the same time. Very different skills." Another good thing about Big Tech, she added, was the amount of time she got to work on projects that were under less pressure to immediately succeed. "Different upside for Big Tech in general was that you do get to work on more moonshot projects that aren't making money today," Kalra said. "They don't have the same existential crisis that startups do." And then, of course, more practically, were the financial upsides — including potential prestige. "There are also practical, good reasons to go to Big Tech," Kalra added. "I'd get my green card faster. I'd get paid more on average. And the unfortunate reality, I think, is that the role does hold more weight. People are more excited about hiring an L5 Google engineer versus an L5 from a startup, especially if that startup doesn't become very successful." Still, Kalra said, there are "great reasons" to go to a startup, like the sheer amount of experience you'll get with programming itself. "First, you just ship so much code, right?" she said. "There are more problems than people, and so you get access to these zero-to-one greenfield problems that you wouldn't necessarily get at Big Tech maybe where there are more people than problems." She said another advantage is the wide array of challenges that'll be thrown at you, allowing you to develop expertise on several fronts. "Second is the breadth of skills — and this is not just in the software engineering space," she said. "Right from a software engineering space, maybe one quarter you're working on a growth hacking front-end feature, and the next quarter you're writing Terraform. But even in terms of the non-technical skills, you get an insight into how the business works." Startups also afford you more responsibility, along with a better chance of materially affecting the company with your work, she said. "You just get more agency in what you work on," she said. "You get the opportunity to propose ideas that you think would be impactful for the business and go execute on it." Given the opportunity, Kalra said it's best to gain experience with both startups and larger firms as early in your career as possible. "Given that Big Tech and startups are such different experiences and you learn so much at each, it would be more educational to do one startup internship and one Big Tech internship to get a very robust overview of what both experiences are like very early," she said.
Yahoo
22-04-2025
- Business
- Yahoo
A software engineer's success often depends on their relationship with their manager, says ex-Amazon technical director
An engineer's success is often largely determined by their manager, Dave Anderson said. The former Amazon engineering director said relationships with teammates and supervisors are crucial. On the "Pragmatic Engineer" podcast, he said underestimating a supervisor's influence is a mistake. A large part of an engineer's success is often tied to how they navigate their relationship with their manager, a former Amazon engineering director, Dave Anderson, said. "I would actually say, as a manager, even, like 50% of that performance, frequently, is your relationship with your manager and your team. How will you fit in with the team, with your peers, with your manager?" he said on a recent episode of "The Pragmatic Engineer," a podcast. Underestimating your supervisor's influence can be particularly dangerous at a company like Amazon, he added. As can be the case at other companies, a manager's decisions can shape the future of their direct report, from determining how much they're compensated to how far up the career ladder they climb. The influence a manager can have over an engineer's trajectory isn't something to shrug off, Anderson said. "I think the mistake that people will sometimes make is like, 'My manager doesn't influence my job that much because I can work independently,' or, you know, 'I don't need to figure this out with my manager because I can, you know, work with my peers, or I have this great engineer on my team I can work with,'" Anderson said. But if your manager doesn't like you, he added, you're "never ever" going to be able to snag a promotion. It's also important to remember that managers are often asked to point out a number of team members "who are not doing great," Anderson said. If you are, when compared with other members of your team, the "least effective," he added, you could be on the chopping block. "If you look around the room and you're thinking, 'Yep, I'm the worst one here' — that's not a great situation to ever be in. It's just never safe," Anderson said. "And at Amazon, it's definitely not safe. Some other companies where they just might do layoffs once every four years, you might be safe for quite a while. But Amazon has this sort of regular cycle." In response to a request for comment from Business Insider, Margaret Callahan, an Amazon spokesperson, said Anderson's experiences were his alone. "These claims reflect the opinion of one individual who worked at Amazon years ago. They're not based in fact, and aren't indicative of what it was like to work here then or what it's like today," Callahan said. "We're proud to be one of the most sought-after employers in the world and to have ranked in the top three in LinkedIn's Top Companies for eight years running." Anderson said that if an engineer's relationship with their manager isn't good, there can be an escape hatch of sorts: moving teams before being managed out. "So many times I've had someone who was either doing amazing on one team, they moved to the next team, and they're, like, actually not doing well at all, or someone who was not doing well escapes to another team before they get fired — and they do well," he said. If you start to hear rumblings from further up the chain of command, Anderson thinks it could be in your best interest to make a change, and swiftly. "This is, like, my sneaky recommendation for anyone is like — if you start to hear performance feedback whatsoever from your management chain, if you have any opportunity at all, get off your team fast as possible," he said. In a follow-up email, Anderson told BI that in a "great number of situations," he'd seen success prove itself to be at least partially dependent on team fit. "I've seen poor performers turn into great performers, and great performers turn into poor performers — and the only factor was them switching teams," he said. "In particular, switching teams to a place where they didn't know their manager. I don't think people fundamentally changed — so the only reasonable conclusion is that team fit (in particular, their relationship with their manager) is the deciding factor." Anderson doesn't suggest disavowing any negative feedback you receive and bailing out into a different section of the company on a whim. It depends, he said, on the relationship you've formed with your supervisor and whether you have faith in their advice. "Now, if you trust your manager, they might be actually just giving you honest feedback, which you'd like to be able to receive," he said. "But for the most part, if you've been working for someone for three years and suddenly they start giving you performance feedback, that's a really bad sign." Anderson added: "If you run for the hills fast enough, it's possible you'll get away before they flag you in the system as non-transferable." Read the original article on Business Insider
