Earlier this month, we had Tigran Sloyan, who’s the co-founder and CEO of CodeSignal, back on the ‘Tech Connects’ podcast. We discussed everything from internal skills training to the potential for AI to radically change how tech professionals do their jobs. Here’s the video in case you missed it:
If you don’t have time to watch the video (although you should; we really dig into some fascinating stuff), here are the quick takeaways from the discussion:
- Misunderstanding of AI is creating a gap in adoption: There's a lot of confusion about AI, leading to both fear and excitement. This uneven understanding is causing a gap between companies that are actively using AI and those that are cautious.
- AI won't replace the need for human skills, but will change them: Just like calculators didn't eliminate the need for math, AI won't make human skills obsolete. However, the definition of "being technical" will evolve. The ability to understand and collaborate with AI will become increasingly valuable.
- The education system is failing to keep pace with changing skill demands: Universities are slow to adapt their curriculum to the rapidly evolving job market. This leaves graduates with skills that may not be relevant and creates a burden of student debt for outdated education.
- There's an urgent need for reskilling and upskilling initiatives: As AI automates tasks, there will be a growing need for programs to help workers develop new skills and transition to different careers. We can't assume people will learn everything they need by age 22 and rely on that knowledge for life.
And here are some excerpts (shortened and polished for clarity) of the discussion’s key points. Sloyan makes some excellent points about companies’ speed of AI adoption, as well as the limits of non-technical people being able to code complex apps and services.
Q: Would you say that most companies have woken up to the need to train and boost employees’ skills internally, especially when it comes to AI?
“’Future is here, but it's just not evenly distributed.’ It's always the case that you've got some companies—and we have some customers—who are very aware of this; very aware that, for example, for software development, generative AI is like the introduction of a calculator, and the introduction of the calculator changed how we thought about math skills and what it meant to be good at math. And you're witnessing something very similar with generative AI and large language models (LLMs), but is everybody seeing it? Absolutely not.
“I think you have a big span of fear, excitement, greed, misunderstanding, and it's just a hot mess because it's all happening way too fast, and people are still trying to wrap their heads around it, and it also doesn't help that there's too much misunderstanding of what is AI, what are large language models, how does this all work? What does it all mean and where is it all going? I see part of our responsibility to be not just helping organizations go through this transition, but also educate and help set perspective and shared perspective.”
Q: Are you seeing that tech-focused companies are more on the ball with internal training and figuring out this whole AI thing, and non-tech companies are not?
“Before, when technology was the machines and factories that made the Industrial Revolution possible, being technical meant being able to operate the machine. And the machines became computers and being technical became being able to operate computers. I think you're going to keep seeing that, as technology keeps evolving, what being technical means keeps evolving, as well. So, with tech companies, you usually have more technical people, and naturally those who are more technical tend to adapt to new kind of technology faster.
“Then you also have the how regulated the spaces and how high the cost of failure is, because when you look at finance and healthcare companies, there's a lot more regulation. There's a lot more risk aversion and there's a lot more regulation that says things you can and cannot do, so this naturally leads to operating with more caution, sometimes fear, and it leads to that gap that we talked about where it's very unevenly distributed.”
Q: When you talk to some people, they suggest that the long-term promise of no- and low-code tools that allow everyone to become a citizen developer will finally achieve mainstream acceptance with the help of AI. In theory, you’re going to have companies full of people with relatively little tech background who will be able to do a few simple prompts and spin up an app for finance or healthcare, and that’ll change the game for how tech pros operate within their companies. Do you see something like that happening in the near term?
“There's this massive misunderstanding which I've heard from multiple people already that computer science and engineering is dead. That people shouldn't major in computer science, like that time has passed, and I'm like, that couldn't be farther from truth. Just because calculators came along, we didn't stop learning. In fact, even more people have to understand basic math so they could use calculators, right? Because you can't use a calculator if you don't know any math. And I think we're talking about the same thing here: before calculators can be in any way productive in math, you must go deep. You can't just have baseline knowledge nowadays; you can have baseline knowledge and do the type of computations that people took decades to master with crazy manual techniques, and I think we're going to see that same transition where being slightly technical is so much more powerful in this world than it used to be.
“You're right on the concept of, ‘Can we get more powerful no-code or low-code tools?’ We can for toy apps like most of the no-code/low-code has been so far, like if you want to just create a website for a restaurant or something like that. But when you're talking about massive healthcare or fintech applications… I mean, maybe LLMs will get better, the context windows will get better, but they're not going to be able to spit out a working piece of code that builds a massive few thousand of lines of codes that integrate. You must understand exactly what you're asking, and you must understand what you're getting back.”
