Are you learning how to code? A standard engineering degree may not be enough.
In recent posts, harvard business reviewMarco Argenti, chief information officer at Goldman Sachs, said AI can now write high-quality code, sometimes better than humans, and the technology is only evolving.
So how can engineers catch up? He shared with readers the advice he gave his college-age daughter: “If you want to pursue a career in engineering, you should focus on learning philosophy in addition to traditional engineering classes.”
Argenti is an engineer with a background as a technology entrepreneur who held leadership positions at large companies such as Amazon Web Services and Nokia.
On the left is Goldman Sachs CIO Marco Argenti, and on the right is Warner Music Group CEO Robert Kinkle. Credit: Patrick McMullan/PMC, Getty Images
He said that as AI becomes more complex, run-of-the-mill engineering courses are no longer sufficient, as even if the AI’s code is technically correct, it may not do what it is supposed to do.
If the engineer cannot understand the prompt correctly, the AI will write code that may be somewhat correct in the best case scenario, but completely wrong in the worst case scenario. However, a philosophically based engineer must have the ability to keep up with reasoning and first-principles thinking.
For engineers, the question isn’t “Can you code?” But “can we get the best code out of AI if we ask the right questions?”
Related: What is Prompt Engineering, a hot new tech job in AI?
AI prompt engineering, or communicating with AI in the most appropriate language to get the desired output, is an emerging field. Jaime Teevan, Microsoft’s chief scientist, wrote in December that communicating with AI is different from telling team members what to do. AI requires more context and may need paraphrases and examples before producing actionable output.
Robin Stein, a partner at PricewaterhouseCoopers, told the publication CIO that PwC “has trained 75,000 people to learn the fundamentals of generative AI, how to apply it responsibly, and how to become agile engineers. “We quickly realized that we needed to upskill our employees.”
Related: Two Yale University PhDs aim to reduce AI’s illusions by a factor of 10, allowing small businesses to adopt the technology in weeks instead of years.
Argenti emphasized that in the age of AI illusions, it is increasingly important to ask the right questions and adopt a philosophical mindset.
“Have a clear mental model of the problem, be able to break it down into manageable steps, have perfect first-principles thinking, and be ready (and able) to argue with sometimes stubborn AI — These are the skills that make great people the future engineers, and the same considerations probably apply to many professions,” Argenti wrote.
