I believe AI is a threat for developers, but not for engineers. I might be wrong, but this is my rationale.
Developers develop. Engineers…engineer? How insightful! Let’s try that again.
Developers implement solutions and write code. Engineers solve problems and figure out whether a problem should be solved in the first place. I’m not saying either is better, but the distinction exists.
Engineering is 90% deliberate thinking and 10% execution. Development is almost the polar opposite. LLMs are not quite the “deliberate thinking” machines, but they can certainly ship some stuff out. This is the base of my premise.
I wrote about unusual skills of highly effective engineers before, let’s use that as a barometer. Engineers are creative, disagreeable, have a bias for action, communicate well and persevere through challenging problems refusing to settle for a makeshift solution. Most of the time anyway.
Out of the five qualities, currently available AI can accomplish one. One and a half at best. It would be the bias for action and communication. And the communication would get the incomplete score due to direct disobeying of what’s being asked of it and fairly frequent hallucinative lying.
I’m not so sure I’d reach out for advice to a colleague under the influence either.
The responses look okay, but every single word needs to be double-checked. Likewise, a bias for action exists, but without deliberate thought this action is mostly futile for any problem worthy of solving. It can type, do it fast, and do it for the problems it’s “encountered” millions of times before.
When presented with a novel problem, it’ll spew out half-baked garbage. And no, power-prompts are not the solution. It’s simply not the right tool for the job.
This is where I see an opportunity to thrive in what’s coming. Positioning yourself as a problem-solver, rather than a coder or developer.
From an engineering perspective, it’s about learning how to think wider and deeper, recognizing the overarching problems and learning how to prioritize them. Being able to think of five possible solutions to a problem and evaluate them against each other.
All of this is way more effective when we’re communicating well with other people. Everyone is aligned, has more context and makes the whole thing run smoothly.
Okay, smoothly-ish.
From a business perspective, the opportunity is in connecting with it deeply. Understanding the domain, the company’s priorities and making sure we’re moving in the right direction. It’s about finding clever and creative ways to bypass entire suits of problems.
From all perspectives, the opportunity is in having a voice that isn’t an average of a million voices, but an exponent of a single one.
Wow, this even sounds poetic.
So don’t panic about AI, understand your true purpose and position yourself to be a winner in the new era.
juricake•7mo ago
Developers develop. Engineers…engineer? How insightful! Let’s try that again.
Developers implement solutions and write code. Engineers solve problems and figure out whether a problem should be solved in the first place. I’m not saying either is better, but the distinction exists.
Engineering is 90% deliberate thinking and 10% execution. Development is almost the polar opposite. LLMs are not quite the “deliberate thinking” machines, but they can certainly ship some stuff out. This is the base of my premise.
I wrote about unusual skills of highly effective engineers before, let’s use that as a barometer. Engineers are creative, disagreeable, have a bias for action, communicate well and persevere through challenging problems refusing to settle for a makeshift solution. Most of the time anyway.
Out of the five qualities, currently available AI can accomplish one. One and a half at best. It would be the bias for action and communication. And the communication would get the incomplete score due to direct disobeying of what’s being asked of it and fairly frequent hallucinative lying.
I’m not so sure I’d reach out for advice to a colleague under the influence either.
The responses look okay, but every single word needs to be double-checked. Likewise, a bias for action exists, but without deliberate thought this action is mostly futile for any problem worthy of solving. It can type, do it fast, and do it for the problems it’s “encountered” millions of times before.
When presented with a novel problem, it’ll spew out half-baked garbage. And no, power-prompts are not the solution. It’s simply not the right tool for the job.
This is where I see an opportunity to thrive in what’s coming. Positioning yourself as a problem-solver, rather than a coder or developer.
From an engineering perspective, it’s about learning how to think wider and deeper, recognizing the overarching problems and learning how to prioritize them. Being able to think of five possible solutions to a problem and evaluate them against each other.
All of this is way more effective when we’re communicating well with other people. Everyone is aligned, has more context and makes the whole thing run smoothly.
Okay, smoothly-ish.
From a business perspective, the opportunity is in connecting with it deeply. Understanding the domain, the company’s priorities and making sure we’re moving in the right direction. It’s about finding clever and creative ways to bypass entire suits of problems.
From all perspectives, the opportunity is in having a voice that isn’t an average of a million voices, but an exponent of a single one.
Wow, this even sounds poetic.
So don’t panic about AI, understand your true purpose and position yourself to be a winner in the new era.