Is the Engineer Dead?

Episode 43 Is the Engineer Dead

Will new emerging technologies assist engineers in enhancing their capabilities or will their jobs simply be done for them?

At Collision, 2023, Hessie Jones met with Maxim Fateev, CEO of Temporal and Anand Kulkarni, CEO of Crowdbotics. Both organizations are making it increasingly easier for engineers in light of the advances in Generative AI.

One school of thought: it’s much easier to replace certain aspects of desktop jobs than replace labor skills like carpenters or construction workers. Everything that works well in AI is happening behind the screen currently, but not so much in the physical world and these may be the the jobs most vulnerable.

How has automation impacted current engineering roles and practices? Have there been areas within this segment that have already been disrupted? How do both see the future role of the engineer?

Transcript

Hessie Jones

Hi everyone. My name is Hessie Jones, and we’re talking about an interesting topic today. If everybody knows or you haven’t been hiding under a rock because of generative AI, everybody’s now trying to figure out if there’s going to be job displacement. So the title of this talk is is the engineer dead? And I’m happy to have with me today Anand Kulkarni, who is the CEO of Crowdbotics, as well as Maxim Fateev, who is a co founder and CEO of Temporal AI. So I’m going to start off with asking each of these gentlemen a little bit about their company and how they’re actually making it easier for engineers. So I’m going to start with Maxim.

 

Maxim Fateev

I think it’s mistake actually Temporal IO, not AI. I know AI would be much better I know these days, but in five years can be something else. Temporal is an open source project targeting developers. And if you’re a developer, I absolutely recommend checking it out. We started it seven years ago at Uber and we practically tried to improve developer productivity. And then within three years, we grew up to 100 applications using within Uber. But as it was open source, we started to pick up adoption from other companies like Airbnb, Box, Coinbase, and they started to run mission critical applications on that.

So we started a company. So if you’re an engineer, I can give you like ten second description. What it is. We make your code run in presence of any failure means that your process will be recovered fully, including variables, blocking calls, stack, and it will keep running if anything happens. So you can write function which will run for a year because we will take care of recovering that state. So just check it out or talk to me after that if you’re interested. Okay, thank you.

 

Hessie Jones

Anand?

 

Anand Kulkarni

Yeah. So I’m Anand Kulkarney, CEO of Crowdbotics. We are a six year old venture backed technology company officially headquartered San Francisco but located all over the world. We are a code operations company. What that means is that we help companies of all sizes as big as the US. Government, down to small startups, create applications faster using reusable modular building blocks of code and now increasingly artificial intelligence to generate specifications and components of code soup to nut software creation. We’ve seen over 25,000 applications launched on the platform, including a number of high growth startups as well as large enterprise organizations moving critical application products into Crowdbotics.

We were one of the earliest users of OpenAI as a software generation tool. You can come check out the product in action. We have a booth over at 375 or come talk to us after. Okay, thank you so much.

Hessie Jones

The keywords I got from you guys, I got build apps faster, no coding required, open source simplify coding. It seems like you are making it easier for engineers in light of what’s going on today. So what is your view on the engineers role ten years ago. And how does that compare today? I’ll start off with Maxim.

 

Maxim Fateev

I think that life of engineers actually got harder. If you think about it last ten years, and actually more than ten years, we spent most of our innovation obviously I’m not talking about AI, it’s a separate story. But we spent a lot of time trying to improve deployment and operations.

So now we can deploy much, many more processes much faster, and you can run hundreds of thousands of processes faster. But the core business abstractions didn’t change. We actually became building distributed services more. So every developer right now is distributed system engineer, at least like back end developer. And it means that life of developers got harder because you don’t have transactions anymore. You need to call all these APIs. Consistency is a problem.

Failures is a problem. So my point is that life of developers is much, much harder now.

 

Anand Kulkarni

Okay, Anand so I think the fun thing about being an engineer is that the more things change, the more you have to keep abreast of these developments and remember what is actually hard to do and what is becoming easier. So what we’re seeing now is a sea change in how people, as engineers are forced to think about where their role is in the software development process. Whereas before you might have just needed to be an architect, a creator, looking at individual lines of code, individual functions, now, because tools are able to assist in letting us create more and more rapidly, we have to become editors as software engineers. We have to become system level thinkers, as opposed to individual sort of hands on keyboard becoming the most important. And that’s a surprising shift.

 

Hessie Jones

Okay, so you guys think that there’s still relevancy when it comes to the engineer’s role. So I’m going to throw a little bit of a wrinkle in this. I spoke with Juergen Schmidhuber, I’m sure you both know him. He’s a renowned researcher. He did some early renditions of Gans as well as transformers, and he’s actually in his lifetime trying to develop AGI before he retires. So when he talks about the replacement of jobs, this is what he says. He says it’s much easier to replace certain aspects of desktop jobs than replacing a carpenter, for example, because everything that works well in AI is happening behind a screen currently, but not so much in the physical world.

So what are your thoughts as it pertains to the future of engineering?

 

Maxim Fateev

I do think that if you work with a computer, things change much faster. It’s true, physical world doesn’t change as fast, but I don’t think we will lose some jobs and get others. But I think as it belongs to engineering, I don’t believe engineering will change. It can change, but it’s not going away. It’s just what you’re doing might change. And also it heavily depends on the specific area and specific domain. Closer you are to areas which can be automated.

Like if you’re doing think about where AI is useful, when you can see output right away, you ask AI to generate UI, you can see it. You ask AI to generate something which has visual representation, you can see it if you ask AI to generate driver and you still need to see the code because it can dream up some driver for something which doesn’t exist. So you still need to be pretty much involved there. Okay, thank you. It’s funny to hear a claim like this from someone at Uber because Uber has been at the forefront of taking physical jobs and figuring out ways to remove the driver from the equation. At least in San Francisco, we’ve seen that was the first role to try and be eliminated from the Uber experience. For the developer, though, I think you’re going to see a transformation of the role, not an elimination.

 

Anand Kulkarni

So the engineer is not dead, they’re just doing a different job than they were ten years ago. And if you look at similar domains like banking, where you had a rise of automation, the number of teller jobs and banking jobs went up after ATMs came out, not down, because there was more opportunity that was created for everybody in the landscape. And that’s likely what we’ll see in software as well.

 

Hessie Jones

Okay, so let’s talk about generative AI, because we’ve seen a lot of mass adoption currently in a short amount of time, but we’re also seeing a lot of kinks in the system. Everybody talks about hallucinations, they talk about IP issues, they talk about manipulation and people using it for bad things. So as companies start to develop this stuff, do you think the engineer’s role is going to be partly remediation in anticipation of some of the issues that we’re seeing in generated AI?

 

Maxim Fateev

Absolutely. But I think it is like with self driving cars, like last 5% of making it right will take like 95% of effort.

And I think what we will see is that domains when you don’t need to be precise, for example, image generation, we clearly see that movie generation, all these things. I’m pretty sure we will see more and more of that and will become much, much better very fast. And it can certainly change completely change the way we do graphic design, movies and all of that. But as far as you move to the areas when it should be more precise, you certainly should be much more careful, especially if you want to give it to your customers. And if you are like airline or bank and you want to just replace your customer service with AI, I don’t see that current state actually can work without at the same time, AI as helper. When human can judge it, copilot is the best example. I think we will see more and more of that, but human will always be in the loop.

So I think this is the best application when human is in the loop and human is qualified to judge the quality of that.

 

Hessie Jones

Okay, thank you. So I’m going to throw this to you, Anand. We’ve already seen in big tech, especially in the last, I guess, couple of years, that there is this tension between what’s right for the end user mitigating those harms versus the profitability of the shareholder. What are the potential ethical considerations that are arising within the engineering role and within the community as new technologies evolve? Are frameworks starting to come up that allow them to actually think about how their technology is affecting people?

 

Anand Kulkarni

So I think this is a great question. Every CIO’s office is now trying to come up very quickly with an AI strategy for their organization. And every engineer inside that organization is also being forced to help answer this question of how can our organizations use AI safely and ethically? And it’s interesting to see that the frameworks that are coming up are really being driven by the platform providers, people like OpenAI or related groups that are trying to advance the narrative. But those questions are still being left more or less open for the end users to help address.

So engineers have an increasing ability to try and answer these questions in their own products as they build them for the first time, which is very interesting to see. When you’re writing prompts, you can decide, or when you’re fine tuning a data set, you can decide, am I going to work with one that mitigates bias or mitigates risk or accelerates my ability to build software? And that’s an interesting trend because those decisions are often not made at the engineering level in other sea changes. I just want to add to that. There needs to be more, I would say CEO C suite buy in to being able to say it’s okay if there’s bias, and yes, that makes us more profitable. But we need to now think of the long term implications.

I think you have to answer this question at the top, early in the process of understanding your AI strategy. It’s too late. If you’ve already got a system in the wild that’s generating profit, then there’s going to be difficulty in pushing back and saying, okay, let’s rein things in. And that can be honestly damaging to the company. At the end of the day, once they’ve got something that’s out there and making headlines in bad ways, it’s much safer for the company to think about this early from the sea level down and say, let’s take an intelligent approach to mitigating the risks of the system that we’re trying to create while taking advantage of the acceleration we can get from software creation or other kinds of AI.

 

Hessie Jones

Okay, perfect. Thank you. Okay, so Maxim like long term, what are the skills and competencies that make it increasingly important for engineers to have to actually remain relevant? And we have this conversation backstage because everybody’s going to say, well, you know what? You don’t really need an engineer because now there’s low code, there’s no code. You got wix to do a drag and drop to build your website.

Do you have any opinion on whether or not that stuff is actually making your profession more vulnerable?

 

Maxim Fateev

Like talking like WordPress didn’t exist, right? WordPress exists for a long time. There are millions of websites built without coding. So it doesn’t change much there. Again, it applies to systems. When you can see your result right away. If you ask AI, make me a query and you see result, you don’t know what it’s doing there, right. You cannot really trust it.

When we talk about code, imagine you go to AI and say generate me operating system. What it means, like any function in this operating system kind of implementation has millions of decisions there. Even if AI makes either AI makes those decisions for you, or you need to ask you and you need to make those decisions. And I think a lot of those will be around deciding what is actual. Kind of like I don’t believe this thing will just dream up things, right? Like you need to give it feedback. How this result of the AI generation? It still needs to be something precise, right? You cannot generate something like, I don’t know, a bunch of numbers looking at human looks at that. I don’t know if it’s true or not.

Right. It should be something which is practical, right? And maybe language will change, but it still will be some language which AI will generate. Human will look at that. Say yes, it makes sense to me. Especially for things we should be precise. You cannot drive airplane by just asking a bunch of numbers. Do something, right.

You need to know that it will do the right thing. So for me, I think only for areas when you can apply this kind of low code no code solution. Like AI is just another low code no code solution to me. You cannot really and again, autopilot when it helps you to build like every little piece of that bigger puzzle you are assembling as a human. Obviously, this piece will become larger and larger, so AI will be better helper, but I don’t think it ever will go and fully replace what any engineer is doing. Okay. Yeah.

 

Anand Kulkarni

So our take on this is a little different. I think AI is a fundamentally different innovation than low code no code. So the promise of low code no code was great. We had this idea that you could have people who are not technical build powerful software systems and apps. And all engineers know this is sort of a myth. It’s proven that we can build out simple things using low code. But no great products have been built on low code.

No code systems ever. And certainly no high growth startups can be created on a low code, no code platform. What you can see though is that with AI we can revisit this idea of democratizing access to software creation because systems like GPT Four or Codecs can actually generate code. I think they are fundamentally lowering the barrier to entry for people to create software and software that’s actually meeting the standards of real software engineers. Not just simple drag and drop interfaces, but actually writing commits inside git launching production systems. I think this is still early days and we’re seeing already impressive results. So I’m optimistic on where this could go.

 

Maxim Fateev

But I think it’s actually making full code more accessible as opposed to saying this is a low code alternative. Okay. I think the difference is the target, right? If your target is professional developer versus citizen developer, right. Obviously AI will open more application of citizen developers. But I don’t see replacing professional developers. I think that is kind of my main idea there is. Who’s your target? I mean, one of the biggest apps that have been picked up is actually and this is probably one of the least headline making but most productive application is GitHub copilot.

Right? And look, that’s the developer being made more powerful with AI. So I think this is the kind of thing we will expect to see much more of as the systems evolve and get better.

 

Hessie Jones

So it seems like both of you don’t see any kind of real inherent risks, at least in maybe the next decade. But engineers, I guess in general, over time will have to embrace this continuous mindset of innovation, how they themselves remain relevant. So what is your view in how they need to create this new mindset and be proactive for the onslaught of oncoming change in innovation?

 

Maxim Fateev

I think we always had these two types of engineers. Engineers who learn all the time and people who are doing UIs like web frameworks, they change all the time. It’s normal.

Certainly certain class of applications will probably will be solved by AI much more. And again, WordPress is out there, we are still writing a bunch of websites from scratch using the existing technology. So it will be kind of similar thing.

But at the same time, I think if you’re still operating on the back end, building this complex system, which AI will be huge helper, Copilot will be absolutely there and I’m pretty sure Kapil will be much more useful. But that engineer will be assembling those solutions maybe on a little bit higher level of abstraction.

 

Anand Kulkarni

Well, so I think there’s actually a little more inherent risk than we realize. I have a friend at a fang company. He said when he saw GPT4 writing code for the first time, it was the first time in his life as a software engineer. He said, I need to think about my future and get ahead of this. And I think what this means for the engineer is it’s adapt.

Right. This is a new way to think about building software, and it’s a new technology to understand and master. And in that sense, it’s very similar to the last big changes we saw with the emergence of web and the of mobile. But this one is important to stay on top of, because to be relevant, you’ll need to know how these tools work and really master them.

 

Hessie Jones

That’s great. Well, that’s it for this panel. Thank you very much for listening to us. And thank you, Anand as well as Maxim for joining me.

Host Information
Hessie Jones

Hessie Jones is an Author, Strategist, Investor and Data Privacy Practitioner, advocating for human-centred AI, education and the ethical distribution of AI in this era of transformation. 

She currently serves as the Innovations Manager at Altitude Accelerator. She provides the necessary support for Altitude Accelerator’s programs including Incubator and Investor Readiness. She will be the liaison among key stakeholders to provide operational support and ultimately drive founder success. 

LinkedIn

You can also listen to this podcast on Transistor.

Please subscribe to our weekly LinkedIn Live newsletters.