Some Thoughts on Academia vs Industry(with a focus on Computer Science)

I'm a 2nd year Computer Science PhD student, I also spent some time in the industry(part-time and internships, not a full-time-employment), I am also a somewhat reactionary person, I criticize a lot, I try to view problems of the processes that I reside in, I try to see what the originating reasons for such phenomenon is.

Today, there was a short Twitter discussion on senior academics leaving academia to join industry as Software Architects, and I bluntly answered with the obvious "money motive". It's not a secret that industry pays better, but that's not the only relevance of money. Industry also spends more money too. Industrial labs get more clusters, more engineers, more scientists, working on ambitious projects in periods of time many academic labs would not be working on. In short, they make more investment to their projects than small academic labs, at least most of the time.

But is that the only reason, can we really condense all of these different people leaving academia to just one reason? If we give the same amount of money to an academic lab, is that going to fix everything? I don't think so. Let's explore what other reasons people might have, what are the differences between academic and industrial labs, and how deep are these reasons lie in the foundations of the current academic system.

Before moving on, I must make it clear that this is simply an opinion piece, I'll try to use as many facts as I can that exist on the literature, I'll try to make my arguments as general as possible, but my opinions still rely on my experience and beliefs. I must warn the reader that the sources I quote will reflect my views, the arguments I build will be based on facts that reside on my side. I welcome all different views on the topic together with any type of criticism.

One thing I do a lot is a comparison between alternative realities. Let's compare two realities of the same person, making one choice differently. One of them is a successful academic that has published many papers, made considerable impact, recruited many bright students; the other one is a successful engineer working at a company they are happy with, has a hardworking team beneath them, they are able to effectively create value inside their company.

Note that even this is pushing the boundaries of my arguments into certain corners, but let's worry about that later on.

The academic has;

• Considerable pay.
• Tenure(they are basically guaranteed employment)
• Freedom:
  • to build a vision of research they believe in
  • to work on what they want(to the extent their grant agency provides them, which is usually a lot)
  • to recruit students they see fit to themselves(to the extent their budget allows them to)
  • to work with external collaborators they see fit
• Prestige(being a professor is rather cool)
• A lot more responsibilities that the engineer. Admin stuff, budget stuff, department stuff, teaching stuff, research stuff, community stuff...

You'll realize that this is one model of funding, which is very frequent in CS, so I'm going to stuck with that.

The engineer has;

• Usually greater pay than the academic
• A chance to work on a product that people actually use
• Large teams with experienced engineers working on a single product
• A rather easy process switching companies
• A lot more external pressure than the academic. If your manager says finish this project by this time or you're done, than you're done.

To me, these provide 5 potential avenues of comparison.

• Stability/Security
• Dynamism/Adaptability
• Freedom/Pressure
• Output
• Team
• Facilities

Let's go over each of these avenues

Stability/Security

It should be clear that I believe an academic job is much more stable, and somewhat more secure than the industrial one. Once the academic has tenure, they are basically guaranteed lifetime employment, and most universities tend to have very long lifetimes. Academic has a considerable pay, in most places enough to make them happy, and they usually make money off external consulting too.

Although, I must mention that this security/stability is hard to come by. A very small percentage of CS PhD's are able to get academic jobs, a much smaller percentage gets tenure. So even getting to this stage is a pretty insecure and unstable process. The pressure of working with many other bright young scholars forces one to work longer and longer hours, focusing on a simple objective, getting tenure. Whereas, the industrial version, making more money working 8 hours days might secure their future by carefully planning their future, investing their surplus wage, using the extra hours to put themselves in "hireable/valuable" positions.

When we talk about senior academics moving to industry though, I believe this stability/security aspect rather disappears from the discussion. Many accomplished academics will have their future already secured, so the displacement of employment guarantee should not really worry them.

Dynamism/Adaptability

I believe industrial careers are more dynamic, and industry engineers are more adaptable to new conditions. Even though the academics have much freedom over basically anything, they are bound by their expertise. Expertise in one academic area does not easily convey to another, as academic expertise tends to form over time and it usually builds upon a very sharp line. On the other hand, many industrial products use similar building blocks, meaning expertise in building one product (in my very immature and unfounded opinion) translates much better to building another product.

To me, this means that an industrial engineer working at Company X on product Y today can easily change their job to work in Company A on product B, whereas it would be much harder for the academic to change the area they work on. They need to cultivate new relationships with new external collaborators, recruit new students, learn te culture of a new academic community, adapt to the conventions and the systems of that community, funding agencies that provide funding to that specific area, and so on... I believe all of these force academics to stay committed to certain communities and domains, putting high thresholds and gatekeepers in the way of changing areas of expertise.

I think in the context of our discussion, this is an important avenue where industry leaves academia behind. If an academic wants to change their domain of expertise, maybe it's easier to move to industry? I'm not really sure about this part, but I think overall this could be a big part of the discussion.

Freedom/Pressure

I believe academic careers have much much much more freedom on almost any avenue. As you see from my passion in writing 3 much'es, I don't even think this is really debatable. I, as a PhD student, am able to pick my own projects as long as my advisor approves it, which he is more than happy to do. I see this is a similar case for many of my friends, especially as students get more senior, they draw their own path, differing from their advisors, and that is expected, it is asked for, even required.

Academia has the pressure of the race you're in. A Ph.D. student is racing for the cunning job market, a tenure-track professor is racing for the tenure spot. People who submit to the same journals and conferences race for the acceptance of their paper. The important thing is, this pressure is mostly self managed. You set your own limits, your own objectives, and the pressure is mostly shaped by your hands, especially as you get more senior.

In the industry, one gains freedom through seniority, but pressure resurfaces in many ways. There is the pressure from one's manager, the pressure from the market, the pressure from the race with others in your position for promotion, the pressure from investors. In short, pressure is an important part of the job. It is also shaped by your hand in some cases, but I believe not as much as it is in academia.

I think, in the context of our discussion, this is particularly important and, to me, surprising. I don't really see how academics leave their freedom and put themselves under the pressure that as seniors, go under the pressure that they were freed by their positions in their universities. This could be due to policies of companies(maybe academics are promised more freedom and less pressure), or some other factor that I cannot foresee right now. We should probably ask them about this.

Output

In my experience, the output incentives, and as a result the outputs themselves, are very different in industry and academia.

Academic work tends to have many smaller outputs, gradually migrating over time, divided into many papers. The output product is usually not consumer grade, it's designed with an academic mindset to be used by other academics. It tries to disrupt existing beliefs, concepts and limits, to create novelty one might say. The industrial work focuses much more on a designated consumer base and the problems of those consumers. It focuses on providing solutions to urgent consumer needs, novelty is an afterthough.

To exemplify what I mean, think about creating a new web interface for an existing database system. None of the components or ideas might be new, maybe the interface already exists for a different database, and you just create a very similar one for a different database. This would be out of the purview of academic work, it lacks novelty in all senses. Yet it might be a very good idea for an industrial product, as users of the database would like to use a better interface of couse.

In the context of our discussion, I think this is mostly about individual beliefs and desires. Every person has different motives, and maybe professors moving to academia is sometimes only about building products that solve urgent problems, rather than creating novelty. Maybe we are just putting too much thought into these decisions.

Team

Academic work is mostly done in small teams, it doesn't scale up too much. Industrial teams tend to be larger, focus on a single product for a longer period of time, and they usually have senior IC's(Individual Contributors). Academics have the advantage of self-recruitment, the ability to work with external collaborators. but they aren't usually able to build very large and effective teams with more than 2 senior members working on the same product for a long time.

In the context of our discussion, I think this might be on either side. Working with strong-willed students that work with you for a rather unconditional 5 year span is amazing, you build strong bonds, work towards a common goal, or sometimes even a dream, and once they are a faculty or engineer somewhere else, you get to continue working with them if they still do research in similar areas. Yet, industrial labs have the advantage that seniors don't have to spend their days doing admin or teaching stuff, there are people who do full-time research and coding with more than 10 years of experience. This is something almost no academic lab will ever have. Yet, these industrial labs are mostly contained to themselves, no external collaboration, even publishing papers is usually hard due to company regulations.

Facilities

Industrial labs tend to have greater clusters, machines, a larger workforce. These are facilities academic labs may lack. I'm not really sure how important this is for our discussion, but I don't think it really is much. There are industrial labs with small budgets and academic labs with huge budgets, and everybody somehow gets work done. So I won't go into much detail on this.

Summary

As a conclusion, I think the dynamic nature of industrial jobs, combined with the ability to work with senior individual contributors and the ability to create impactful products that solve people's urgent problems might be the key point that incentivizes senior academics to move to industry. But, I think we should just ask them, they'll probably give us more insights than we could ever find out by discussing amongs ourselves.