Kotlin "J God" Two-Hour Controversial Interview: I Would Rather Lose 80% of Job Opportunities Than Use AI for Programming Under Any Circumstances

At a time when the entire software industry is betting heavily on AI, publicly stating that you "firmly refuse to use AI for programming" almost voluntarily narrows down your career choices.

But Jake Wharton still did it.

This Android and Kotlin open-source guru, known as "God J" by domestic developers, listed nine job selection criteria when searching for a new position. The first item is: do not join AI companies, do not participate in AI-centric products, and do not accept mandatory AI usage for engineers imposed by employers. According to his estimation, this requirement directly excludes around 80% of potential employers.

Wharton is not short of options. He says the current base annual salary for senior Kotlin engineers in the US usually starts at 200,000 US dollars, and they may also get around 100,000 US dollars in equity per year, which vests over four years. Entering popular sectors like AI may bring higher returns, but also greater risks. When he was younger, he was willing to bet on the future with equity; now he values stable cash income more, as well as the freedom to choose employers.

This attitude also echoes the reason why he left large corporations. Wharton admits that Google's Android team is extremely excellent, but he dislikes the way Google locks users into its own ecosystem through products and services. As the company scales up, bureaucracy rises, product iteration slows down, and the management becomes more eager to acquire new users and revenue. In contrast, he prefers to stay in small teams with clear goals.

Wharton's attitude towards Silicon Valley trends is almost sharp. He believes that cryptocurrency "has basically become a Ponzi scheme" that creates little real value; while large language models are not only accompanied by disputes over copyright and open source licenses, but also entering software development in an abnormal way: in the past, new languages, new IDEs and new tools were usually promoted bottom-up by engineers, but now the management bypasses engineering judgment and requires everyone to use AI.

He has never used AI Agent to write code so far. In his opinion, AI can help engineers handle repetitive work, but the most irresponsible usage is to let AI generate code beyond the user's ability. If you cannot complete tasks without large language models, developers often cannot truly understand, verify and maintain the outputs. Once dependence is formed, engineers hand over not only part of their work, but also their own capabilities and bargaining power, and eventually they may become a gear that is easier to be replaced at low cost.

Wharton does not think AI will disappear, nor does he worry that programmers will lose their positions soon. "AI has become disturbingly powerful", but that does not mean it can undertake the entire software engineering process stably and flawlessly. For him, AI can at most be a tool in the toolbox, rather than the foundation of engineering capabilities.

From rejecting AI and questioning cryptocurrency, to discussing whether Kotlin still has advantages and why he prefers Rust, the following are the full answers of Jake Wharton in a recent two-hour interview.

1  Kotlin is like "what Java wants to be ten years from now" 

Host: Java was already the most mainstream programming language in the world back then, and Android was also built on top of Java. Why was Kotlin still needed?

Jake Wharton: When Kotlin emerged, the development of Java language was in a relatively stagnant stage. At that time, Oracle was gradually taking over Java, and the version release cycle was still as long as two years. For the Android team, the usable language features were largely limited by Java itself.

At the same time, programming languages outside the Java ecosystem were evolving rapidly, with many new features that developers wanted. The community also began to explore other alternatives, some used Scala, while others used Groovy. Everyone was looking forward to a more modern language, and Kotlin just appeared at this point in time.

JetBrains itself is a development tool company, so it can provide IDEs and supporting infrastructure while launching a new language. This makes Kotlin more acceptable to developers than languages that need to build the entire toolchain from scratch.

More importantly, adopting Kotlin does not mean abandoning existing Java code and libraries. Developers can continue to use the existing Java ecosystem, gradually introduce Kotlin into original projects, and then complete the migration slowly. This progressive compatibility capability is an important reason why Kotlin has become popular.

Host: You later became the first engineer on Google's Kotlin team. How did this opportunity come about?

Jake Wharton: Before joining Google, I had worked at Square for five or six years. About three years before that, we began to adopt Kotlin internally in the company.

To get company approval, I wrote a document explaining why we should use Kotlin. Since the arguments in it were not only applicable to Square, I later made this document public. It boosted the Android community's interest in Kotlin to a certain extent, and I got more and more deeply involved in Kotlin-related work, and kept promoting this language to developers.

As community attention rose, some people inside Google gradually realized that Kotlin could bring great help to Android development. Some Google employees I knew at that time began to hint that Google might officially support Kotlin, and asked if I would like to participate.

For me, there is no greater recognition than joining the company that is responsible for the entire Android toolchain and promoting Kotlin to become the official first-class language for Android. This opportunity was almost impossible to turn down.

Host: Did the Google Android team have any resistance to adopting Kotlin at that time?

Jake Wharton: It was more caution than resistance. At that time, Android had been built around Java for about ten to twelve years. Java was deeply embedded in the entire tech stack, and the team was very familiar with it. In contrast, Kotlin was still a relatively young, not fully verified language, which had only been stable for one or two years.

Therefore, there were indeed risks in adopting Kotlin, and there were concerns within the team. Google is not completely risk-averse, but it must make sure that all aspects are fully considered.

The introduction process of Kotlin was very prudent, and it was not completed with a hard switch overnight. The Android team initially built supporting libraries separately for Kotlin, instead of immediately rewriting existing Java libraries in Kotlin. The development team did not need to transform the entire codebase at once, but could adopt it gradually.

From another perspective, this also proves Kotlin's respect for Java's legacy. It allows developers to migrate slowly without overturning the existing system.

Host: Is Kotlin only a language for Android development?

Jake Wharton: In the first few years, because Kotlin gained massive support in the Android community, the outside world formed a wrong impression that it was a language specifically designed for Android. There were even articles claiming that Kotlin was created by Google for Android, which is obviously not correct.

Later, Kotlin began to enter more fields. Backend services that might have been built with Java in the past are now increasingly using Kotlin. Many capabilities provided by companies like Spring also prioritize Kotlin.

JetBrains is also continuously promoting Kotlin Multiplatform, so now Kotlin can also be seen in iOS and desktop applications. For example, the Toolbox App for managing and installing JetBrains IDEs is a desktop application built with Kotlin.

Android is more like a catalyst for Kotlin to enter other ecosystems. Take Square as an example, we were initially only allowed to use Kotlin in Android projects. But after backend Java engineers saw the mobile team's enthusiasm for Kotlin, they also wanted to use it. Later, Kotlin was gradually approved for backend services.

Host: If you were to recommend Kotlin to Java developers in one minute, what would you say?

Jake Wharton: Kotlin is very much like "what Java wants to be ten years from now".

Java must try its best to avoid breaking existing code, so the evolution speed of the language is relatively slow. Many modern language features that may enter Java in the future are already available in Kotlin now.

Adopting Kotlin also does not require projects to run on the latest version of JVM. In reality, many codebases still stay on Java 8 or Java 11, and Kotlin is also compatible with these environments. Developers can use more modern language capabilities on older JVMs.

Host: What large companies are using Kotlin in production environments now?

Jake Wharton: Meta, Google and Amazon are all using Kotlin. The initial applications were mainly concentrated on the mobile side, but now backend services are also increasingly adopting Kotlin.

These companies are also investing in related tools. For example, in the week we recorded this episode, Meta contributed its own Kotlin formatting tool to the Kotlin Foundation, which may become the official Kotlin formatting tool in the future.

These companies have the largest codebases in the world. The fact that they can gradually introduce Kotlin on top of existing languages and systems, and are willing to make long-term investments, is itself a verification of this language.

2  The annoying part of Google — locking users into its own ecosystem 

Host: Google is the dream employer in the minds of many engineers. Why did you leave after working for two years and nine months?

Jake Wharton: Before joining Google, I worked at Cash App. We promoted the adoption of Kotlin, and also wrote many open source libraries in Kotlin. Later, Google decided to make Kotlin the official language of Android, and this opportunity was too rare to come by.

However, when I joined, I told my colleagues that I would probably leave the office in two years. My plan was to help Kotlin become a first-class language for Android, push it to succeed as much as possible, so that I could safely rely on this language later, and then return to the front line to actually build products with these tools.

I love developing tools and libraries, but I don't want to only make tools forever. I also want to be a user of these tools, and truly enjoy the value brought by the early investment.

When the two-year period expired, I was originally prepared to leave, but Google happened to approve another project that I really wanted to work on, so I stayed for another nine months. Eventually, I thought I had achieved my original goal, so I decided to go back to product development.

Host: What parts of Google do you dislike?

Jake Wharton: Google is a very large company. The Android team is relatively independent, and the colleagues I met there are all very excellent. I still miss that period of work to this day.

But I have objections to some practices of Google as a company, for example, it locks users into its own ecosystem through products and services. However, these views are mainly aimed at the way Google operates as a company, not at the Android team.

Host: When you left Cash App in 2025, the technology industry was experiencing massive layoffs. How difficult was it to find a job at that time?

Jake Wharton：Finding a job was not easy, but there were not no opportunities in the market at all. I had some specific requirements for my next job, so I chose very carefully.

I was lucky that I left voluntarily according to my own plan, and there was no pressure to find a job immediately. But the current job market is indeed not like it was five years ago. Back then, an Android developer could easily be recruited by a company and get very high pay.

Now, job hunting requires more effort, takes longer, and you have to find ways to prove your differentiated value. I fully understand the pressure that engineers who are forced to enter the job market are facing. However, I don't think the entire market has completely collapsed, job opportunities still exist, but they are harder to find.

Host: What is the approximate compensation for senior Kotlin engineers in the US at present?

Jake Wharton：The base salary I usually see starts at 200,000 US dollars, and in addition it may include equity, for example, around 100,000 US dollars of equity per year, vesting over four years.

Of course, there are many positions below this level. Nowadays companies will consider the flow of funds more carefully, and no longer generally pay extremely high compensation as before.

If you want to get higher returns, you may need to enter the most high-profile technical fields with higher potential returns. Such companies may offer more cash and equity, but the risks are also greater: the equity may go to zero, or grow tenfold.

Ultimately, it depends on your personal life stage and risk tolerance. When I was younger, I was willing to take this risk; but now, I prefer to get stable and predictable cash income.

Host: By 2026, can mastering Kotlin still bring career advantages to software engineers?

Jake Wharton：I'm not sure it can still be considered an advantage. In a sense, this instead reflects the success of Kotlin: it has become a default requirement, rather than a scarce and hard-to-find capability.

This is mainly because Google established Kotlin as the official first-class language for Android. Now when you enter the field of Android development, you usually learn Kotlin directly, and employers also default that job seekers have mastered it. The same goes for Jetpack Compose: about four years ago, it might have been a relatively rare skill, but now it has gradually become a default requirement.

If developers master not only regular Kotlin, but also relatively advanced capabilities such as Kotlin Multiplatform, it can still constitute a differentiated advantage and make you stand out from other Kotlin developers. But just "knowing Kotlin" is hardly a clear career plus point, because it is so popular.

3  Why refuse to use AI for programming? 

Host: When you were looking for your next job, you publicly listed nine requirements for future employers, the first of which is "do not go to AI companies, do not make AI-based products, and do not accept companies forcing the use of AI for development". Why is this point so important to you?

Jake Wharton: It is placed first for a reason. This is probably the most important criterion for me when looking for a job.

We are in a period of rapid change. Not only is the way of software development changing, but many companies are also forcibly stuffing AI — more accurately, large language models — into places that might not have needed it originally.

I don't think the overall quality of software has improved in the past few years, nor do I think such tools will improve software quality in the long run. But some company executives see a different picture: they think AI can make engineers work faster and produce higher quality code. My judgment is exactly the opposite.

In addition to that, I also have strong objections to the ethical issues behind these tools. In my opinion, they are built on the massive use of copyrighted content. Model training ignores the copyright of artists, and also ignores the licenses of open source code. Being able to access a piece of content online does not mean you have the right to recombine, rewrite it, or even output it verbatim.

Therefore, I don't want to be associated with these things. It is already getting harder and harder to completely avoid AI now, so I put this requirement first. It also directly excludes about 80% of the companies that I might have joined.

Host: Apart from copyright and software quality, are you also against the way companies force engineers to use AI?

Jake Wharton：Yes. Traditionally, a development tool is usually promoted bottom-up by engineers.

It could be an engineer falling in love with a new language, like Kotlin; or a new IDE, like IntelliJ IDEA; or a new build system. Engineers find it valuable in their actual work, the enthusiasm gradually infects others, and then everyone convinces engineering managers to approve its adoption.

But the promotion direction of AI tools seems to be completely opposite. Why should the CEO of a company tell engineers what tools they should use in their daily work?

Company leaders usually don't directly force you to use statically typed languages, nor decide how you should use the type system to express business requirements, or how to handle null values. These are originally engineering decisions. Engineers understand business requirements, and then choose tools that can get the work done effectively.

If Jack Dorsey had required me to use Eclipse back then, I might have resigned as well. The problem is not just AI, but the company's management bypassing engineering judgment and forcing developers to use a certain tool.

Host: Why does the management's strong promotion of AI remind you of layoffs in the tech industry?

Jake Wharton