AI has been the buzzword for the past 2 years. Everyone is talking about it, many are embracing it, and a small percentage are even worried about its implications.
Since OpenAI dropped the ChatGPT bomb in late 2022, I've been learning a lot about AI. I've read articles, listened to podcasts, attended webinars, and had interesting chats with AI stakeholders.
Most importantly, I've been using AI regularly and observing how others use it.
Today, I want to share three perspectives on AI as a founder in the software engineering space.
Some of them are strong opinions, others are merely open-ended reflections.
If you have FOMO on AI, you're right. Missing out on AI is like missing the web in the mid-1990s.
ChatGPT has become a household name despite its somewhat unfriendly product name. Even my 75-year-old mom recently asked me to teach her how to use it.
While I mostly use LLMs as a thinking partner, software engineers are using them to 10x their productivity.
However, some people remain stuck in the mindset that LLMs cannot produce good code. They avoid AI-generated code entirely.
This reluctance might stem from a purist view of coding or discomfort with the disruption such technology brings to their daily workflows. After all, coding is central to a software engineer's workday.
Two analogies come to mind:
In hindsight, we all agree that using calculators and Excel save us time that would have been otherwise spent on low-value tasks.
In a few years, we will all agree that using AI to assist software engineers in coding was the right move.
The question is no longer whether to use AI but how to use it effectively.
Vibe coding - letting AI code on your behalf - is a hot topic causing debates everywhere, including within my own team.
From a founder's perspective, this approach seems promising. But what does it mean for engineers?
Research suggests that productivity gain from GenAI coding could be worth $12 billion annually in the United States alone.
Sundar Pichai has recently stated that AI systems are now responsible for generating over 25% of new code for Google's products.
Coding is just one (although major) part of a software engineer's role. The ultimate goal remains problem-solving.
I believe AI will soon replace developers - not software engineers.
While it will be able to generate code, software engineers will still be needed to review that code, debug issues, define product architecture, and decide on functional priorities for products.
AI is your coding companion. Guide it to ship faster.
In conversations with engineering leaders at startups and Big Tech companies alike, one recurring question emerges: will AI replace junior engineers?
A friend at Google told me that their team hasn't hired a single junior engineer in the past 18 months, which raises some valid concerns about the future of the role.
If AI is currently doing the job of junior engineers, what should they focus on instead?
This is where I start to realize the impact of AI on our lives.
Education has to drastically change: we should be teaching computer science students how to build scalable software, not the syntax of Java and Node.
Junior engineers entering teams will need higher-level thinking skills.
They will be assisting with building products that people want and enjoy using. They will be solving problems that have a direct business outcome.
This is where the big question lies: how much time do we need for this shift? And do we have enough time?
Book a call and get matched with engineers in 24–72h.
