Imagine that you studied for five years, struggled through exams on data structures and algorithms, landed a dream job at a tech company with a salary that made your friends a little bit jealous. And then, just when life starts to feel really good, the man who created one of the most used AI tools for coding tells you that your job title might not exist by the end of the year. It is 2026, and welcome to reality.
Table of content
Boris Cherny, the creator of Claude Code at Anthropic, recently said in an episode of Lenny’s Podcast that the job title “software engineer” will soon disappear. It will be replaced by something more democratic and somewhat more intimidating: “builder.” Everyone codes, everyone is a product manager, and the distinction between those who understand technology and those who don’t is disappearing rapidly.
What exactly is Claude Code?
Claude Code is not an ordinary AI chat tool where you type a question and get a mediocre answer back. It is an agentic tool, which means it can independently perform tasks with minimal human intervention. It writes code, tests it, fixes bugs, and delivers finished solutions. A senior Google engineer stated that the tool recreated a full year’s work in just one hour.
Cherny said that he himself hasn’t written a single line of code by hand since November. Claude Code has done everything. He still reviews the code to ensure it is correct and safe, but the heavy, repetitive work is handled by the AI. It is one thing for an AI enthusiast on Twitter to claim that AI will replace all jobs. It is quite another when the creator of the tool says it themselves.
Analogy Book printer and writer fate
Cherny used an interesting historical comparison to explain what is happening. Before the invention of the printing press, it was the writers’ job to copy books by hand. They were a small, privileged group who mastered something most did not. When the printing press arrived and literacy spread, the role of writers changed radically. They stopped copying books and instead began focusing on what they truly enjoyed: bookbinding, illustration, and artistic expression.
The analogy is worth considering. Just as the role of the writer changed, the role of the software engineer will change. The mechanical, repetitive coding will be taken care of by AI. What remains is the creative, strategic, and human aspects.
It is a comforting thought, if you choose to see it that way. A little less comforting is the fact that the writing profession as a career more or less ceased to exist. But perhaps that is a detail you don’t need to mention at the Friday coffee break.
Many will unfortunately be affected
Cherny is not naive about what this means. He said outright that the change “will be painful for many people.” It is a rare honesty from someone who is at the epicenter of the industry and whose product is one of the driving forces behind the transformation.
Anthropic states that it takes the issue of job transformation very seriously and hires economists, policy experts, and social analysts to understand the consequences. At the same time, the company shows no sign of slowing down the pace of technological development, especially with a planned stock market listing in 2026. It is a classic paradox of the tech industry: it is aware of the social consequences, it has empathy for those affected, but it pushes forward anyway.
With old jobs new interesting ones emerge
The history still favors the optimists. Every time technology has taken jobs, it has ultimately also created new ones. Industrialization took away weavers’ jobs and created factory work. The computer took away typists’ jobs and created a whole ecosystem of IT professions. AI appears to be repeating this pattern, albeit at a pace that makes history dizzying.
According to World Economic Forum forecasts, the labor market between 2025 and 2030 is expected to create approximately 170 million new jobs globally, while about 92 million disappear. The net is positive, but the transition in between is what Cherny rightly calls painful.
New professions already emerging
AI Training Specialist Responsible for shaping, testing, and refining how AI models learn and behave. It involves understanding where the model fails, why it does so, and how to correct it. Three years ago, the title barely existed. Today it is one of the fastest-growing professions within tech.
Prompt Engineer Sounds strange, but it is serious. A Prompt Engineer knows exactly how to communicate with AI systems to get the best possible results. It is a combination of linguistics, logic, and psychology, but directed at a machine. Many laughed at this profession when it first appeared. Most of them are now looking for jobs.
Agentic AI Engineer The hottest profession right now. An Agentic AI Engineer builds and maintains AI agents that independently perform tasks, make decisions, and interact with other systems without a human needing to monitor every step. It requires a deep understanding of how autonomous systems behave, where they can go wrong, and how to set the correct boundaries for what they are allowed to do. With tools like Claude Code accelerating in adoption, demand for this skill is currently almost impossible to satisfy.
Agentic Systems Architect Where the Agentic AI Engineer builds individual agents, the Agentic Systems Architect designs the entire ecosystem of agents that collaborate. Imagine a network of AI agents where one handles email, another updates databases, and a third communicates with external systems, all coordinated and secure. Someone needs to map out how it should work. This is that profession.
AI Agent Quality Assurance Specialist When an AI agent acts autonomously and makes hundreds of decisions per day, a small error can have large consequences. The AI Agent QA specialist tests, evaluates, and certifies that agents behave as they should, even in edge cases and unexpected situations. This role combines classic testing with a deep understanding of how AI systems reason and where they tend to lose their footing.
AI Orchestration Engineer Responsible for coordinating and optimizing how multiple AI agents and systems work together in real time. It involves ensuring that the right agent does the right thing at the right time, identifying bottlenecks, and ensuring the entire automated flow operates efficiently. As companies build increasingly complex agentic pipelines, the need for this role is growing explosively.
AI Ethics Analyst Appearing on more and more company organizational charts. The task is to ensure that deployed AI systems do not discriminate, deceive, or accidentally do something spectacularly stupid that ends up on the headlines. A combination of law, philosophy, and technical understanding that the job market desperately needs more of.
AI Product Manager The role that understands both business needs and what AI can actually deliver, and most importantly what it cannot deliver. It is one of the hottest roles right now because the gap between what executive teams believe AI can do and what it actually does is huge. Someone has to stand in that gap and explain both sides.
AI-focused Data Analyst AI generates enormous amounts of data and insights, but they are useless if no one can interpret, contextualize, and communicate what they actually mean. The analyst who combines statistical skills with the ability to understand how AI systems reason is in high demand and is expected to continue being so.
AI Security Specialist Cybersecurity has always been important, but when AI agents start making autonomous decisions and interacting with sensitive systems, entirely new vulnerabilities emerge. The AI Security Specialist is a hybrid between a traditional security expert and AI engineer, focusing on preventing systems from being manipulated or abused. With autonomous agents having access to emails, files, and business systems, this profession only grows more important every month.
Human-AI Interaction Designer An evolution of classic UX design, but focused on how humans interact with AI systems in an intuitive, safe, and effective way. It is not just about interfaces but also about behavior, trust, and how to design systems that people actually rely on and understand.
AI Upskilling Coach With millions of people needing to learn new skills at record speed, the demand for qualified trainers and coaches has exploded. Those who combine teaching ability with deep understanding of AI tools have a profession with a secure future, which is somewhat ironic given the context.
Winners in the new work landscape
If you choose to be optimistic, there is practical advice to take with you. Cherny believes that those who will succeed best are not only those who can use AI tools but those who are genuinely curious and broad in their knowledge. A product manager who can code a little, an engineer who understands design and business strategy, a finance person who can communicate technical solutions. Generalists, in other words.
Claude Codes entire team codes, regardless of position. Product managers, business developers, everyone. Not because they are experts, but because the tools make it possible. It is a paradigm shift that changes what it means to be competent in a professional context.
Cherny recommends everyone, regardless of industry, to start experimenting with the tools now. Not out of fear but curiosity. Those who wait too long risk waking up one day to find the job market has moved on without them.
The shift is happening now
What is striking is how quickly the change is happening. Just a year ago, agentic AI was a niche term that most engineers barely understood. Today it is the standard method for how work is carried out on Cherny’s team. The same change is now spreading to semi-technical and non-technical professions because Claude can now interact with Google Docs, email, and Slack.
The big question hanging in the air is not technical but political and social. Who is responsible for the millions of people whose jobs are changing or disappearing? Cherny says it is a question for society to solve, not for the tech companies. It may seem like a disclaimer of responsibility, but it also contains a truth: technology is moving faster than politics has the capacity to respond to.
The software engineer is perhaps the first professional group to seriously feel the change. But hardly the last.
Sources
F.A.Q
Kommer mjukvaruingenjörer att försvinna helt?
Inte nödvändigtvis försvinna, men titeln och rollen kommer att förändras drastiskt. Boris Cherny, skaparen av Claude Code, menar att titeln “mjukvaruingenjör” ersätts av något bredare: “builder.” Det handlar mindre om att skriva kod för hand och mer om att förstå vad som ska byggas, varför det ska byggas och hur AI-verktyg bäst används för att bygga det.
Vad är Claude Code och vad gör det annorlunda mot vanliga AI-verktyg?
Claude Code är ett agentiskt AI-verktyg, vilket skiljer det från vanliga chatbaserade AI-assistenter. Det kan självständigt utföra hela arbetsflöden: skriva kod, testa den, hitta fel och leverera färdiga lösningar utan att en människa behöver övervaka varje steg. Boris Cherny har inte skrivit en enda rad kod för hand sedan november 2025. Allt sköts av Claude Code.
Hur snabbt går förändringen egentligen?
Snabbare än de flesta tror. För ett år sedan var agentisk AI ett nischbegrepp. Idag är det standardarbetssätt på ledande techteam. En senior Google-ingenjör uppgav att Claude Code återskapade ett helt års arbete på en timme. Det är inte en gradvis förändring, det är ett skifte.
Vilka nya jobb skapas av AI-utvecklingen?
Flera helt nya yrkesroller växer fram i snabb takt, däribland Agentic AI Engineer, AI Orchestration Engineer, Prompt Engineer, AI Ethics Analyst, AI Security Specialist och AI Upskilling Coach. Gemensamt för dessa roller är att de kräver en kombination av teknisk förståelse, affärssinne och förmågan att arbeta med och runt AI-system snarare än mot dem.
Vad är en agentisk AI och varför är det viktigt?
En agentisk AI är ett system som självständigt kan planera och utföra uppgifter, fatta beslut längs vägen och interagera med andra system, allt utan konstant mänsklig styrning. Det är skillnaden mellan ett verktyg som svarar på frågor och ett verktyg som faktiskt utför arbete. Det är den agentiska förmågan som gör att hela yrkesroller nu kan automatiseras, inte bara enskilda uppgifter.
Vad ska man göra för att framtidssäkra sin karriär?
Cherny ger tre tydliga råd: experimentera med AI-verktygen nu istället för att vänta, bli mer av en generalist som rör sig över disciplingränser, och odla nyfikenhet som en grundläggande arbetskompetens. De som klarar sig bäst kommer inte bara att kunna använda AI-verktyg utan att förstå de bredare problem de löser och kunna kommunicera det till andra.
Gäller det här bara mjukvaruutvecklare?
Nej, långt ifrån. Förändringen startade inom kodning men sprider sig nu till alla jobb som utförs framför en dator. Claude kan redan idag interagera med e-post, Google Docs och Slack, vilket innebär att administrativa, analytiska och kommunikativa arbetsroller också påverkas i allt högre grad.
