Software program engineers working in an open workplace area
getty
AI has modified coding endlessly. Simply final month, Sam Altman, cofounder and CEO of OpenAI, posted a thank-you be aware to builders on X marking an enormous shift within the trade.
“I’ve a lot gratitude to individuals who wrote extraordinarily advanced software program character-by-character. It already feels tough to recollect how a lot effort it actually took. Thanks for getting us up to now,” the post stated.
The implication of the submit was that the times of guide coding have been coming to an finish. A brand new era of AI-driven instruments like Claude Code, Cursor and Codex at the moment are able to producing code from scratch. Now, many up and coming builders are seeing much less demand for his or her abilities.
In response to a Federal Reserve Financial institution of New York study launched in 2025, recent pc science graduates are going through unemployment charges of 6.1% to 7.5%, greater than double that of biology and artwork historical past majors. As AI begins to automate code creation, many are beginning to ask, is software program engineering “cooked”?
The Software program Engineer Function At present
The way forward for software program engineering has been a frequent matter of debate amongst AI leaders. In February 2026, Boris Cherney, creator and head of Anthropic’s Claude Code, said that the software program engineer title goes to begin going away by the tip of the yr and “it’s going to be painful for lots of people.”
His feedback align with Dario Amodei, cofounder and CEO of Anthropic who predicted on the World Financial Discussion board that “we may be 6-12 months away from fashions doing all of what software program engineers do end-to-end.”
Whereas Cherney and Amodei’s feedback might be dismissed as an try and hype code creation instruments, there appears to be a rising variety of software program engineers struggling to search out work out there.
Garrett Rose, a full-stack engineer with three and a half years of expertise within the trade, just lately put out a YouTube video claiming that the stress of the job market had contributed to a subconjunctival hemorrhage. Rose was laid off from his earlier job 8 to 9 months in the past, a layoff which he attributes to AI. Rose is now one in all many engineers struggling to search out roles within the trade.
“I’m certainly very involved with the way forward for software program engineering. I perceive that expertise shifts and turns into extra summary from its core implementations. Nevertheless, what AI is doing is coaching off of each software program engineer that ever wrote code pushed to the net,” Rose informed me by way of e-mail.
Whereas Rose doesn’t imagine software program engineering has been made out of date by AI but, he believes it might be sooner or later. He additionally stated that AI’s capacity to finish easy work has led to firms solely hiring “unicorns,” skilled builders with the flashiest resumes.
“Proper now, software program engineers have transitioned from writing code manually on a keyboard a few years in the past to now talking to agentic AI, describing precisely what they need to construct and construct it. The engineering half has solely turn into the planning part, and roughly everyone seems to be an architect,” Rose stated.
Full Steam Forward On AI Coding
As builders throughout the market really feel the stress of automation, many firms proceed to go full steam forward with AI. Nowhere is that this extra clearly demonstrated than with the tokenmaxxing development, the place engineering groups are burning by as many tokens as doable to streamline product innovation.
Nevertheless, although many software program builders are being displaced by AI, a variety of trade leaders counsel that engineers nonetheless have a crucial function to play. Peter Hill, CTO at Synthesia, an AI video platform valued at $4 billion, informed me by way of e-mail that he didn’t imagine software program engineers had turn into out of date.
“No, and I believe individuals saying which are complicated the job with one particular a part of the job. For a very long time, the worth of a software program engineer was closely tied to their capacity to translate concepts into working code. That half is getting dramatically cheaper. However the purpose firms wanted that code within the first place hasn’t modified. Somebody nonetheless has to determine the fitting drawback, what an excellent answer appears like, and when one thing is able to put in entrance of a person,” Hill stated.
Hill famous that AI hadn’t simply modified the pace of writing code but additionally inverted the connection between planning and constructing. Synthesia can now prototype in hours what used to take weeks with instruments like Claude Code and Cursor. As an illustration, at a latest firm hackathon, one workforce of coders managed to rebuild the corporate’s video editor from scratch in simply a few days.
He stated that Synthesia has been utilizing instruments like Claude Code, Cursor and different AI coding instruments with promising outcomes, corresponding to quicker function work and simpler automation of routine duties. Hill additionally informed me that at a latest firm hackathon, the profitable workforce rebuilt the corporate’s video editor from scratch in a few days.
When requested if he would have a look at guide coding exams when hiring new builders, Hill stated, “I believe the trustworthy reply is that pure coding assessments have gotten much less helpful as a sign.” As a substitute, he’d reasonably see how somebody thinks by a posh product drawback and makes use of the instruments accessible to them to reply.
What Does Software program Engineering Look Like At present?
Synthesia isn’t alone in turning to AI to automate code. Nancy Wang, CTO of password administration large 1Password, informed me the corporate polled a few of its greatest engineers to ask how a lot code they have been writing manually and located that the highest 10% have been writing zero p.c of code manually.
“I really made the choice to drag guide coding exams from our software program engineering hiring loop,” Wang informed me in a video interview. “I need to see the way you’re enthusiastic about issues, much less so are you able to simply generate a ton of code? As a result of guess what? LLMs are going to be far more efficient at simply producing pure traces of code than a human engineer will.”
In a world the place writing code manually is turning into much less vital, Wang says builders want a mixture of the systems-level considering that may include a standard pc science curriculum, plus AI proficiency. “We’re transferring into totally agentic workflows, and so as to construct the fitting workflows, it is advisable know the place the methods are bottlenecked and the place totally different workflows may fail,” Wang stated.
“At 1Password, we’re requiring, you realize, AI proficiency and AI fluency for all incoming engineers, as in, we is not going to rent anybody who shouldn’t be no less than proficient with AI tooling. And in the event that they’re solely proficient, meaning they must be additionally proficient and have depth in one other space, whether or not it’s databases, whether or not it’s cryptography and so forth, so fourth,” Wang stated.
Fluency with AI instruments isn’t nearly understanding use Claude Code, however with the ability to create multi-step workflows and seize the state of a number of methods. Wang would additionally search for engineers with nuanced approaches to automation, corresponding to utilizing Claude Code to generate code and designs after which Codex to spin up take a look at harnesses.
Because the product improvement lifecycle bleeds into the software program improvement lifecycle, Wang factors to the delivery of a product engineer function, somebody who builds prototypes, has the flexibility to speak to prospects and might step into the footwear of a product supervisor.
Life After AI
The rising recognition of instruments like Claude Code has modified coding workflows for good. As firms like Synthesia and 1Password look to automation to streamline product innovation, builders will likely be underneath higher stress to show proficiency with AI instruments.
Within the quick time period, builders will play a crucial function by working alongside AI brokers, double-checking code, and growing advanced workflows that hook up with downstream methods. However the way forward for the software program engineer’s function is unsure. If these instruments turn into superior sufficient to write down and deploy code unbiased of human supervision, there may also be much less demand for builders.
Software program engineering is “cooked” within the sense that builders are going to must work a lot tougher to show their worth out there. If engineering groups can do extra work with AI brokers and fewer general headcount, there will likely be fewer roles accessible to human engineers.