With AI developments, software program professionals might must brush up on pure language greater than programming ones.
For Feroz Sheikh, chief data and digital officer at agricultural-tech firm Syngenta, this supposedly new give attention to utilizing pure language to program is a callback to an idea known as “literate programming,” launched greater than 4 a long time in the past by Stanford researcher Donald Knuth.
With a “literate” improvement method, you clarify this system in pure language, like an essayist, utilizing each phrases and code snippets. This forces programmers to pay attention not on instructing a pc, however explaining to people what a pc can do.
“My applications will not be solely defined higher than ever earlier than; in addition they are higher applications, as a result of the brand new methodology encourages me to do a greater job,” Knuth wrote in a 1984 paper.
Like shoulder pads and energy fits, Sheikh sees this ’80s method making a comeback, as AI simply accomplishes the what, whereas the human developer performs the important function of contemplating the why. “However precisely why you wrote [a particular code loop] is usually misplaced in that deterministic instruction to the machine,” he stated.
Sheikh and different tech leaders spoke with IT Brew at Amazon’s AWS Summit, hosted in New York Metropolis on June 17, the place the corporate announced new agentic tools, options, and type components. These automations, aimed toward making life simpler for builders, go away software program professionals with essential duties involving the communication of intent.
Proud to announce. Listed below are just a few notable instruments introduced by AWS this week:
- AWS Continuum. The device, obtainable in gated preview, finds code vulnerabilities, prioritizes them, and affords remediation help.
- Kiro, obtainable on iOS. This prompt-to-spec improvement device is now supplied proper out of your iPhone.
- Amazon Bedrock Managed Information Base. With this function (and an agentic “retriever” of related context) builders construct GenAI apps utilizing their very own knowledge sources.
Brokers at work. So, what’s left for a developer to do when AI finds your bugs, builds your app, and connects you to your organization’s related knowledge repos?
It’s not unusual for David Yanacek, a senior principal engineer at AWS engaged on agentic AI, to have a half-dozen coding brokers operating without delay—one constructing a function, possibly others doing a repair or finishing up different duties. For fine-tuning his plain-language or command-line specs, he depends on Kiro, which turns specs (descriptions of desired necessities, designs, and duties) right into a type of blueprint for the ensuing code era and testing.
High insights for IT professionals
From cybersecurity and large knowledge to cloud computing, IT Brew covers the newest developments shaping enterprise tech in our 4x weekly publication, digital occasions with trade consultants, and digital guides.
With a lot of the work automated, Yanacek finds himself spending extra time contemplating how an agentic answer ought to behave, and reviewing specs to get to that answer: “I’m doing loads of determining what to do and the way I need to go about it.”
In an October 2025 survey from software-security firm Sonar, which polled over 1,100 world builders, the duties ranked most extremely efficient by the builders included writing documentation, explaining or understanding present code, and vibe-coding or creating new tasks with principally AI-generated code. Three out of 4 builders stated that AI diminished their time spent on “toil work.”
Asking round. On the AWS occasion, Cristina Pieretti, basic supervisor and head of digital content material and innovation at knowledge, analytics, and intelligence agency Moody’s, stated she sees builders freed up by AI to deal with essential challenges like analyzing outputs and stopping hallucinations.
“Possibly you’re going to do much less coding, for instance, however you need to just remember to have the correct governance on that coding,” Pieretti stated.
Jimmy Hatzell, CEO and co-founder of AI platform Hatz AI, sees IT professionals changing into extra like a business-process supervisor, that’s, a marketing consultant targeted on programs design quite than “the technical nuts and bolts.”
Hatzell appears for a extra summary talent from right now’s software program professionals: an creativeness. “As organizations undertake AI, having the ability to rethink processes and the way issues are finished in an AI‑first method turns into crucial,” he informed us on the AWS occasion.
Syngenta has deployed its share of brokers for duties like serving to farmers with seed choice or checking out supply-chain bottlenecks. These sorts of automations have led to an rising talent that Sheikh desires on his groups: linguistics. If an agent receives a poorly written coverage doc, for instance, it might misread the info.
People must put the directions into clear language.
“Creating the agent will not be sufficient. You want people who find themselves doing this work of reviewing it, figuring out gaps or points in that coverage doc, both rewriting it or having a dialog with the enterprise companions that you should change the coverage,” Sheikh stated.