While Anthropic CEO Dario Amodi Announced that Ai would Write 90 % code Within six months, the coding world was widely ready to become extinct. But inside Sales forceA different fact has already taken shape.
“All about 20 % Apex code Was written in the past 30 days Agent force“Senior Vice President of Sales Force AI, Jiash Govindrajan, told me during a recent interview. His team has not only been produced, but the code has been reality.
Yet the sales force developers are not disappearing. They are getting ready.
Govindarjan acknowledged, “The vast majority of development – at least whom I call the first draft of the code – will be written by AI.” “But what developers do with this first draft are basically changed.”
From code lines to strategic control: How are the developer becoming a technology pilot
Software engineering has always mixed creativity with the tedium. Now Ai later handles the developers to the former.
Govindarjan explained, “You purely move toward more strategic than a technical technical role.” Just ‘I have nothing to make, so I will make it,’ but ‘what should we make? What does the user actually want? ‘
This change mirrror other technical obstacles. When the calculators replaced the manual count, mathematics did not end – they dealt with more complex problems. When digital cameras killed the Dark Rooms, the photography spread instead of the deal.
The sales force believes the code works like this. Since AI software reduces the cost of creation, developers get that they always lack: time.
“Once it takes weeks to make a working prototype, it takes hours,” said Govindrajan. “Instead of showing users a document, you can hand them over to working software just. Then you repeat on the basis of their reaction.”
‘VIBE coding’ is here: Software engineers are now orching AI instead of typing each command now
Coders have begun to adopt what they say “Web coding“-A term developed by the co-founder of Openi, Andridge Carpathi, involves giving the AI a high level direction instead of the exact instructions, then it is to improve its preparation.
There is a new type of coding that I call “vibing coding”, where you give fully to the web, embrace the efficacy of the vocabulary, and forget that the code is also available. This is possible because LLM (such as cursor composer W. Swant) is getting great. Also I just talk to the composer with the Super Wasper …
– Andridge Carpati (@carpet) February 2, 2025
“You just give it a high level direction and allow AI to use your creativity to develop the first draft,” said Gwanderjan. “It won’t work as you want, but it needs to play something. You improve some of its parts by saying, ‘it looks good, do more of it,’ or ‘the buttons that are, I don’t need them.’
He compared the process with musical support: “AI fixes rhythm while the developer fixes the rag.”
While AI takes the lead in producing straightforward business applications, Govindrajan admits that it has boundaries. “Are you going to create next -generation database with web coding? Not likely. But can you really make a cool UI that calls the database and produces an excellent business application?”
New standards Mandatory: Why Testing Strategies should develop because AI produces more production code
AI does not just write the code differently – it requires different quality control. The sales force made its progress Agent Force Testing Center New methods of verification have been called after discovering the machine -made code.
“These are the stockstick systems,” said Govindarjan. It may fail in step 3, or step 4, or 17 steps in 17 steps.
The unofficial nature of AI outpots means that developers should be experts in boundary testing and gurdel sequence. They not only need to know how to write the code, but also to evaluate it.
Beyond Code Generation: AI how the entire software is compressing the Development Life Cycle
The change goes beyond the initial coding to surround the full software life cycle.
“In the construction stage, the tools understand the current code and increase it with intelligence, which intensifies everything,” said Gwanderjan, which accelerates everything. ” “Then the test comes – giving birth to regressive tests, creating test cases for the new code – all of which can handle AI.”
This comprehensive automation creates something that is called “significantly tough loop” between governing Idea and Implementation. Fastest developers can test and improve, as they can become more expensive.
The algorithmic thinking is still important: Why the basic principles of computer science are essential in the AI era
Govindrajan is often in the field of anxious questions about the future of software engineering.
“I am permanently asked if people should still study computer science,” he said. “The answer is yes, because the algorithmic thinking is essential. Breaking major issues into manageable pieces, understanding which software can solve which issues, modeling user needs – these skills are more valuable, not less.”
What are the changes how this skill shows? Instead of typing the role of each solution in terms of character, developers guide AI tools toward maximum results. Man provides a decision. The machine provides speed.
“You still need a good intuition to give the right instructions and to evaluate the output,” said Govindarjan. “It takes the real taste to see what AI produces and recognizes what works and what not.”
Strategic height: How is the developer becoming a business partner instead of technical implementation
Since coding itself becomes commodities, the developer’s roles are more directly connected to the business strategy.
“Developer is playing the role of caretaker, leading the agents working on their behalf,” said Govindarjan. But they are responsible for what they are deployed. The deer still stops with them. “
This height keeps the developers closer to the decision makers and closer to the details of the implementation.
The Sales Force supports this transition with tools designed for each stage: Agent Force Code for Developers handles generation, the agent builder enables customization, and the Agent Force Testing Center ensures reliability. Together, they create a platform for developers to increase these expansion roles.
The company’s vision offers the “developers ruined” statement. Instead of codeing to be obsolete, the software engineers who are adapted may find themselves more important than ever.
In a field where the renovation is normal, AI represents the most powerful composition – not only how the code is written, but who writes and why. For developers who want to upgrade their mental models, the future looks like less and exceeded.
Credit : venturebeat.com
