Future of Software Engineering in the Age of AI

James Gress

Emerging Technology and AI Lead / Accenture

LinkedIn: - jamesgress
GitHub: jmgress
X.com: @jmgress
Tampa Bay Generative AI Meetup
Tampa Bay DevOps Meetup
Tampa Bay Platform Engineering Meetup

The Evolution of Software Engineering in the Age of AI

center

center
center

The Evolution of Software Engineering with AI

November 30, 2022: ChatGPT Launches

  • OpenAI released ChatGPT to the public as a free research preview.
  • Marked a significant milestone in AI, introducing a conversational interface for users.
  • Sparked widespread interest and rapid adoption in various industries.

2023: AI-Powered Code Completion Becomes Mainstream

  • Tools like GitHub Copilot, Tabnine, and Amazon CodeWhisperer gained popularity.
  • Provided developers with real-time code suggestions within IDEs.
  • Enhanced productivity and streamlined the coding process.

2024: Conversational Coding Interfaces Emerge

  • Integration of AI chat interfaces into development environments.
  • Tools like GitHub Copilot Chat allowed natural language interactions with code.
  • Revolutionizing the way developers write and refine code.

2025: Agentic AI Systems Transform Development Workflows

  • Emergence of agentic AI systems capable of autonomous task handling.
  • Systems interpret high-level objectives and modify code across multiple files.
  • Transformed software development by reducing manual intervention.

2025: Emerging Generative AI

  • A2A - Agent to Agent Communications
  • MCP - Model Context Protocols
  • LOKA - Layered Orchestration for Knowledgeful Agents
  • Trusted Agent Huddle

Jensen Huang - CEO NVIDIA

"AI is not going to take your job. The person who uses AI is going to take your job."

"You probably recall over the course of the last 10, 15 years almost everybody who sits on a stage like this would tell you that it is vital that your children learn computer science. Everybody should learn how to program. And in fact, it’s almost exactly the opposite."

"Software is eating the world, but AI is going to eat software."

— Various speeches, 2023–2024

Satya Nadella - CEO Microsoft

"AI is the runtime that will shape all of what we do"

— December 2024

"I think what'll happen is these crud I mean SAS applications are a crud database with a lot of business logic so the crud database will then get orchestrated outside of the business logic tier of just the SAS application"

— January 2025

Eric Schmidt - Former Google CEO

"We believe as an industry that in the next one year the vast majority of programmers will be replaced by AI programmers."

— August 2024

Dario Amodei - CEO of Anthropic

“I think we'll be there in three to six months, where AI is writing 90% of the code.
And then, in 12 months, we may be in a world where AI is writing essentially all of the code.”

— March 2025

LlamaCon 2025 LIVE: Mark Zuckerberg chats with Microsoft CEO Satya Nadella

Fake software company with AI agents

Researchers at Carnegie Mellon staffed an entire fake software company with AI agents from OpenAI, Google, Anthropic, and Meta to see how they'd perform in roles like software engineers, project managers, and financial analysts.

Spoiler: If you've been worried about ChatGPT stealing your job, you can relax — turns out the bots couldn't even steal pretend jobs.

Sam Altman – CEO of OpenAI

Questions was asked "What does my future look like?"

"There are sort of two views you can take.
You can say, ‘Oh man, it’s doing everything I can do,’
or you can say, ‘Look at all the new things I can now do.’"

— April 2025

Vibe Coding

Initial Concerns on Vibe Coding

  • Do we care about the code?
  • What if the code accesses sensitive data?
  • Loss of architecture and design principles
  • Erosion of team understanding and collaboration
  • Testing becomes an afterthought
  • Skill atrophy for developers

Do We Really Care About the Code?

  • Do we care about the code — or just the outcome?

  • AI has already shifted part of our mindset — treating code as a temporary utility, not a permanent product.

  • AI is already being inserted into code bases and we don't know even what it really does.

center
center
center
center
center

Vibe Coding for the Enterprise

Image center

Image center

Initial Findings on Vibe Coding

Strengths:

  • Excellent at adding utility features
    • Command line argument handling
    • Logging implementations
    • Basic functionality enhancements

Limitations:

  • Struggles with advanced architecture
    • Provider patterns
    • Object-oriented design principles
    • Complex software architecture

Agentic Coding

  • Definition:
    Agentic Coding refers to the use of AI systems that can autonomously interpret high-level objectives and execute tasks across multiple files or systems without requiring step-by-step instructions.

  • Key Characteristics:

    • Operates with autonomy to achieve defined goals.
    • Capable of multi-step workflows and decision-making.
    • Integrates with existing systems to modify, test, and deploy code.

Why Agentic Coding Matters

  • Efficiency: Reduces manual intervention in repetitive tasks.
  • Scalability: Handles complex, large-scale changes across systems.
  • Innovation: Allows developers to focus on high-level design and strategy.

"Agentic Coding is not just about writing code—it's about enabling AI to act as a collaborator in the software development lifecycle."

Image center

Image center

Image center

Image center

Task Role Copilot Played
Project structure setup Auto-suggested folders and files
Backend implementation Generated API structure, models, DB init, endpoints
Frontend implementation Created React components and state logic
Styling Wrote modern, responsive CSS
Troubleshooting Detected and fixed package issues
Fallback HTML Generated a fully functional vanilla JS fallback UI
Integration Helped with Axios/fetch API setup
Documentation Drafted detailed, well-structured README
Scripts Wrote CLI launcher script for local dev

How do we embrase AI for Software Engineering?

Calibrated Trust Between Humans and Machines

  • Trust isn’t binary—it's a spectrum
  • We must know when to rely on AI and when to verify
  • Align confidence with competence
  • Design workflows that include human-in-the-loop oversight

“We don’t need blind trust—we need calibrated collaboration.”

AI as a Teammate, Not a Tool

  • AI isn’t replacing your team—it’s joining it
  • Like a great teammate, it doesn’t wait for instructions
  • It watches, understands, and suggests—proactively
  • But like any teammate, it still needs guidance

“Think of AI as the pair programmer that never sleeps.”

AI Writing Code ≠ Ignoring DevOps

"When we say AI will write the code, we don’t mean we’re skipping the discipline."

AI is transforming how code is created
but code quality, security, and delivery pipelines still matter.

  • DevOps isn't going away
  • Guardrails are still essential
  • Testing, observability, and governance remain critical

We’re not removing the need for rigor —
we're elevating it to meet the speed of AI.

Don’t Just Automate the Past—Reimagine the Future

“It’s like putting a jet engine on a horse-drawn carriage. Sure, it moves faster, but you’re missing the point—the whole vehicle can be reinvented. Generative AI gives us a chance to reimagine how we deliver, not just speed up what we already do.”

From Asking to Advising: A New AI Interaction Paradigm

  • Stop building chatbots. Start building outcomes.
  • Don’t make users ask—let AI advise
  • Systems should anticipate needs, not just respond
  • The future UX is ambient, embedded, and contextual

Preparing for the Future

  • Learn agentic systems, system thinking, and architecture
  • Embrace change: experimentation and transparency
  • Grow internal champions for AI adoption

Training Agents using Agents with Memory

The End of "Tech for Tech’s Sake"

  • Old mentality: "I just code, not business"
  • AI needs business context to generate useful output
  • Engineers must understand goals, not just syntax

“In the future, the best engineers won’t just know how to code — they’ll know why the code matters.”

Closing: A Call to Lead

"AI is the runtime that is going to shape all of what we do…" — Satya Nadella

  • Will you be a passenger — or help engineer the future?
  • Start learning. Start building. Start leading.

“The next chapter of software engineering is being written now — and we’re all holding the pen.”

Questions?

Done 100's of Prototypes Taken 10 applications to Production ranging from simple RAG to more complex Agentic systems

Software engineering was split - naysayers and enthusiasts Companies were concerned about data leakage copyright infringement ability to copyright code it generated 99.99% of code people write, is useless to others. Many code suggestions were not great Able to complete comments IntelliCode – Announced in May 2018 IntelliCode was introduced as an experimental feature in May 2018. It leveraged machine learning to offer context-aware code completions based on patterns learned from thousands of open-source projects. GitHub Copilot – Integrated into Visual Studio on March 29, 2022 GitHub Copilot became available for Visual Studio 2022 on March 29, 2022. This AI-powered tool, developed by GitHub and OpenAI, provided advanced code suggestions and completions. https://en.wikipedia.org/wiki/GitHub_Copilot?utm_source=chatgpt.com

This really allowed things to be productive Tie Back to Stack Overflow Things started to really change and opinions started to change, and more of the naysayers started to turn and get on board.

With some of these advancements we seen many of the early skeptics realize we are not going back

MCP is like RPA (Robotic Process Automation), but instead of mimicking user actions through the UI, it gives AI a direct, intelligent channel into the application—so different AI agents can talk to each other and to the app itself in a smarter, more coordinated way. Carnegie Mellon University April 28ish 2025, Layered Orchestration for Knowledgeful Agents (LOKA) is a proposed framework designed to enable AI agents to operate collaboratively, ethically, and securely across diverse systems and environments. Introduced by researchers from Carnegie Mellon University in April 2025, LOKA aims to address challenges related to identity, accountability, and ethical alignment in autonomous AI agents Trusted Agent Huddle https://newsroom.accenture.com/news/2025/accenture-introduces-trusted-agent-huddle-to-allow-seamless-first-of-its-kind-multi-system-ai-agent-collaboration-across-the-enterprise

Jensen Huang is the founder and CEO of NVIDIA, a global leader in AI and GPU computing. He has been vocal about how Generative AI is reshaping the future of programming. His vision suggests a world where AI bridges the gap between human intent and technical implementation, making traditional programming optional for many tasks. These quotes were made during keynotes at events like Computex, GTC, and various interviews in 2023 and 2024. Huang emphasizes that the future of software development lies in human-centric interaction with machines, moving from code to natural language prompts.

In the interview December 2024, Nadella elaborated on how AI agents are poised to revolutionize the traditional software model: Essentially he said Agents will replace Software​ https://www.youtube.com/watch?v=9NtsnzRFJ_o&t=1122s <!-- Satya Nadella on the Future of SaaS, How 2025 is the year of Agents, Advice for Indian Engineers https://www.youtube.com/watch?v=GuqAUv4UKXo&t=103s

Eric Schmidt is the former CEO of Google (2001–2011), where he helped transform the company into a global tech powerhouse. After his CEO tenure, he served as Executive Chairman of Google and then Alphabet Inc., contributing to the company’s growth and strategic vision. He co-founded Schmidt Futures, a philanthropic initiative focused on advancing technology and talent to solve global problems. Eric has also chaired the U.S. National Security Commission on Artificial Intelligence. This quote was made during a talk at **Stanford University on August 15, 2024**. In his remarks, Schmidt discussed the rapid advancement of AI and its implications on software engineering. He highlighted the concept of recursive self-improvement in AI and predicted that AI systems would soon outperform the majority of human programmers. The talk was later removed from public platforms at Schmidt’s request due to its provocative nature.

Dario Amodei — Anthropic CEO Dario Amodei is the co-founder and CEO of Anthropic, an AI safety and research company best known for creating Claude, a family of large language models. Prior to founding Anthropic in 2021, Dario was VP of Research at OpenAI, where he led the development of GPT-2 and GPT-3. He has a background in physics and machine learning, and has been one of the most vocal advocates for aligning AI systems with human values and ensuring they can be controlled and understood. This quote was delivered during a panel at the Council on Foreign Relations on March 10, 2025. In it, Dario predicted a near-term future where AI writes 90% or more of all code, with human engineers focused more on high-level goals, design inputs, and oversight rather than the manual construction of software. AI companies and government need to stop "sugar-coating" what's coming: the possible mass elimination of jobs across technology, finance, law, consulting and other white-collar professions, especially entry-level gigs.

Satya 30% code is being written by AI During this discussion between Meta CEO Mark Zuckerberg and Microsoft CEO Satya Nadella at LlamaCon 2025, they primarily focused on the transformative impact of AI, comparing it to previous technological shifts like client-server, the web, mobile, and cloud. They explored how AI necessitates a rethinking of the entire technology stack and discussed Microsoft's strategy for providing world-class infrastructure and tools through Azure to support the development of AI applications and agents. A significant portion of the conversation centered on the importance of open-source AI and the concept of a "distillation factory" for creating smaller, more efficient models from larger ones, which they believe will be crucial for widespread accessibility and innovation in the AI ecosystem. They also touched on the impact of AI on developer and knowledge worker productivity, citing examples within their own companies and envisioning a future where developers work with AI "agents."

https://futurism.com/professors-company-ai-agents

Sam Altman is the CEO of OpenAI This quote was delivered during a live interview at **TED2025 in Vancouver on April 11, 2025**. In his remarks, Altman addressed growing concerns about AI taking over human tasks. He framed the moment not as a loss of human agency, but as a transformative leap—where AI becomes a partner that expands what humans can achieve, rather than a replacement for their work. https://www.youtube.com/watch?v=5MWT_doo68k A bit proactive and right now I would say to use AI effectively you do need a base understanding on how it works and the best ways to get the most out of it. I do feel very quickly some of the capabilities will become embedded into the applications and be a bit more automatic Like is a voice system that can determine intent and offer up suggestions to a customer service representative to queue up suggestive actions

I generated HTML code for the flashcards. Specifically, it's a complete single-page web application that includes: HTML - The structure and content of the flashcards CSS - Styling for the flip animations, colors, layout, and responsive design JavaScript - Interactive functionality like: Card flipping animations Navigation between cards Shuffle functionality Keyboard controls (arrow keys, spacebar) Dynamic content updates

Story about adding sarcastic comments in the code messed up the generation

- **Examples**: - AI agents that refactor codebases to improve performance. - Systems that generate and validate unit tests automatically. - Tools that handle repetitive coding tasks, freeing developers for creative work.

**Speaker Notes:** This slide is meant to clarify a misconception: just because AI can generate code, that doesn’t mean we’re abandoning the rigor of software engineering practices. In fact, it's the opposite. As AI accelerates delivery, DevOps becomes even more important. Guardrails, automated testing, static analysis, CI/CD pipelines — these are all essential for ensuring what AI creates is robust, secure, and maintainable. The mindset isn’t “AI replaces everything,” but “AI amplifies our ability to deliver — within a system that ensures quality and trust.”

Corralite the early days of Agile Too often, companies approach generative AI as a tool to accelerate old, linear processes—just plugging AI into the same waterfall pipeline: requirements, designs, coding, testing, and deployment. But real transformation comes when we realize we can rebuild the entire vehicle, not just add horsepower to the old one. Generative AI allows us to move to more agile, feedback-driven delivery models. We can shorten feedback loops, continuously refine requirements and implementation, and focus on outcomes over outputs. Let’s challenge ourselves and our organizations: Instead of simply making old processes faster, how might we deliver value in entirely new ways—ways that only become possible because of AI?

Simple tactical stuff I just end up doing it myself the real skill will be understanding the business

Previous Questions Need a slide to address, coding for weapons is different that a CIS system Address the Legal Concerns on Generated Code and Legal Stack Overflow Need a slide to show the difference between AI writing code AI being the code

Needed for mermaid, can be anywhere in file except frontmatter