Podcast Summary: The Cutting Edge of Software Development in the AI Era

Podcast: Bloomberg Businessweek
Hosts: Carol Massar & Tim Stenovec
Guests: Ryan J. Salva (Senior Director of Product, Google), Mandeep Singh (Bloomberg Intelligence)
Air Date: November 10, 2025

Overview

This episode explores the profound ways artificial intelligence is transforming software development. The discussion focuses on how AI tools—particularly those deployed by Google, such as Gemini Code Assist—are reshaping developer workflows, accelerating productivity, altering required skill sets, and influencing the broader job market within technology. The guests also touch on the challenges and opportunities associated with legacy system modernization in the AI era.

Key Topics and Insights

1. AI’s Impact on Developer Productivity

• AI Assistants Streamlining Code Writing
  Ryan J. Salva explains how AI tools enable developers to shift their focus from repetitive coding tasks (like “if/then/else” loops) to higher-level problem solving.
  • [02:54] “So often developers and organizations are really just trying to deliver user requirements... they're able to use AI and large language models to write those requirements in natural language, translate that to code, and through that ultimately accelerate their pace of iteration, their pace of learning.” —Ryan J. Salva
• Widespread Adoption and Internal Metrics
  • [03:55] “DORA effectively surveys thousands and thousands of engineers... One of the things that we're seeing is that today roughly 90% of developers are integrating AI into their everyday work... At Google, what we see is today roughly 50% of our code is being written by AI. And I want you to like, stop and maybe imagine... 5, 0, 50% of code is being written by AI.” —Ryan J. Salva
  • This shows a tidal wave of adoption and how AI is enabling faster iteration and less “preciousness” about each line of code.
• Memorable Analogy
  • [05:41] The co-host likens the shift to AI-assisted coding to moving from film cameras to digital, where experimentation is easy and costless:
    “It's akin to when we got digital cameras on our phone... Now I just take a million photos. I do wonder though, if we're less precious, we're more efficient, we're more productive…”

2. Changing Skillsets & the Developer Role

• From Syntax to Systems Thinking
  • [06:08] “It's not so much that the developer's job is any less important, but what it does mean is that our job requirements are changing... We're really asking developers to think more like architects, to think about systems design, to think about negotiating the contract between components.” —Ryan J. Salva
  • The role of a developer is evolving from “coding in a language” to solving problems and thinking about large-scale system design.
• Hiring Trends in the AI Era
  • Despite headlines about declining tech jobs, Ryan clarifies Google is actively hiring engineers, including product managers and architects, emphasizing the evolving nature of essential skills:
    • [07:46] “The headline's a little bit easier to grab attention than the ground level reality... writ large across the industry, I'm still seeing a very, very healthy engineering ecosystem.” —Ryan J. Salva
    • Entry-level jobs are changing rather than disappearing—skills in systems thinking and problem-solving are increasingly valuable.

3. AI’s Role in Code Quality and Validation

• Determinism as an Advantage in Software Development
  • [09:39] Ryan discusses how software development is uniquely suited to benefit from large language models because quality and safety can be objectively measured via tests and analysis:
    • “We have decades of deterministic quality measures... so what I see is a lot of organizations using AI... to accelerate their engineering life cycle because they can deterministically say, this is of good quality, this is of bad quality...” —Ryan J. Salva
  • This is contrasted with creative tasks (like writing an email), where quality is subjective.

4. Token Consumption and AI Scale at Google

• AI’s Massive Scale
  • [09:07] “1.3 quintillion [tokens].” —Mandeep Singh (Carol jokes about explaining this number to her son)
  • Token consumption at Google is staggering; coding assistants contribute heavily, but the line between use cases (search, coding, homework) isn’t strictly tracked by token.

5. Modernization & Legacy Systems

• AI as a Tool for Modernization
  • [11:37] Ryan highlights the critical role of AI in helping organizations modernize or migrate legacy systems, especially as skills for maintaining old systems dwindle and the cost of “code rot” rises.
    • “Migration and modernization is one of the areas where I see the most interest among large engineering teams today... AI is good in that without consuming too many tokens or increasing your bill.”
    • In fact, companies are often willing to process huge amounts of tokens because the cost of maintaining old infrastructure is so high, both in terms of talent and operational overhead.
  • [12:14] “If you even just take the cost of maintaining them to the side, the fact that you're not able to carry those applications forward and innovate with them and do new things, often that's the real cost.” —Ryan J. Salva

Notable Quotes & Moments (with Timestamps)

• On AI’s Share of Google’s Codebase:
  “At Google, what we see is today roughly 50% of our code is being written by AI. And I want you to like, stop and maybe imagine... 5, 0, 50% of code is being written by AI.” —Ryan J. Salva [03:55]

• The Digital Photography Analogy:
  “It's akin to when we got digital cameras on our phone... Now I just take a million photos.” —Co-host [05:41]

• On Entry-Level Tech Job Fears:
  “The headline's a little bit easier to grab attention than the ground level reality... writ large across the industry, I'm still seeing a very, very healthy engineering ecosystem.” —Ryan J. Salva [07:46]

• On Deteminism in Software Quality:
  “We have decades of deterministic quality measures... organizations using AI... to accelerate their engineering life cycle because they can deterministically say, this is of good quality, this is of bad quality...” —Ryan J. Salva [09:39]

• On Costs of Maintaining Legacy Systems:
  “If you even just take the cost of maintaining them to the side, the fact that you're not able to carry those applications forward and innovate with them and do new things, often that's the real cost.” —Ryan J. Salva [12:14]

Segment Timestamps

• [01:35] Episode content begins, noting Alphabet’s investment in AI and introduction of guests.
• [02:54] Ryan J. Salva explains AI's role for developers.
• [03:41] Discussion of productivity metrics and use cases.
• [05:41] Analogy for changing developer practices due to AI.
• [06:08] Impacts on developer hiring and required skills.
• [07:16] Questioning narratives about shrinking tech job opportunities.
• [09:06] Token consumption and Google’s scale in AI.
• [11:37] Migration and modernization of legacy software systems.
• [12:14] AI’s value in modernizing legacy apps and controlling costs.

Summary

The episode provides an in-depth look at how AI tools like Google’s Gemini are revolutionizing software development by automating code generation, enabling rapid prototyping, democratizing problem-solving, and driving major shifts in developer skillsets. The tone throughout is enthusiastic about AI's benefits, while also realistic about ongoing changes in job requirements and legacy systems. For listeners, the conversation offers both reassurance and guidance on how to thrive as the industry evolves.