TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
The Pragmatic Engineer published a podcast interview with Titus Winters, a former Google engineer and lead author of Software Engineering at Google, now at Adobe. They discuss Google’s C++ refactoring effort, testing and code-review practices, computer science education and the mixed effects of AI on software development. The interview does not announce a new Google policy or a finalized change to university curricula.
The Pragmatic Engineer has published a podcast interview with Titus Winters, discussing engineering practices he worked on during more than a decade at Google and his role in updating undergraduate computer science curriculum guidance. Winters, now a senior principal scientist at Adobe, describes a large-scale C++ refactor, Google’s testing and code-review approaches, and the pressures AI tools may place on software quality. The interview is a discussion of his experience and views, not an announcement of new Google policy or an adopted curriculum change.
Winters joined Google in 2010 and worked on the company’s readability program and Abseil, its open-source foundational C++ libraries, a topic explored in software engineering. The report says he was also the lead author of Software Engineering at Google. In the interview, he recounts a project to update C++ APIs across a codebase that then contained close to 250 million lines and was edited by about 10,000 engineers. A team of roughly a dozen people carried out the work.
According to Winters’ account, the team used language parsing and automated edits: it parsed build targets into abstract syntax trees, matched particular code patterns and generated changes for engineers to review and apply in batches, an approach that connects with pushing software engineering limits. The old and new APIs had to coexist during the rollout, and the team worked to prevent the old pattern from returning before removing the old API. The report links the project’s needs to the creation of the clang-format toolkit. Winters likened the task to replacing a building’s foundation without closing it.
The conversation also covers Google’s use of tests and code review. The report says Google made tests mandatory for changes after Search struggled to maintain its previous release cadence, and describes a review model requiring sign-off from a project owner and a language-specific readability reviewer. The report says more than 10,000 engineers were enrolled in the readability program around the time Winters left Google. These details are presented through the interview and the publication’s account, rather than a newly released Google statement.
How Google Built Safer Code Changes
The interview’s central engineering lesson is that safe change depends on safeguards, not just individual coding skill. Tests, structured review and tools for broad mechanical edits can help large teams modify shared systems while reducing the risk that a change will break other work. That matters to organizations managing extensive codebases, where a seemingly local update may affect multiple products or teams.
Winters also raises a question for teams adopting AI coding tools: whether higher code output is producing more useful software or adding operational risk. The report attributes to him an observation from conversations with engineers at other companies that outages were up about 30% while feature output remained static. The source gives no defined sample, measurement period or comparison method, so this should be read as his reported observation, not as a broad industry statistic. It nevertheless frames a practical issue for engineering leaders: productivity measures should account for reliability and the work required to maintain generated code.
For students and educators, Winters’ involvement in curriculum guidance connects the discussion to a wider question about what new engineers need to learn. The interview raises the need to revisit conventional computer science teaching as software development changes, but the supplied report does not set out a finalized replacement curriculum.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Amazon Kindle Paperwhite Signature Edition (12th Generation, 2024), 32 GB, Metallic Black
Logitech MX Keys S Wireless Keyboard, Graphite
As an affiliate, we earn on qualifying purchases.
From Google Refactoring to Curriculum Work
The report places the C++ refactor in 2010, shortly after Winters joined Google. Rather than relying on today’s large language models, the team used parsing, exact pattern matching and generated edits to change code at scale. The old and new interfaces had to remain compatible during the rollout, allowing the system to keep operating as changes were distributed across teams.
The testing discussion reaches back to around 2007, when, according to the report, Google Search could no longer ship as often as it had previously and was releasing to production roughly quarterly. The report says Google required changes to include tests, with Search later returning to a faster cadence. Winters’ discussion of education draws on his work updating ACM/IEEE undergraduate computer science curriculum guidelines; the source does not say that any proposed revision has been approved or implemented.
“We were replacing the foundation of the Empire State Building. We were not allowed to close the building or block traffic. Oh, and we discovered nuclear waste in the basement.”
— Titus Winters, as quoted in The Pragmatic Engineer’s account of the C++ refactor
Limits of the AI Outage Estimate
The report does not provide a recording transcript or supporting data for the roughly 30% outage increase Winters mentions. It does not identify the companies, time period, outage definition or baseline, and the observation cannot establish that AI tools caused outages or that the pattern applies across the industry.
It is also unclear from the supplied material what specific changes Winters advocates for the undergraduate curriculum, which organizations have reviewed those proposals, or whether any recommendations have been adopted. The report’s account of Google’s historical engineering practices does not establish whether those policies remain unchanged today.
Curriculum Proposals Still Need Detail
The immediate next step for readers is the full podcast and transcript referenced by The Pragmatic Engineer, which may provide more detail on Winters’ views about curriculum design, AI and engineering productivity. The supplied report does not identify a scheduled publication, approval vote or implementation date for curriculum revisions.
Further evidence would be needed to evaluate the outage claim, including a defined time window, a comparison baseline and data showing how the cases were selected. Until then, the interview is best understood as a practitioner’s account of engineering trade-offs and an argument for closer attention to testing, review and reliability as development tools change.
Key Questions
What is the news development?
The Pragmatic Engineer published a podcast interview with Titus Winters about his Google engineering work and his views on computer science education and AI-assisted software development.
Who is Titus Winters?
Winters spent more than a decade at Google, where he led its readability program and worked on Abseil. The report identifies him as a senior principal scientist at Adobe and the lead author of Software Engineering at Google.
What did the Google C++ refactor involve?
Winters described a team of about a dozen engineers using code parsing and automated edits to update APIs across a codebase reported to contain close to 250 million lines. The work was staged so old and new APIs could operate together during the change.
Does the report prove AI is causing more outages?
No. Winters reported an approximate 30% rise in outages from conversations with engineers at other companies, but the report gives no sample, time window, baseline or causal evidence. It is an attributed observation, not a verified industry-wide finding.
Has a new computer science curriculum been adopted?
The source says Winters worked on updating ACM/IEEE undergraduate curriculum guidelines, but it does not report an approved or implemented revision or specify a timetable.
Source: newsletter.pragmaticengineer.com
