Dev with Serdar

Learn coding in Python, Go and Rust from Serdar Yegulalp, software dev specialist and senior writer at InfoWorld.

Ep.277

What AI really is doing to software development (and developers)

A newly released report from Faros crunched two years of live, in-the-wild data about AI usage and adoption. The results are eye-opening. No question AI is having a major transformative effect on software dev. But many of the gains are being drastically offset by losses elsewhere in the dev cycle ... and many of those losses are invisible and insidious.

Ep.277

Sep 29, 2026 5 mins

Developer

Ep.276

Did OpenAI STEAL Math Researchers' Work? The AI Controversy Explained

A pair of math researchers working on a hard problem found out OpenAI had a solution to the same problem they were working on. Was their work co-opted without their knowledge? It's complicated -- and the implications for how much we can trust AI services with our work just got a lot murkier.

Ep.276

Sep 16, 2026 6 mins

Python

Ep.275

Why Software Developers Are “Token Maxxing” AI—and What It Means for Jobs

Software developers are using AI at a level that other professions may never match—but what does that mean for the future of software engineering? In this episode, we look at why coding has become such a massive driver of AI usage, and why the way developers use AI may be fundamentally different from how AI will be used in fields like finance, law, and consulting. We also examine the growing push by major software companies to shrink engineering teams and rely more heavily on AI. Meta’s attempted cuts to its engineering workforce offer a cautionary example—but the risks may go far beyond employee morale or productivity. Cutting software teams can also mean losing three things that are difficult, if not impossible, for AI to replace: diversity of experience, domain knowledge, and human discernment. So while software developers may be “token maxxing” AI today, that doesn't necessarily mean the rest of the working world will follow. And even within software development, there may be hard limits to what AI can actually replace. What do you think? Is AI going to make software development teams dramatically smaller, or are companies underestimating what they lose when they cut experienced engineers?

Ep.275

Sep 11, 2026 7 mins

Python

Ep.274

WebAssembly is getting WILD these days

The things you can do in a web browser now are mind-boggling, thanks to how far WebAssembly has come. In this video I look at a few examples of some of the more powerful things that can be done: running the Postgres database as a WebAssembly app, running Jupyter notebooks with LLVM-powered JIT compiling for wicked fast computations, and deploying Rust apps to run in the browser without major rewrites. The Numba notebook demo can be found at https://notebook.link/@anutosh491/numba-ecosystem

Ep.274

Sep 3, 2026 6 mins

WebAssembly

Ep.273

Star Citizen: The Worst Case of Scope Creep Ever?

The ongoing saga of Star Citizen is an object lesson in why big software projects need constraints and realistic goals. If this isn't the worst example of "scope creep" in the history of IT, it's easily in the top five. Worst of all: there's no end in sight. What went wrong, and what can we learn from this fiasco? (And would you be shocked to learn an IBM veteran warned against all this forty-plus years ago?)

Ep.273

Aug 31, 2026 9 mins

Python

Ep.272

The 10 biggest new things in Python 3.15

Python 3.15's first release candidate is out, and is worth trying with your own programs. Here's the 10 biggest things you'll want to know about and get your hands dirty with. More on the "lazy imports" feature: https://www.infoworld.com/article/4145854/speed-boost-your-python-programs-with-new-lazy-imports.htm... More on the "frozendict" type: https://www.infoworld.com/article/4152654/get-started-with-pythons-new-frozendict-type.html All of what's new in Python 3.15: https://www.infoworld.com/article/4166693/the-best-new-features-in-python-3-15.html

Ep.272

Aug 19, 2026 7 mins

Python

Ep.271

AI trends that need more attention

Two things are happening in the AI world that deserve more discussion: the move towards decoupling models from workflows, and better work to make models smaller -- both to run on local hardware and in the big clouds.

Ep.271

Aug 4, 2026 5 mins

Python

Ep.270

Who's leaving GitHub and why

A few high-profile developers, and their projects, are abandoning GitHub for other code hosting sites or even their own self-hosted solutions. Why are they leaving? And what's it like to hop the fence and escape the GitHub (and Microsoft) monolith?

Ep.270

Jul 29, 2026 7 mins

Python

Ep.269

Typst, the programming language for documents

Languages like TeX can be used to create documents with elaborate formatting, like scientific papers or anything with math formulae. Typst aims to be as powerful, but by way of a modern programming style, opening the door for programmatic ways to create richly designed books, manuals, documentation, or even web pages.

Ep.269

Jul 23, 2026 7 mins

Python

Ep.268

UnigetUI: Package-manage all the things! (In Windows, at least)

Like Windows's native package manager, winget, but are not crazy about wrangling it through the command line? Want to handle all the other package management utilities on the same platform -- pip, Cargo, PowerShell's packages -- all in one place? This free and open utility gives you one UI to rule them all.

Ep.268

Jul 20, 2026 5 mins

Python

Ep.267

Sony's Killed Game Discs - What Will Be Killed In Enterprise IT?

Sony's plans to wind down disc production for its game consoles have met with fury (and memes) across the board. But this decision was driven by many of the same factors squeezing IT generally: spiraling hardware prices, weakening economies. And enterprise IT may be next to have a major rug pulled.

Ep.267

Jul 10, 2026 6 mins

Python

Ep.266

Why local AI's the way forward, and the best way period

With hardware prices spiraling, AI vendors ramping up token costs, and models becoming drastically slimmer and more economical, running AI models locally isn't just going to be a good idea whose time has come. It's fast becoming the most productive way forward for AI. Here's why.

Ep.266

Jul 2, 2026 3 mins

App Testing Python

Ep.265

How to Use Your Own Local AI in Visual Studio Code Chat

Visual Studio Code recently made it possible to use locally hosted AI, or AI from any endpoint, as part of its built-in chat system. It doesn't take long to get it set up, and while it has some limitations, you can get a lot out of it right away

Ep.265

Jul 2, 2026 5 mins

Python

Ep.264

4 things that need to change about gen-AI

Generative AI isn't going away, that's for sure. But that doesn't mean we have to live with the most dangerous and problematic aspects of gen-AI. Here's four things that need to happen for generative AI to go from potential menace to manageable.

Ep.264

Jun 23, 2026 4 mins

Python

Ep.263

Gemma 4 runs general-purpose AI locally (and quickly)

Google's Gemma 4 model promises new architectural improvements to process images, video, and audio faster, and to deliver quicker responses. It also comes in a variety of quantizations that allow it to run well on commodity PC hardware. This video runs through some example tasks like code generation, tool use, image captioning, and code analysis. For more details on Gemma 4's new "prediction drafter" architecture: Multi-token-prediction in Gemma 4

Ep.263

Jun 18, 2026 5 mins

Python

Ep.262

A first look at Pyrefly 1.0

Pyrefly, the Python type checker and linter from Meta, has just dropped its first full 1.0 release. It's intended to be a solid alternative to existing type checking/linting tools like Mypy and Pylance, and it offers many of the same features as those tools plus some intriguing experimental features, like automatic type stub generation and type coverage reports.

Ep.262

Jun 9, 2026 5 mins

Python

Ep.261

A look at Mojo 1.0, a Python-like Rust contender

The Mojo language is finally available in a 1.0 release. It's billed as "write like Python, run like C++", and it offers compatibility with the Python ecosystem along with many memory safety metaphors borrowed freely from Rust. Here's what it looks like to work with a simple program written in Mojo, that also interoperates with the Python world.

Ep.261

Jun 2, 2026 5 mins

Python

Ep.260

The AI rewrite of Bun in Rust is making shock waves

The alternative JavaScript runtime Bun, originally written in Zig, got an AI-assisted rewrite in Rust in the past week. That's startling enough by itself, but the way it was rolled out unilaterally by Bun's maintainers has left many people questioning the future of the project — me included.

Ep.260

May 26, 2026 4 mins

Python

Ep.259

Why developers are so divided about using AI tools

AI code tools have a divisive, even polarizing, effect on developers. Some feel empowered and rejuvenated by such tools; others feel disgusted and drained. What explains such violently divergent positions? Here's five things we could ask developers to better understand their reactions.

Ep.259

May 19, 2026 6 mins

Python

Ep.258

Profile Python Functions With ZERO Extra Code

Ever wanted performance stats for just one function, or a few of them compared against each other? The third-party tprof library uses newer (3.12 and higher) Python profiling technology to garner runtime statistics about functions — and doesn't need you to edit your code to do it.

Ep.258

May 12, 2026 4 mins

Python

Show me more

AI could boost software engineer productivity by 32.6%

By Maxwell Cooter

Oct 2, 20262 mins

Artificial IntelligenceDeveloperRoles

Microsoft, Google back Apache Ossie to make enterprise data and AI platforms more interoperable

By Anirban Ghoshal

Oct 1, 20266 mins

Data GovernanceData IntegrationData Management

Google makes Gemini 4 AI model available to a trusted few

By Nidhi Singal

Oct 1, 20265 mins

Artificial IntelligenceSecurity

What AI really is doing to software development (and developers)

Sep 29, 20265 mins

Did OpenAI STEAL Math Researchers' Work? The AI Controversy Explained

Sep 16, 20266 mins

Why Software Developers Are “Token Maxxing” AI—and What It Means for Jobs

Sep 11, 20267 mins