RedMonk February 2025 Update
A small, sharp research firm focusing on developer-led technology adoption and developer culture. We help folks understand the industry by understanding you.
Greetings from the AI mines. Kate here.
It may only be February, but 2025 is already shaping up to be one hell of a year. Many of us are grappling with family colds (possibly an exciting new Flu strain). Cross-country school closures for winter weather have affected us in Maine, Georgia, and Colorado. And we are all closely following the news and deeply concerned about ongoing political upheaval in the US.
While these issues are certainly top of mind, in the tech and AI space the biggest story is Deepseek. Stephen O’Grady’s “DeepSeek and the Enterprise” does a fantastic job contextualizing why everyone is talking about this Chinese lab and its low-cost, high-quality models. Now that the dust is beginning to settle, Deepseek’s sociopolitical, economic, and technological implications are also coming into focus. To begin, rumors of Deepseek’s astronomically low training costs have likely been overrepresented. Next, there are concerns around data privacy and bias owing to the Chinese government’s involvement. Moreover, intellectual property, which has long been a sore, and even litigious, point around training data, has now expanded to include accusations of distilling data from OpenAI’s models in violation of OpenAI’s terms of service.

This far from an exhaustive list points to what makes Deepseek so damn interesting. Whether this company and its models represent an existential threat to incumbents remains to be seen, but, as I’ve said repeatedly, following this story exemplifies what makes being an industry analyst so exciting.
Finally, hats off to Rachel Stephens, who was promoted to Research Director at RedMonk. Congrats Rachel!
That’s a wrap,
Kate Holterhoff
Links Roundup
You Need Data to Write a Fair Engineering Performance Review: I found this article about constructively using engineering metrics in performance reviews from the DORA community forum. Both the article and the discussion around it are thoughtful and worth your time.
If Even 0.001 Percent of an AI's Training Data Is Misinformation, the Whole Thing Becomes Compromised, Scientists Find: Did your provider answer your latest health question using the MyChart (an extremely popular and prevalent app used in medical settings in the US) platform? If so, you might want to reach out to them personally to double check on the information that was shared with you. According to this article, it only takes an infinitesimal amount, proportionally (0.001 percent) of misinformation to pollute an entire LLM and this has already had demonstrably negative effects in the form of a disproportionate amount of hallucinations
Can a headline alone be eligible for a Pulitzer? We’re all observing the shenanigans at Meta, but some of us are contemplating deleting our personal accounts.
Google engineers detailed how LLMs cut costs for large codebase migrations: A prime example of the kinds of use cases we should be looking for with AI.
Recent RedMonk Research
Wrote up my thoughts on GitHub npm and the JavaScript package registry startups vlt technology inc. vsr & Deno JSR. There's a lot going on around JS/TS, Node.js package management and registries—and a lot of history. Is npm Enough? Why Startups are Coming after this JavaScript Package Registry RM clients mentioned: GitHub, GitLab, and Microsoft
Stephen O’Grady offers a coherent and accessible summary and analysis of Deepseek, focusing particularly on how the models offered by this lab will impact the enterprise “DeepSeek and the Enterprise” RM clients mentioned: AWS, Google and Microsoft
Readers interested in the convergence of data platforms and AI, including a trip down memory lane through some Hadoop history, should check out Stephen O’Grady’s The Dream of Hadoop is Alive in AI RM clients mentioned: AWS, Google, IBM and Microsoft
Recent Videos and Media Appearances
Emilio Salvador, VP of DevRel & Strategy at GitLab, joined RedMonk’s Rachel Stephens to discuss current trends around agentic AI: how do agents differ from code assistants; how can AI come into play in the larger software development lifecycle (beyond writing code); where is the “human in the loop” in all of this?
Exclusively on the MonkCast

A RedMonk Conversation: Zack Chapple on Zephyr Cloud, Micro Frontends, and Module Federation. Zack Chapple, Co-Founder and CEO of Zephyr Cloud, discusses how micro frontends and module federation are empowering developers to work more efficiently. Zack shares insights from Zephyr Cloud’s recent Acceleration Week, the importance of open source, the role of AI in development, and how Zephyr Cloud is positioning itself to meet the challenges of the evolving tech landscape.
A RedMonk Conversation: Cassidy Williams Talks Companionable Coding & GitHub Copilot. Cassidy Williams, Senior Director of Developer Advocacy at GitHub, discusses GitHub Copilot’s role as a coding companion. She tackles skepticism about AI coding assistants among developers and engineering leadership and emphasizes GitHub’s commitment to the open source community and the accessibility of Copilot for all developers.
RedMonk Recommends
Jason Johns is a longtime product focused engineer with a track record of project leadership in multiple domains. Most recently at Coalition, Inc, Jason was a subject matter expert in several core components of highly asynchronous event driven workflows facilitating over $500MM in annual gross written premium. He also initiated and facilitated a backend-focused guild that met regularly to facilitate silo reduction and increase individual awareness and communication across disparate teams. Before that, he was at O’Reilly Media as a major contributor in search, metadata and search relevancy for the learning platform. Jason is always open to new concepts and opportunities for learning and growth, but areas of particular interest include search and distributed event driven architecture as well as observability, monitoring and enhancing engineering efficiency.
Jason is currently looking for full time remote senior to staff opportunities in Series B/C and higher as well as small/midsize companies in python or golang technical stacks. With his experience and expertise, he is well suited for an organization that has strong business vision with a need for technical leadership and desire for a high quality engineering division to facilitate business growth by leveraging best practices to ensure engineering efficiency.

Monki Gras
Monki Gras 2025 is just around the corner: March 27th/28th in London. This year's theme is Sustaining Software Development Craft. Early bird tickets are now sold out, but you can purchase mid-bird tickets online here.
If you're interested in sponsoring this year's Monki Gras, please contact Morgan Harris for a list of available opportunities.
Meet the Monks
Events we'll be attending:
AWS Developer Innovation Day: February 20, 2025, Hybrid: Seattle, WA / Virtual
ACM SIGCSE Technical Symposium 2025: February 26 - March 1 2025, Pittsburgh, PA
Devnexus 2025: 4 - 6 March 2025, Atlanta, GA
Salesforce TDX: 5 - 6 March 2025, San Francisco, CA
KubeCon + CloudNativeCon EU 2025: 1- 4 April 2025, London, UK
Google Cloud Next ’25: 9 - 11 April 2025, Las Vegas, NV
Events we'll be hosting:
Monki Gras 2025: 27 - 28 March 2025, London, UK
Our mailing address is:
RedMonk
411 Congress Street
Portland, ME 04101
Copyright (C) 2024 RedMonk. All rights reserved.