I regularly read Ed Sim's What's Hot in Enterprise IT. Not only does it keep me up-to-date with the latest tech/VC happenings, but I also frequently get ideas and experiments from Ed's weekly writing. This weekend, I was intrigued by the mention of Google's NotebookLM, a "personalized AI research assistant powered by Google's most capable model, Gemini 1.5 Pro." I was keen to experiment further.
I signed up for NotebookLM using my Google account and pointed the tool at a series of Syntasso blogs and webinar recordings. The results were epic—especially the auto-generated podcasts—and I say this as a regular user of ChatGPT and other LLM services.
Podcast # 1: Discussing platform engineering, platform decay, platform as a product, and more
Without further ado, I'll share the first podcast I generated based on our platform engineering blogs and webinars:
This is undoubtedly the best auto-generated audio stream I have heard. The hosts sound natural, the discussion is focused on the topic at hand, and most points appear correct.
Podcast # 2: Kratix 101 - Building composable internal developer platforms (IDPs)
My colleague, Shane Dowling, was also playing around with Notebook LM. He pointed the tool at our Kratix documentation and resources and generated this podcast:
Again, the results are incredible. Although the first podcast I generated discussed the relatively high-level topic of platform engineering, this second one focuses on a specific open source tool for building platforms.
Briefing Doc: Platform Engineering - Navigating the "Trough of Disillusionment"
Although I was particularly interested in podcast generation, NotebookLM also allowed me to create FAQs, study guides, and briefing docs. Here is an example of the latter:
This briefing document reviews common themes and key takeaways from various sources on platform engineering, focusing on the challenges organizations face as they move past the initial hype cycle.
Main Themes:
The "Trough of Disillusionment": Platform engineering, like many promising technologies, is entering a phase where initial enthusiasm gives way to the realization that it is not a silver bullet. Gartner predicts that while 80% of organizations aim to establish platform teams by 2026, many will struggle to demonstrate clear ROI, leading to disillusionment.
"The 'build it and they will come' mentality...is going to be [challenged]... there's going to be some challenges up and coming" - Mark O'Neil, Gartner (Platform Engineering Hype Cycle)
Brownfield Reality: The ideal of building a platform from scratch rarely exists. Organizations must grapple with integrating legacy systems, diverse team workflows, and existing technical debt. This necessitates a shift in mindset from "greenfield idealism" to "brownfield pragmatism".
"Most platforms are brownfield...the old and the new co-exist. [We need] to meet the existing complexities of the legacy enterprise and its infrastructure and systems" - Cat Morris (Turning Legacy to Leverage)
Platform as Product: A consistent theme is the need to treat internal platforms with the same product-focused mindset applied to external products. This means understanding developer needs, gathering feedback, iterating on features, and demonstrating tangible value.
"Your platform is not an off-the-shelf piece of software... it is an evolving set of reusable services...that creates valuable outcomes for your business. The platform’s capabilities should change in response to the needs of its users" - Colin Humphreys (Five Lessons Learned)
Measuring Success: Simply building a platform is insufficient. Organizations must define clear goals, establish relevant metrics (like deployment frequency, lead time, MTTR), and track the impact on developer productivity and business outcomes.
"Establish clear goals and then get a baseline of your metrics and look for improvement...leading indicators [like] adoption rates...lagging indicators [like] app retention rates, incident rate" - Daniel Bryant (Platform Engineering Hype Cycle)
Avoiding Platform Decay: Over time, platforms can become bloated, difficult to maintain, and misaligned with user needs. Organizations must proactively manage platform decay through continuous improvement, embracing commoditization, and effective fleet management.
"The better you get at migrations, the better you feel as a platform engineer...If you're not migrating you're stagnating." - Abby Bangser, (The #1 Platform Engineering Problem)
Key Takeaways:
Focus on Developer Needs: The primary users of internal platforms are developers. Understanding their pain points, removing roadblocks to productivity, and empowering them through self-service are critical success factors.
Codify and Share Best Practices: Platforms offer a powerful mechanism to codify organizational policies for security, compliance, and best practices. This ensures consistency, reduces cognitive load on developers, and simplifies governance.
Embrace Inner Sourcing: Encourage contributions from different teams to foster a sense of shared ownership and leverage domain expertise. This can accelerate platform adoption and ensure it meets diverse organizational needs.
Think Long-Term - Day 2 and Beyond: Design platforms for maintainability, scalability, and ease of evolution. This includes embracing automation, adopting fleet management practices, and planning for ongoing upgrades and migrations.
Leadership and Communication are Crucial: Securing executive buy-in, communicating the value proposition effectively, and fostering collaboration between platform teams and other stakeholders are essential for success.
Recommendations:
Conduct a thorough assessment of your organization’s existing systems, workflows, and developer needs before embarking on a platform initiative.
Start small, iterate, and demonstrate value quickly. Focus on solving a specific problem for a specific team and then build upon those successes.
Adopt a product mindset. Gather feedback, measure usage, track metrics, and iterate based on data to ensure the platform remains valuable and aligned with evolving needs.
Invest in platform engineering expertise. Building and managing successful platforms requires specialized skills. Hiring experienced platform engineers or partnering with external experts can accelerate your efforts.
Foster a culture of collaboration and transparency. Platform engineering thrives when teams communicate effectively, share knowledge, and work together towards common goals.
Conclusion: Thumbs up for NotebookLM
Google's NotebookLM looks like a very interesting tool! I could see myself using this as a learning aid, as I enjoy supplementing my reading with podcasts. Being able to curate the exact topics covered within a podcast is something that I hadn't given much thought to until now.
Comentários