In this month's issue, you'll look at how soft skills help you get better results from coding assistants and how you can help people find your content by bringing it to them.

Help People Find You By Bringing Content To Them

You're creating great content and following SEO best practices, but your traffic is dropping. Something fundamental has shifted: AI summaries are reducing blog traffic across the board. In fact, a recent study found that AI summaries are causing a "devastating drop" in online news audiences. When people search for information, Google's AI summary gives them what they need. They don't see a link to your carefully crafted blog post.

Your content strategy needs to evolve. Instead of hoping people will find your content on your site, you need to take your content to where people already are. This doesn't mean you have to abandon your blog or overhaul your SEO strategy. LLMs like ChatGPT and Claude still rely on traditional web crawling and search infrastructure to discover content, so your existing well-optimized blog posts remain valuable. You're going to expand your reach and build on your solid foundation.

Find where your audience actually spends time and meet them there

You have to meet people where they are.

The most effective Developer Relations teams figured out years ago that they have to go where the community is. They don't sit in their company headquarters hoping developers will stumble across their code and blog posts. They pack up their demos and get on planes to developer conferences. They contribute to open source projects where developers actually work. They show up authentically in existing communities. That's what you need to do with your content.

You've likely identified your target audience and their needs, but now it's time to map where they gather outside of Google searches. Are they active in industry-specific LinkedIn groups? Do they hang out in particular subreddits? Are they watching YouTube to solve problems? Do they participate in Slack communities or Discord servers?

Once you know where your audience gathers, start building a presence there. First, extend the life of content you've already created. Turn that comprehensive guide into a LinkedIn post series. Break down your case study into a YouTube video. Share key insights from your research in relevant Reddit threads. This approach helps you get backlinks while reaching people who might never find your blog. And it's working. In a recent episode of the Talking Too Loud podcast, Josh Bliskel from ProFound shared data showing that Reddit citations in ChatGPT have increased 400% in just the past few weeks. AI systems are actively seeking out authentic community discussions to inform their responses. You need to be there with your content.

But don't just repurpose your existing content. Create first-run content specifically for these platforms. Write LinkedIn posts that feel native to the platform rather than just linking back to your blog. Create YouTube videos that take advantage of YouTube's growing search traffic. Start genuine conversations in Slack communities with questions and insights. Those conversations can help you develop more relevant content you can then repurpose.

This approach is especially important if you're the only one driving content at your company. You need to create a network of content that works together to build recognition for your brand.

Include calls to action that build awareness and track results

When creating content on these platforms, include clear yet subtle calls to action. Instead of "buy our product," try "learn more about this in our documentation" or "we dive deeper into this on our blog." You're building awareness and creating multiple touchpoints that reinforce your brand's expertise. You're not making hard sales pitches; you're helping search engines and LLMs understand your relevance.

Focus on engagement that shows genuine interest: people clicking through to learn more, signing up for your newsletter, or downloading your resources. This isn't about vanity metrics like followers or views. You want to see if this expanded approach is actually creating the kind of awareness that leads to business results.

Start small to avoid getting overwhelmed by all the data. Pick one or two platforms where your audience is most active and focus your tracking efforts there first. Set up UTM parameters so you can see which LinkedIn posts drive the most traffic to your blog. Monitor which YouTube videos generate the most documentation visits.

Check your metrics weekly or monthly rather than daily. Tools like Common Room can help you track community activity across platforms, but you can also start by manually monitoring the communities you're active in before investing in specialized tools. Once you have a handle on what's working on those platforms, you can expand to others. The goal is sustainable growth, not perfect measurement across every possible channel.

Building relationships and establishing credibility in communities doesn't happen overnight. You're creating a network of touchpoints that reinforce your brand's expertise and authority over time. But your audience is out there having conversations about the problems your product solves. Make sure you're showing up in those conversations.

Things To Explore

• This Dot Labs maintains a list of Tech Slack communities. Find your tribe.
• kew is a music player for your terminal.

"Soft Skills" Make or Break AI-Assisted Coding

If you're following discourse on LinkedIn, you'll see two narratives about AI coding assistants. Some engineers are getting great results with tools like Claude Code, GitHub Copilot, and Cursor. They're shipping features faster than ever and tackling problems they've never worked on before. But other equally talented developers are struggling. They get mediocre suggestions, spend more time debugging AI-generated code than writing their own, and wonder what all the fuss is about.

The difference isn't technical skill. It's not about knowing more algorithms or having memorized more syntax. The engineers succeeding with AI assistants share something else: they're good at the "soft skills" our industry has spent years downplaying. But successful software developers have always known that writing code isn't the most important part of the job; it's knowing what code to write.

Two approaches to programming

Most developers lean toward one of two approaches when tackling a new problem. Some get an idea and head straight to the keyboard. They start coding immediately, iterating and refining until something works. They're often brilliant problem-solvers who can debug complex issues and build impressive systems. But ask them to explain how their code works to a colleague, and they might struggle. The documentation comes later, if at all.

Other developers lean toward planning first. They sketch out the system design, think through edge cases, and can explain their approach before writing a single line of code. They might take longer to start coding, but when they do, they usually know exactly where they're headed.

Both approaches can produce great software, and most developers use elements of each depending on the situation. But when it comes to working with AI assistants, one approach has a clear advantage.

AI coding tools work best when you can clearly articulate what you want to build. They need context about your system, specific requirements for the feature, and details about how different components should interact. They're able to implement a lot of features for you, but only if you can explain those features clearly.

The "straight to code" programmers often hit a wall here. They know what they want to build, but they're used to figuring out the details through trial and error. When they ask an AI assistant for help, their prompts are vague. The AI does its best, but without clear requirements, the results are a bit off. Sometimes they're laughably bad.

Meanwhile, the planning-oriented developers are having a completely different experience. They're used to thinking through requirements and breaking down complex problems. When they prompt an AI assistant, they provide context, specify edge cases, and explain how the new code should integrate with existing systems. The AI has what it needs to generate exactly what the developer wants.

Communication skills suddenly matter

This shift highlights something the industry has been slow to recognize: writing code was never the hardest part of software development. Writing the right code was. And that requires understanding requirements, designing systems, and communicating clearly about complex problems.

AI assistants excel at the mechanical aspects of coding. They can generate boilerplate, implement algorithms, and even debug syntax errors. But they can't think, and they can't read your mind. They need you to be explicit about what you want, just like any other collaborator would.

The same skills that make someone good at writing PRDs, technical documentation, or design specs are the skills that make someone effective with AI coding tools. This is why product managers are having success with AI coding assistants when creating prototypes. They're already used to writing detailed requirements documents and breaking down complex features into specific, testable pieces. When they work with AI assistants, they naturally provide the kind of detailed context that leads to good results.

For years, the tech industry has treated these communication and planning skills as nice-to-haves. Job descriptions focused on technical requirements like algorithms and data structures. Code challenges tested implementation skills, not the ability to gather requirements or design systems. The "brilliant but introverted programmer who can't communicate" stereotype became a badge of honor rather than a limitation. Companies reinforced this stereotype by creating distinct tiers: "technical" roles for those who wrote code, and separate positions for marketers, technical writing, and other roles. The implicit message was clear about which skills the industry valued most.

AI is flipping this dynamic. The mechanical coding skills that we've prioritized are exactly what AI assistants are best at. The planning, communication, and systems thinking skills we've treated as secondary are now the differentiators. These aren't "soft skills" at all. They're core engineering skills that our industry has systematically undervalued. They're the "human element" that AI can't replace.

It's not prompt engineering, it's communicating well

If you've always been more of a "straight to code" developer, this shift might feel frustrating. But the good news is that you can develop these skills with practice, and it's worth doing because you'll get better results with the robots and with other humans.

Begin by adjusting your approach to new features and bug fixes. Before opening your editor, spend five minutes writing down what you're trying to accomplish. What should the end result look like? What edge cases might come up? How does this change fit into the broader system? Write these things down. That's the first draft of a spec.

The same thing goes when someone asks for a new feature. Get curious. What specific problem do they want to solve? What would success look like? What constraints do you need to work within?

Getting really clear on your requirements will help you write better prompts. Instead of asking for "a function to handle user input," you ask for "a function that validates email addresses, handles both Gmail and corporate domains, and returns specific error messages for common typos." The more you practice translating vague requests into specific requirements, the better you'll become at working with AI tools. The specificity makes all the difference in the quality of code you get back.

Practice explaining your code to others. Start writing brief comments that explain not just what your code does, but why you made specific choices. When you can clearly articulate your reasoning to a human colleague, you'll be able to provide that same clarity to an AI assistant.

Get comfortable with iterating on your requests when working with AI assistants. This is the same skill you already use when building software: you create something, get feedback that it's not quite right, and adjust your approach. When the generated code misses the mark, look at what went wrong and refine your prompt. Maybe you didn't specify error handling, weren't clear about the data format, or forgot to mention a key constraint. This back-and-forth teaches you to communicate more precisely about technical requirements.

As you refine these skills, you'll get better and better results from AI coding tools because you're providing the clear context and specific requirements they need. And this makes you a stronger developer overall because you'll also communicate more effectively with your colleagues, write better documentation, and design more robust systems.

So-called"Soft skills" are fundamental engineering competencies that make you better at every aspect of your job.

Parting Thoughts

Here are a couple of things to think about over the next month:

1. Has working with coding assistants changed how you approach building software? What have you noticed, and why do you think that is?
2. Take two pieces of content you've already written. One should be a piece that performs well, and another should underperform. Identify places where you can syndicate that content. Medium or Dev.to are good places for syndication because they can use canonical URLs to point back to your original posts. After a month, look at inbound traffic to your site. Did you get referrals? Did you start seeing increased traffic?

As always, thank you for reading.

I'd love to talk with you about this newsletter on Mastodon, Twitter, or LinkedIn. Let's connect!

Please support this newsletter and my work by encouraging others to subscribe or by buying a friend a copy of Exercises for Programmers, Small, Sharp Software Tools, or any of my other books.