By now, every dev (and business leader) feels some type of way about AI. How could we not?

Every time you open the news, listen to a podcast, or visit your favorite social media site, you're hearing something about AI. If you've kept up, you might be burning out; if you haven't, you might be scared to dip your toe in, even if it's now feeling like a requirement to do your job.

Let's sit down and take a peaceful, friendly-for-all-skill-levels look at LLMs as a tool in software development workflows.

In this five-part series, we’re going to examine AI as a tool, as it exists when this series was written. We’ll revisit some of the fundamentals of LLMs, so that we can (1) stress-test our best practices for incorporating AI into our workflows, and (2) be better equipped to address questions and diagnose problems as the technology evolves:

What are the capabilities and limitations of LLMs, present and future? What are the benefits and risks of using AI in our workflows, short-term and long-term? If these are the sort of questions that’ve been bugging you, but you’ve been too skittish to ask, you’ve come to the right place.

Who is this series for?

• Senior developers who’ve mostly dodged the AI hype (but now feel pressure to be informed)
• Junior-to-mid-level devs who’ve used AI a bit (but haven’t thought much about the "but why?" and "and how?” of it all)
• Business leaders who want a pitch on AI that’s certified 100% Snake Oil-Free
• Tech-adjacent non-devs who want a grasp of the basics of AI without needing a Computer Science degree to do so

What will we learn in this series?

• How to understand and talk in clear, simple terms about AI with teammates, clients, and bosses
• How to be pragmatic in any conversation about AI (which, increasingly, feels like a superpower)
• How to separate AI hype/hate from AI reality
• Best practices for incorporating AI into our app dev workflow

Why does this matter?

It’s so easy to get caught up in the latest buzz and look past the present reality of anything that’s surrounded with this much hype. And it’s also easy to stick our heads in the sand. But what is best for us—as individuals, as people navigating our careers, as members of the global community—is to be clear-eyed and level-headed about AI and any other such tools. We want to preserve our craft—and our humanity—while also expanding our capabilities.

It’s how we become the sort of people who cut through the noise with clarity, confidence, and calm.

Note: This series is an examination of how we at Tighten understand and use LLMs in our workflow as an expert consultancy. It is not a series of tips, tools, or techniques for integrating AI into web apps. But we also know plenty about that! Stay tuned for future posts about incorporating AI into your app.

Overview

1. A Pragmatic Look at AI and LLMs in Software Development Workflows

• Preface: On AI Doomerism, Boomerism, and Pragmatism

2. What Even Is "AI"? Defining Key Terms in Plain Language

• What Is “AI”?
• Practical AI vs. Theoretical AI
• What’s a Large Language Model (LLM)?

3. Why Developers Should—and Shouldn’t—Use LLMs in Our Development

• Why devs have to choose how to use AI
• Why devs should use LLMs
• Why devs shouldn’t use LLMs

4. How to Expertly Use LLMs in Development Workflows

• Playing with AI vs. Working with AI
• Agentic Coding: How to Expertly Use LLMs in Your Dev Workflow
• A Process for Agentic Coding with LLMs
• Other Use Cases for AI in Software Development

5. How Will LLMs Transform Us? AI as a Tool in the Future of Development

We have a lot to tackle here. But before we begin, let’s acknowledge what we’re not doing here.

Preface: On Doomerism, Boomerism, and Pragmatism

The AI discourse is often like a cable news debate.

One pundit—the Doomer—fumes over the ways AI ruins, well, everything. We’re only a few years into this technological revolution, but already we’re reckoning with the very legitimate negative consequences of AI; for example:

• Energy consumption and electric grid strain
• Ethical violations over copyrighted training data
• Environmental devastation, especially in disadvantaged communities
• Economic bubble/debt spiral risk
• Job obsolescence narratives and/or AI-justified mass layoffs
• Productivity gap (hype vs. reality)
• Erosion of critical thinking in students and piling of new burdens onto already-overburdened teachers

The other pundit—the Boomer—evangelizes the early wins that make an AI techno-utopian future seem as desirable as it is inevitable. They steer the discourse towards the incontestable good AI has delivered:

• Earlier, more accurate cancer detection
• Zero-cost detection of Alzheimer’s and other dementias
• Smart buildings cutting energy use
• Accessibility tools for differently abled people
• Inclusive education for neurodivergent learners
• Rhino poaching reductions
• Faster, greener, more accurate weather forecasting

These two both seem to start with facts. But, like anyone with an extreme view, they’re often operating with as much baked-in preconception as they are fact-based decision-making. And as they spar over AI’s wins and losses and try to declare themselves the winner, they increasingly grow annoyed with each other, until their analysis curdles into insult.

We’re not here for spicy takes. The world, in general, needs less knee-jerk rhetoric, and more measured deliberation.

This matters even more in these early days of AI adoption: No one knows exactly how it will all unfold because the future isn’t here. AI is. As developers, we can’t go all-in or spurn any emergent computer technology—let alone a generational one—without understanding the fundamentals. And we absolutely can’t look the other way.

Thomas Dohmke, former CEO of GitHub, wrote, “For these developers at the forefront, making AI a core part of their work is not a distant, long-horizon future, but a change that is happening today.” That was in August 2025. But it already feels ancient. Early on in 2026, AI-powered development is practically all anyone in the dev world is talking about.

As a wise figure who also found himself alive during a world-altering technological revolution once said: "All we have to decide is what to do with the time that is given to us." We hear the futurecasting of Doomers and Boomers. But respectfully: We’re hitting mute. For now.

Ok. Let’s dive in: We begin with the most urgent/absurd question of the day: What Even Is AI?.