The Modern AI Work Stack

The conversation around artificial intelligence and income has shifted. A couple of years ago, the focus was on experimentation seeing what the models could do, testing boundaries for the sake of it. Now, the landscape is more settled. The tools are more capable, and the ways people integrate them into their work have become more structured. It is less about the novelty of the technology and more about the repeatability of the systems built around it.

There are two primary approaches emerging. One is skill-based, where an individual uses AI to enhance an existing craft writing, design, analysis to take on more complex projects or deliver work faster. The other is system-based, where the focus is on building automated workflows that handle tasks from start to finish with minimal direct involvement. Neither is inherently better; they suit different temperaments and goals. The skill based path leans into the human-as-artisan model, using AI as a sophisticated tool. The system-based path treats the AI as a semi-autonomous agent, with the human acting as an architect and quality control.

Content creation remains a viable avenue, but the workflow has changed. A writer might use language models to generate research summaries, outline structures, or even draft variations of a headline. The value is no longer in the raw act of typing, but in the direction, the editing, and the unique angle that only comes from lived experience. Similarly, in visual work, AI tools handle the rendering, but the human still holds the creative direction, the composition choices, and the narrative intent behind the image.

Automation is a different lane. Service based workflows can be partially or fully automated for clients. Think of a system that monitors data sources, generates a summary report, formats it into a presentation, and emails it to stakeholders. The human role is in setting up the initial architecture, ensuring the data connections are secure, and handling the exceptions that the automated system flags. This creates a service that runs on a retainer model rather than a per-hour model, changing the nature of the income entirely.

A common structural gap appears when people treat AI as a single, monolithic tool. In practice, effective work involves a stack of different models and interfaces. A language model might be best for brainstorming, another for structured data extraction, and a third for summarization. Learning which tool fits which part of a process is a form of fluency that develops over time. It is not about mastering one interface, but about understanding the capabilities of different underlying models and how to route tasks between them.

Short-term tactics often involve exploiting a temporary gap a new model release with specific capabilities, or a platform that hasn't yet restricted certain uses. These can yield quick results but are unstable. Long-term sustainability comes from building around the capabilities that are likely to persist: the ability to process language, to summarize information, to translate between formats. These are the durable foundations. The specific interfaces and models will change, but the underlying utility of being able to manipulate language and data programmatically is not going away.

The infrastructure for working with AI is also fragmenting in interesting ways. The comparison between tools like Google Antigravity and Cursor highlights different philosophies. Antigravity represents a more ambient, integrated approach AI assistance woven into the fabric of existing Google Workspace tools. Cursor, on the other hand, is a dedicated environment built from the ground up for AI interaction, specifically for code. The choice between them isn't about which is better, but about workflow fit. For someone embedded in document-centric work, the ambient integration might be seamless. For a builder focused on software, the dedicated environment offers more control.

In service-based work, the effective workflows often look like assembly lines. A raw input comes in, passes through a series of AI-powered stations cleaning, analysis, formatting, summarization—and a human inspects the output at key checkpoints. This hybrid model scales because the AI handles the volume, and the human handles the judgment. The income generated is a reflection of that judgment, not the throughput.

The mistakes tend to cluster around the same points: assuming the AI understands context without being explicitly told, treating its output as finished work, and failing to build in feedback loops where the system improves over time based on corrections. The people who integrate these tools into sustainable income streams are the ones who treat the interaction as a conversation, not a command. They refine, they iterate, and they maintain a clear sense of what they are trying to build.

The current moment feels like a settling period. The rapid release cycles of 2023 and 2024 have given way to more measured, applied iterations. The models are getting smarter in ways that are less about flashy demos and more about consistent, reliable performance on narrow tasks. For someone looking to understand this space, the focus should be on workflows, not features. The feature set of a model matters less than how it fits into a sequence of actions that leads to a finished outcome.

The OpenAI work in health, the prompt packs circulating in specialized communities, the different interaction modes in Gemini, the competing philosophies of integration versus dedicated environments in tools like Antigravity and Cursor these are all pieces of a larger puzzle. They are not announcements of a new era. They are incremental adjustments to a landscape that is already here. The question is not what the technology can do, but what someone wants to build with it.

For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.

LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.

The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.

Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

Learn more.