Taro Logo

A Framework For Integrating AI Into Your Workflow

AI will not replace you. A person using AI will.

This tweet was seen 5M times because it captures a transformation that will happen in the workplace over the next 5 years.

This video covers a practical framework to apply AI tools like ChatGPT, Bard, or Github Copilot to become more productive in your job.

The goal for many of us is to become the mythical 10x engineer – they have 10x as much impact as their peers. These engineers do exist, but they don’t write 10x more code or fix 10x more bugs. Instead, they understand the concept of leverage, which manifests in two ways.

  • Take a challenging problem, and solve it in a 10x or 100x better way
  • Make 10 or 100 people around you more productive

Shreyas Doshi has a framework called LNO to identify where you should spend your effort:

  • Leverage (L) tasks are tasks that have huge consequences and you should do a great job.
  • Neutral (N) tasks are tasks that are important, but don't have a big payoff. You should do a good job.
  • Overhead (O) tasks are tasks that are a tax on your time. You should just get it done.

Let's go through an example in each bucket:

In the leverage bucket is creating the design or architecture for the project. This is high leverage not only because architecture is inherently difficult to change, but it’s also an important trust-building exercise.

  • Ask AI to identify what could go wrong with your general approach and then add that as a discussion.
  • Ask AI to summarize documents related to what you’re trying to build
  • Ask AI to solve the problem and then sanity-check. The results here won’t always be good, but at least you should be able to explain why you’re choosing one path over the other.

Some coding tasks are in the neutral bucket. AI gives us the opportunity to dramatically speed up how long some coding tasks take.

  • Use tools like Copilot to write boilerplate code much faster
  • Instead of having to go to StackOverflow, you can get answers inline
  • Ask for things like regexes or how to do something in idiomatic Kotlin

Finally, there are tasks in the overhead category that AI can help minimize

  • The classic example is writing your performance review – you’re not actually having any business impact.
  • I recommend every engineer has a brag document of what they did throughout the half or year. Then you need to massage all your work into whatever template your company uses.

With the rise of AI, asking good questions is more important than ever. Artificial Intelligence cannot read your mind (at least not yet 😅), so the quality of your question dictates the quality of your response. Learn more about asking good questions below: