Software Engineering is characterized by Feedback Loops. The faster the Feedback Loop, the faster the Dev Velocity. (Wouldn't that be 'Eng Velocity' for Engineers? I guess Dev Velocity is catchier..)

Who do you think develops a product faster: Engineer 1, or Engineer 2?

Fast Feedback is a 'Law' of Software Engineering, similar to what Gravity is to Physics. The shorter the fall, the less it hurts.

To define the terms in the image above:

1. Engineer: Represents a Software Engineer working on the product.
2. Feedback Mechanism: Represents the mechanism by which that Software Engineer gets feedback
3. Feedback Time: How long it takes the Software Engineer to receive feedback from a Change they have made.

Techniques that Help Reduce Feedback Time

Continuous Integration

Continuous integration provides fast feedback on the integration process, which used to be colloquially referred to as 'Integration Hell'.

Back in the day, you were merging massive changes together from multiple sources simultaneously and it was painful.

Continuous Deployment

Continuous deployment provides fast feedback from a production environment. Known as 'Prod' to Software Engineers, feedback from actual users is invaluable.

There are ways to fail fast in Prod without affecting the majority of your users, such as Canary Test- worth an investigation to manage thoughtfully for your use case.

Take a moment to think about what you could learn if your product was in front of a user today versus the pain of learning that same lesson a year from now.

Dave Farley is a great source on CI/CD.

Test Driven Development (TDD)

The image above is a big reason why teams of any size benefit from certain types of testing.

Even if you feel like TDD itself isn't for you, thinking about how automated regression testing might save your QA team time and get your dev team feedback immediately instead of days later.

If you feel like checking out TDD, Martin Fowler has an article on it.

Conclusion: Remember the Principle, not the Techniques

I purposely don't go into huge detail here because the important part is understanding the physics of Fast Feedback, not the techniques.

The techniques help, but as long as you remember the principle you can research and critically evaluate which techniques work best for your situation.

I love mixing an old parchment and ink style with a circuitry style.

Sage's Bookshelf

This is the grab bag of cool things. I felt like 📖 emojis are on-brand.

📖 AI News: Amazon is heavily investing in Anthropic (Notable if you work with AWS)

📖 Software Engineering: The Gift of Constraints (Bane of Scope Creep, Bringer of Completed Projects, Defeater of Delayed Deployments!)

📖 ERP Tools: Odoo Raised $527M, Valued at $5.26B (Odoo has a suite of highly useful tools for all kinds of businesses - think Email Marketing, Website, Forum, Invoicing, etc. Between these jawdropping numbers and their 40% growth rate, it's likely you don't have to worry about them disappearing anytime soon.)

📖 Rosenblatt AI: A Great Article from Ryan on Benchmarking AI Models (It is always important to fact check the numbers folks put out- thanks for the insights!)

📖 State of AI Agents (LangChain): 2024 AI Agent Statistics from LangChain (Cursor seems like a favorite AI IDE for Devs, and Safety mechanisms such as Human-in-the-Loop are mentioned)

📖 RAG: A Guide/Insights on Evaluating Retrieval Augmented Generation (RAG) Systems (I always appreciate folks sharing lessons learned on new technologies)

📖 Startup Process: My Article on Business Automation (Single source of truth, Knowledge Management, Critical Thinking, Slack has Automations, and more- I believe in relying on systems, not people. A system will be around when someone joins your company and it will be around when someone leaves your company.)

📖 State of AI in Enterprise (Menlo): 2024 State of AI in Enterprise (A 'Survey of 600 IT decision-makers regarding AI', has some interesting points.)

📖 Chat GPT API: Structured Outputs (This is one of those 'everyone struggles with making Chat GPT output a consistent syntax, but there's a tool that makes it far easier' situations.)

📖 Jaguar Ad: Copy Nothing (Is Jaguar's ad genius or a flop? I'm more interested in watching everyones reactions while I eat popcorn than getting 'upset' over the ad. Lots of people getting upset about it to the point that I'm hearing about it from an almost exclusively tech LinkedIn feed. If that isn't reach, I don't know what is.)

Building AI: Interested in help building AI & Software? Check out Rosenblatt AI via Website, or check out the Upwork. Let them know I sent you. :)

#startup #cto #ceo #fullstack #fullstackengineering #softwareengineering