Welcome to WebDev of Reality!

That is what Adrian Toomes says to the Tinkerer in Spider-Man: Homecoming before he decides to become The Vulture. That quote has more weight to this whole WebDev of Reality than I first realized.

1. Spider-Man is my favorite Hero — have been a comic book nerd since watching the 90s Spider-Man cartoon. Of course I would reach for a Spidey quote first. That quote by Toomes feels very relevant today in the fast changing world of AI.
2. I love to Tinker — fitting Phineas T. Mason A.K.A the Tinkerer would be in this scene. I have always been a tinkerer. I remember taking apart electronics and trying to build new things. When I was 9 I took a Sony Walkman, with a broken headphone jack, apart. Then, stripped the wire of a speaker from a “boombox” and soldiered it to the Walkman. I had myself my own portable speaker before Bluetooth was a thing!
3. Spider-verse OS — for me to truly understand something I need to break it down to the smallest possible unit, then rebuild it back. To truly understand computers I am building my own Operating System called Spider-Verse OS. More on this in a future post!

“Who the hell is this guy, and what makes him an expert?!”

Short answer is I’m not! Definitely not an expert but I have been building software for over a decade. My name is Anthony Renteria, I am a Senior Software Engineer at Olo. Olo has been (what my computer science professor, Dr. Beard would call) a trial-by-fire 🔥. Olo gives restaurants the ability to provide their customers with online ordering. Before DoorDash, Ubereats and even the iPhone, there was text-based mobile ordering by Olo!

“What could be so complicated about ordering food” is what I thought when was first going through the technical interview rounds for Olo. 7 years later know my answers. Everything. I have learned so much from the being in the 🔥. Knowledge absorption from some of the most brilliant engineers I have ever met. I have also learned so much from doing things the wrong way. Can you really be an “expert” if you haven’t broke ordering in production for 1000s of customers? 😬

As mentioned earlier I am a certified Tinkerer. AI has reinvigorated that love for technology and innovation. Over the last year and a half I have been working with AI to figure out how to best utilize it for my work. I was a skeptic in those earlier days. “It’s just a bullshit generator” but then again so am I. It was once I started experimenting with workflows and agent orchestration, I saw real benefits. I created a workflow called “The council of Elrond” (yes I name most of my tools after pop culture).

AI agent Elrond’s job was to analyze the “dossier” I prepared for him, determine from the pool of subagents we had built, who was best suited to investigate the task. The council would convene and investigate. Each council member investigated from a different angle based on its agentic directives. Then they would come together, discuss findings, and create a plan based on what they found. The council would have disagreements and bring the other councils members in to help resolve disagreements. The final report would give list of where the council agreed, also where they disagreed and why. I used this for a spike (technical investigation) ticket. It came back with a very detailed report in under 10 minutes. The report noted a disagreement with the docs agent (whose knowledge was rooted in Olo documentation) and the back-ended agent (rooted in the Olo codebase). Documents said one thing but the actually code behaved differently. Something I would not have caught, that could have left the implementer of the ticket spinning their wheels because I made an assumption.

“Isn’t AI killing our planet and making us dumber”

Yes, probably. I believe it doesn’t have to, though. Part of this blog is to help remove the unknowns around AI. I want to open up conversation on how do we navigate AI in an ethical way. Our billionaire overloads seem to think data centers are their way to cash in on the AI hype. Very unimaginative… models keep getting more efficient with a smaller environmental footprint. My prediction is the future of AI will be on local modals on your local hardware. Smaller models are getting more efficient and smarter. With the right harness and ai tooling, smaller models will be usable for 90% of AI requests.

This blog and all of WebDev of Reality will have AI directives baked in. The repo for this blog is public. Take a look at the agent.md file. This instructs any AI working in this space, the directives it must follow

1. Transparency: label and sign all AI content
2. No AI-generated art — ever
3. Human authorship is the default

This blog should also serve as a resource for newer engineer who are trying to navigate the hellscape that is the tech career field. Mass layoffs, AI replacing everyone, everyone being hired back because AI is too expensive. I’ve noticed engineers complaining that they don’t want to be AI babysitters. That because coding is essentially “solved” there is no need for engineers anymore. Good News, you get to skip the code monkey phase of becoming a software engineer!

My career followed this path.

1. Junior developer — Code monkey. Get it to work by any means necessary. Spaghetti code and all.
2. Mid engineer — Start planning implementation. Clean, simple, maintainable Architecture.
3. Senior — Deep in planning and architecting. Up and down the entire stack. Higher level planning work that can easily be implemented by multiple engineers in Parallel.

Domain knowledge is your super power. AI is good, real good at coding. Maybe that does mean the junior developer who gets code working by any means necessary, is becoming obsolete. Someone technical still needs to be in the build loop. Needed to interpret the product requirements and convert them into technical directions. Your understanding of the business from a technical angle, will take you farther, than your ability to solve any interview question with a hashmap. When in doubt hashmap it out

Stick around, let’s learn and break some shit together!