lesterbotello.dev

AI is not replacing you, it's promoting you

Sun, 08 Mar 2026 00:00:00 -0500

If you’ve been following the recent labor market impact assessments—specifically the latest research from Anthropic—the numbers look startling at first glance. Computer programmers are listed at the very top of the “exposed” list, with a staggering 75% of our tasks identified as being within the reach of AI automation.

But as someone who has navigated the industry for over two decades, I can’t shake the feeling that I’ve kind of seen this movie before. Long before I started coding, assembler experts were alarmed with the rise of COBOL and Fortran. Flashforward a few decades and the rise of IDEs were the new cause of panic, followed by the “low-code” movement, and the offshore outsourcing waves a few years later. Each time, the narrative is the same: The end of the software development career is near.

Beating AI Choice Paralysis - How to Choose the Best AI Model for Your Needs

Fri, 19 Dec 2025 00:00:00 -0500

If you’ve taken a look at the current landscape of Artificial Intelligence lately, you’ve probably felt it. You have a compelling use case, you know what problem to solve and you know you can use AI to solve it. But then, it hits you: “Should I use GPT-5.2? Wouldn’t it be too expensive? What about Gemini? OK but WHICH Gemini? Flash, Pro? Which version? Or should I probably go with something better trained for my specific needs?”

Setting up the new Uno Platform's MCP Servers

Sat, 29 Nov 2025 22:00:00 -0500

As I write this, .NET Conf is just behind us. As it’s mandatory every year, there is an endless string of announcements around .NET 10, C# 14, AI building blocks, and a revamped and rebranded Aspire, among many other topics. I’ll be covering some of that in the next few articles, but in today’s post, I’d like to take a closer look at a different announcement from the Uno Platform team: Uno Platform 6.4. This release introduces (among many other things) two new MCP Servers, which promise to make it a lot easier to create awesome stuff by pairing Uno Platform with your favorite agentic AI tool.

Make your AI prompts collaborative with VS Code and Promptitude

Tue, 04 Nov 2025 09:26:00 -0500

If you’re working in a project in active development, you’re probably doing a lot of tasks that you’ve done before and/or will likely be doing again in the future. For example, you might need to add functionality to your app to call a given REST API, deserialize its JSON response to create a DTO and then transform it into a domain object. You’ll likely need to do the exact same process everytime you need to connect to a new endpoint, since the endpoint, the returned data and its purpose itself will be different, but the implementation steps will remain unchanged. Interestingly, this particular task is very common in software development in general, and one could argue that automating this process would be beneficial for any project of the same type.

An MCP Server written in C#? Yes, sir!

Wed, 22 Oct 2025 10:00:00 -0500

If you’re a C# developer chances are you already have the tools you need to build an MCP server from scratch. No, really. If you’re a C# dev, it’s likely that you already have VS Code and .NET installed in your system. This is really enough. In this article I’ll walk you through creating your own MCP server using just that.

What you’ll need

.NET 9.0 SDK or higher..
Visual Studio Code is recommended, but you can also use any code editor or .NET IDE you want.

Preparing the project.

For this example, we’ll build a super simple MCP server that we’ll use to convert between currencies using fixer.io. It’s a free, simple and lightweight API that provides current and historical exchange rates between any currencies.

Yes, your AI agent can be a .NET cross-platform expert

Sun, 17 Aug 2025 10:00:00 -0500

Today I’d like to show you how to make your AI agent better at ANY framework or programming language you choose for your project and to get the optimal results by leveraging the tools these AI agents put at our disposal.

Introduction

It’s no doubt that coding is one of the primary targets of leading AI models. Recent iterations of known models such as GPT-5 are significantly more focused on helping developers deliver quality code. However, the knowledge of these models seem to be better when the platform of choice is less complex and requires less “homework”. But what happens when you NEED to take advantage of a powerful but complex cross-platform framework in which a little bit of documentation can make everything easier?