Hey!
Summary of this email:
- Things got a little heated with the Event Bus in Database 😅 (+ published 100% on Codely Pro Standard)
- Is the way we code going to change in 2026?
- The joke of the year.
⏱️ Estimated reading time: Just a bit (things get spicy).
😅 Things got a little heated with the Event Bus in Database
A few weeks ago we published a post on LinkedIn discussing Why you shouldn't use RabbitMQ or Kafka for an Event Bus and it sparked quite an interesting debate.
Comments worth highlighting:
- 🔥 This setup only works if you have a single worker consuming.
- 🔐 Lock and deadlock management.
- ⏱️ If I'm going to spend hours implementing this solution, I'd rather spend them on the proper tool.
- 🔔 It may sound great, but when loads go past a certain threshold it doesn't work.
In these comments there are many valid points and also a lot of "it depends on the context."
We wanted to respond to them by explaining in which contexts it applies and in which it doesn't, and so we published a video addressing all these topics.
Also, we've just published the complete Event bus in database course on the Codely Pro Standard plan. Build a scalable event bus until you need to migrate to another system.
🤔 Is the way we code going to change in 2026?
To try to answer this question, we need to look back and understand at which moments in the past this has happened:
💻 April 2008: Launch of GitHub
This was the pivotal moment for Open Source. Suddenly code became very accessible and dependency managers started appearing based on public GitHub repositories.
It also standardized the use of Git as a version control system. Before that, people used alternatives like SVN or simply zipped up their code. ✨
📚 September 2008: Launch of Stack Overflow
2008 was a year that quite revolutionized programming.
Before Stack Overflow launched, you had to read the documentation to solve errors. With Stack Overflow, after a Google search, you most likely already had the solution to your problem.
It became the go-to tool for everyone who coded and the productivity boost was massive.
🛩️ June 2021: Launch of GitHub Copilot
13 years passed until the next big revolution.
It feels like AI text suggestions for coding have been with us for decades, but it was in late June 2021 when this first appeared.
Until then, all the autocomplete we had was IDE suggestions. This AI-powered autocomplete suggestion blurred the line between IDE and Editor.
2 days after the technical preview launch, we got access to try it, and we did a livestream with it without having tested it beforehand. We've normalized things that at the time felt like pure black magic.
This is the first moment where we see an AI-generated code suggestion. It blew our minds that from the context of the parameters it knew how to suggest the aggregate instance we wanted to create. Those were the days! 🥲
From that livestream there's another great moment when we asked Copilot what the best z-index is and it said 9999 (sorry, Núria 😬).
We weren't aware that right at that moment everything was about to start changing at a dizzying pace.
🗣️ November 2022: Launch of ChatGPT
Almost a year and a half after the launch of GitHub Copilot, ChatGPT came out. We had little idea at the time about Microsoft's deal with OpenAI and that Copilot's engine was an early version of Codex.
Suddenly we went from using Stack Overflow for everything to asking ChatGPT directly.
We did a livestream with Bea discussing how to get the most out of it, and it became one of our most watched. The interest was real.
🦙 February 2023: Launch of Llama
Meta released the first major open AI model. This triggered the appearance of tons of tools (llama.cpp, Ollama...) and made it possible to code with local models.
✨ 2024: The year of Cursor
This was the year of Cursor. It launched with "tab tab" autocomplete. You accept a suggestion and it suggests the next one to make.
Suddenly, the unbeatable king of AI, GitHub Copilot, started having competition 3 years after its launch.
Also, at the end of the year Anthropic launched the MCP spec, although it went largely unnoticed.
🤖 2025: The year of Agents and Vibe Coding
Cursor added support for agents. A mode that could modify multiple files while you only interacted via chat. Suddenly people who didn't know how to code could build things.
Shortly after, in February, the term Vibe Coding was coined. A term that feels like it's been with us for a decade, but it hasn't even been a year.
The MCP revolution. They were like Pokémon. Gotta catch 'em all!
That same year, endless new coding agents came out, Claude Code appeared, incredibly powerful models. Agents that could work for hours to do what you asked.
There was a point in May where we had to stop and analyze everything that had happened, and we did a special livestream sharing our opinion and analyzing all the remote agents that existed.
On top of that, new editors from Google, Amazon... for coding came out. Countless new models, each more powerful than the last.
From this year on, coding without AI is no longer conceivable.
🍍 2026: The year of... indirect agent coordination? Vibe Coding++
We don't know what's going to happen this coming year, but what we do know is that given the pace we're going, it's going to be a year with even more revolutions than 2025 itself.
One area where coding agents currently fall short is that they're too generalist. Sub-agents still aren't mature. So it makes sense that we'll see more specialized agents working together to accomplish better tasks.
Although if there's going to be a major revolution, it'll probably be something that supercharges Vibe Coding. Right now Vibe Coding is limited. If you don't know how to code, there's a point where agents stop working well. Maybe code becomes transparent and we go from Vibe Coding to Vibe Building.
Also, related: Git. For people who don't know how to code, it's very complicated. Maybe we'll start seeing a simpler alternative (or wrapper).
It's certainly hard to imagine what's going to happen. If you had told us all of this back in 2020, we would have laughed at how impossible it seemed.
What we can tell you is that every week we'll be discussing AI, tech, and programming news at Café con Codely, on both our YouTube and Twitch every Friday at 9 AM CET.
See you there. 😊
And since you've made it this far into the newsletter, here's the joke of the year, which I know you've been waiting for:
> - Knock knock > - An async function! > - Who's there? 😂 😂 😂 😂 😂
See you around!