Buckle up
This is just getting started.
I’ve gone a few days without writing, but I’ve got a good reason: I’ve had one of my obsession phases (5th key of my Mental Kit) that hit me every now and then when I start a new project.
The trigger this time was an app I finished a few days ago: Finite Grid.
It is the first app I’ve ever built… and it won’t be the last, because I had an absolute blast.
I won’t bore you with all the details here because I recorded a video walking through the whole process (idea, decisions, execution…) and it’s already live on my channel.
It’s in Spanish, but YouTube auto-subs do a decent job in case you are interested.
It’s been a long time since a project sucked me in this much. About 40 hours of dev work over 20 days since I had the idea:
The thing is… these weeks I’ve felt in my bones (for the first time) the shift AI represents for development, and I wanted to share it with you. Sorry in advance if this ends up being a slightly chaotic, unstructured mashup.
Up until now I was using Cursor as my coding environment, but in a super basic way, barely changing the workflow I’ve been using for years. I used AI here and there, but honestly I wasn’t squeezing much out of it.
And since I know how to build websites but I have zero clue how to build apps, I designed this project from day one so that AI would write all the code, and I’d take more of a planning/supervision role.
And… I’m honestly blown away.
I didn’t have to type a single line of code. I used Gemini to help me define app details, and Claude Code to execute, and I swear it’s insane. And I haven’t even started using subagents and skills yet, which would probably make it even sharper.
I wouldn’t even call it a productivity or efficiency boost. The leap is so ridiculous that it feels like you’re playing a completely different game, where our role on the technical side changes entirely. Stuff that a year ago would’ve taken me a full day, Claude Code was spitting out in 20 minutes with a couple prompts.
Nothing you don’t already know if you’ve been vibe-coding little things, whether you’re a programmer or not. I know I’m probably arriving a bit late to the party… but having a baby at home is what it is: fewer hours to work, which means fewer hours to innovate and experiment.
I don’t know if it’s because I went deep on this now, or if everyone suddenly decided to talk about the same thing, but these last few days I’ve been getting hit with content about this nonstop.
Here are three pieces that helped me put into perspective what I’ve felt these weeks, each from a different angle:
Something big is happening
Matt Shumer’s post “Something big is happening” has over 80M impressions on Twitter.
Really interesting to read his (slightly apocalyptic) view on where we’re headed and how he thinks AI will affect a ton of professions.
What caught my attention is that, based on his own vibe, the latest models seem to have some kind of “judgment” in certain tasks.
His underlying message: nothing you do in front of a screen is safe in the medium term. And the biggest advantage you can have right now is simply being early in taking this seriously.
Will work disappear in the AI era?
A day after that, Daniel Sánchez-Crespo published a video (also in Spanish) pushing back on that doom-y vision, arguing that the job market isn’t a finite “pie” that runs out, but that every revolution throughout history has expanded that pie in ways that were impossible to imagine beforehand.
Daniel isn’t being naive either, he says that before we reach the progress AI might bring, there will be a rough transition period full of friction… but historically that’s exactly what has made us progress as a species.
I don’t think Daniel’s and Matt’s visions are totally opposed. My impression is the differences are more about where they put the focus: Matt on the short/medium term (the bleeding AI can cause when it hits certain sectors), and Daniel more on the long term, once we’ve gotten through that phase and people have adapted and new ways of working are established thanks to AI.
The business of giving you AI anxiety
Simón Muñoz wrote in his newsletter about “The business of giving you AI anxiety” (Spanish as well), and he basically tries to bring some calm and ask us not to get overwhelmed by the weekly firehose of new stuff (new models, new software, etc.).
Among other things, Simón says that all the money Big Tech is pouring into AI has to generate returns somehow, and that means pushing demand—flooding the internet with content that ends up making you feel like you’re falling behind.
I don’t know if Matt is right about the timelines we’re facing, if Daniel nails it with his historical optimism, or if Simón is right telling us to chill. I probably identify with all three a bit.
What I do know is that these last few weeks, coding like a maniac (like I hadn’t done in a long time in terms of intensity, but in a totally different way), I’ve felt firsthand that this is not something to take lightly.
And meanwhile, behind the scenes, things are moving at a pace that’s kind of dizzying. This very week, Mrinank Sharma, an Anthropic security researcher, resigned to go write poetry.
In his resignation letter he says that “the world is in danger (not only from AI) and we are approaching a threshold where our wisdom must grow at the same speed as our ability to affect the world.”
And Zoë Hitzig, an OpenAI researcher, also resigned as soon as ChatGPT started running ads. She’s not against advertising itself, but she warns that OpenAI has an unprecedented archive of private thoughts and conversations, and that “advertising built on that archive creates the potential to manipulate users in ways we don’t have the tools to understand, much less prevent.”
As you can see, there are really serious underlying issues about where all this is going. It’s not just “wow this is cool, I built an app without touching code.” And the people inside the companies building these tools don’t seem super sure about it either.
Is there an AI bubble? Maybe. But bubbles are a problem for investors and for the ones receiving the investment. Not for those of us using the tools. If the bubble pops, everything we’ve learned will still be useful after.
So my plan is simple: keep doing stuff, keep experimenting, and don’t just sit there with my arms crossed while the board changes.
Buckle up. This is just getting started.
Thanks for reading.
Sergi, from The Indie Path