F2F #78: Moving to Claude

For the past years, I’ve relied heavily on ChatGPT for writing, brainstorming, and coding help, but over time its generic tone, weak long‑form reasoning, and unreliable integrations started to show - also they licked Trump's ass and that's a red line for me.

F2F #78: Moving to Claude
Photo by Aerps.com / Unsplash

For the past years, I have relied heavily on ChatGPT for writing, brainstorming, and coding help. It became part of my daily workflow very early as I forced myself to using AI everyday to build habit. I have blogged many times about it. But over time, I started noticing its limitations: the tone sometimes felt too generic, and it struggled with nuanced or long-form reasoning. It also feels like it can't interact with anything outside of its own app reliably.

On the one hand, you've got to give it to them: they have been the original trailblazers for everyone else, the voice mode is very useful, the custom GPTs worked considerably well and they've made a damn good job at pushing consumer AI to cross the chasm.

On the other hand, OpenAI haven't been able to figure out what they excel at. Now, they're giving a shot at e-learning at health, after failing at a few things. Maybe they're only good at being a generalist but that doesn't resonate well with the ever-tiring promise of "AGI soon". Also, the voice mode is very unreliable, their ecosystem of apps is very subpar, integrations don't work and the general feeling is that they've gotten stuck trying to make their financials work instead of shipping useful stuff.

But what happened with Anthropic and the Pentagon deal was the last straw for me. It's yet another episode of season 4 of "spineless big tech CEO selling their users for profit", so I am well trained now. I can switch SaaS's at the drop of a hat, and so I did.

I tried Claude that same day. The difference was subtle but meaningful. Claude felt more conversational, more "human" in tone, and better at following context across long exchanges. It seems to go more straight to the point and doesn't feel like the intern with infinite time that ChatGPT was.

On top of that, Claude offers the Cowork environment (integrations, automations) and Claude Code, which has brought me back to coding after 15 years - with tears in my eyes.

Switching hasn't been about abandoning ChatGPT: it has been about finding a tool that better matched how I think and write and feel.

I know I might be in the 0.0000001% of privileged people that choose who to conduct business with irrespective of their cost or how much money can I make off them. That's why I don't own a Tesla, why I don't invest in real estate and why I will never work for shit companies and why I boycott companies financing war or fascists.

So far, in less than two weeks, I have been able to work in three different coding projects, contribute to Open Source, learn how to work with agents and get me closer to being the 10x engineer.

Can I become a 10x founder? Come with me and let's find out!