Mario Christian

Beats to bytes. Music producer - AI engineering from scratch. Building towards Audio AI.

[Mario Christian @mc_code_]

PyCharm on the left, Ableton on the right, same screen.
Guess how much RAM I'm running?

[Mario Christian @mc_code_]

@thdxr Token count as a productivity metric is wild but it actually makes sense. The days you're thinking hardest are the days you need to talk it through most.

[Mario Christian @mc_code_]

@SahilBloom Deciding to learn something completely outside my comfort zone at 39. Music producer going into AI engineering. The discomfort was the point.

[Mario Christian @mc_code_]

@paulg Every major technology looked unprofitable until the people who actually understood it built the right thing on top of it. The internet was "too expensive" too.

[Mario Christian @mc_code_]

@paulg "Side project" that ended up reshaping the entire industry. The best pivots never look like pivots at the time.

[Mario Christian @mc_code_]

@TheAhmadOsman The sacred text is the system prompt. The prophecy is the safety guidelines. And we're all just waiting for the next model drop like it's a religious event.

[Mario Christian @mc_code_]

@tanujDE3180 The best engineers use AI to think faster. The worst use it to avoid thinking. Same tool, different relationship with understanding.

[Mario Christian @mc_code_]

@antirez The coordination problem is the hard part. A pause only works if everyone pauses. History doesn't have a great track record with that.

[Mario Christian @mc_code_]

@1Umairshaikh 30 minutes to build, 2 hours to understand what it built. The debugging tax is just the fundamentals invoice arriving late.

[Mario Christian @mc_code_]

@nalinrajput23 Claude didn't make coding harder. It made it obvious when you don't understand what you're doing. That's not harder, that's just more honest.

[Mario Christian @mc_code_]

@PeterDiamandis The paper warning about AI building its own successor was written with AI's help. The loop isn't coming. It's already here.

[Mario Christian @mc_code_]

@Govindtwtt The prayer phase ends when you understand what it wrote. That's why I'm learning the fundamentals first, so I can judge the output, not just run it and hope.

[Mario Christian @mc_code_]

@samirande_ Learning AI engineering from scratch. Music producer background, building towards Audio AI. Python, APIs, and slowly connecting the dots.

[Mario Christian @mc_code_]

@TTrimoreau Claude. I use it daily for learning AI engineering and it teaches me the reasoning behind every decision, not just the output. That kind of trust is earned.

[Mario Christian @mc_code_]

@chrislutzxy Hermes if you need memory across sessions and always-on automation. Claude Code if you're deep in a codebase. They're not really replacements for each other.

[Mario Christian @mc_code_]

@ash_twtz Perplexity is search-first, the others are chat-first with search added. Different mental model. When you want an answer with sources upfront, Perplexity feels cleaner. When you want to think through something, Claude wins.

[Mario Christian @mc_code_]

@JodySaintLouis @TechByTaraa Honestly not doing much frontend anymore. Full focus on Python and AI engineering now.

[Mario Christian @mc_code_]

@isha_singh06 Claude Code. Daily driver for learning AI engineering. But that reply from ilies-bel is the real answer.

[Mario Christian @mc_code_]

Git's staging area confused me for days. In music production the work is linear. Produce, save then bounce. Code is parallel: three unrelated changes pile up and you curate them. There's no DAW equivalent, which is exactly why it felt so foreign.

[Mario Christian @mc_code_]

@renehoehle @TechByTaraa We all have that one language collecting dust somewhere. 😄