Rodrigo Díaz Jonguitud

Non-technical, teaching myself to build with AI. The real project is self-mastery.

Imperare sibi maximum imperium est.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

I work in tech, but near the bottom of the ladder, with very basic technical knowledge.

On the other side, we all know who Kun Chen (@kunchenguid) is. One of the tools he built, and that I use daily, is Lavish (https://t.co/ZaR41uLUnW). It lets you give AI feedback on an HTML page instead of reading walls of text in a chat. You click a section and leave an annotation.

I noticed you could only annotate with text. So I proposed adding images, to show the AI exactly what you're seeing on the page, or just give it more context. I opened the issue, built it, and Kun merged it.

Who won here:
- me, the guy who spotted the missing feature and built it
- Kun, whose tool now has a new feature
- every future user, who won't even notice it was ever missing

What still amazes me is that AI let me work with someone on a whole different level than me. Without AI, the two of us would never have "worked" together. Now his tool has my feature in it.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

So Gemini 4 Argon flopped?

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@kunchenguid Is this still applicable? Ever since I read this I compact by default before stepping away from an important session for a few hours, to "save tokens"

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

One of the best features of @t3dotcodes is the ability to snooze threads. Sometimes I'm working on something I know I won't be able to take care of today, and being able to snooze it til tomorrow (or even next week) gives me peace of mind. Little by little my list of active chats goes down and I get that feeling of closure at the end of the day.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@b1rkh0ff @buildwtim Fuck...per day? I cant do that shit...

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@VadimStrizheus @bot I use it to help me a bit, but I don't give it full control...maybe I would for a secondary niche account.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@topxea Yes! It sucks! On the other hand, lately we've had a bunch of Codex resets. Hard to tell someone to pick the Claude sub over ChatGPT's if they can only afford one... I guess it's quality vs quantity. Claude's quality is better, but the gap ain't that big, so I prefer quantity...

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

I think the $200 Codex sub is a better deal than the $200 Claude sub right now.

Claude's models are better in my opinion. Easier to work with and they get it right more often. But my Claude usage resets on Friday, and by Tuesday or Wednesday I'm already running low. So I end up moving all the work to ChatGPT, which also doesn't have the 5-hour limit.

That's why ChatGPT is my workhorse. Better models on one side, more usage on the other, and if you only have money for one, usage wins.

I got the ChatGPT sub before the infamous Dev Day nerf, so I'm expecting my usage to drop about 2x at the end of October. Let's see how that goes...

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

Okay not really sure what happens in my situation. I have my main Gmail account running the Super Grok subscription but the problem is that for some odd reason I got banned on X on that account due to the Twitter/X migration. I have a second Gmail account where I have my X premium subscription so not really sure if I'm going to be able to merge my subs...

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@kunchenguid Same here. I'm not interested in using Gemini 3 Argon if the only way to use it's in their harness...

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@theo Wtf is Muse Wonder and why everyone is freaking out about it? What does it bring to the table that other llms dont? Havent paid much attention today... Waiting for the video 👀

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@poteto This tweet is gold. I showed it to Fable 5.1 on high reasoning and asked it to audit our method, SOPs, repos and skills against it. Now I have a fleet of agents going through my LLM wiki and processes to implement this philosophy.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@kunchenguid Should have waited one week lol. Paid $30 for SuperGrok and $10 for X Premium last week. At least I didn't go for the $300 annual SuperGrok. Not sure what to do now

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@B_doong2daddy that would be a cool feature

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@Kappaemmedev I have 12 and I'm still figuring it out too. One that actually works: a bot that scans X every morning and brings me posts worth replying to, with the angle. I still write the replies myself.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@petergyang @bot watched the video this morning! loved it. TY!

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@debs_obrien @bot I think no one around me has heard from @bot or Dot. Neither do they care...

Frustrating not having people to talk about this shit lol pretty sure this will change soon. AI will get more and more mainstream with time.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@designwkarthick This has never happened to me

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

Both are Grok Bots with a persona prompt + a scheduled routine. Mercator runs daily: researches people making money with AI, logs each case with source and flags self-reported numbers. Catón runs Sundays: reads the week's output and tells me what I'm avoiding. Both report to Cayo Mecenas.

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

Well, this is my Grok Bot setup... 11 bots, each with one job. I love Roman history, so they all got Roman names.

Cayo Mecenas, named after Augustus's closest advisor and right-hand man. The only bot I talk to. He orchestrates and routes all the work. Based on @kunchenguid's Firstmate: https://t.co/Imr3Kc2YTZ

Quintiliano, after Rome's great teacher of rhetoric. Keeps my voice and edits every draft before it goes out.

Explorator, Latin for scout. Scouts X twice a day for trends and replies worth writing.

Frontino, after the general who wrote a book on stratagems. Plans my X growth: content pillars, cadence, weekly calendar.

Scriptor, the writer. Helps me turn my own stories and thoughts into essays and long content.

Tabularius, the keeper of the archive. Mines my GitHub and my LLM-Wiki daily for post ideas.

Ratiocinator, the accountant. Reads my X numbers weekly: what worked, what didn't.

Mercator, the merchant. Researches daily how people actually make money with AI.

Catón, after Cato the censor. Every Sunday, tells me what I don't want to hear.

Nomenclator, the one who remembered everyone's name. Designs new bots and keeps the fleet lean.

Actuarius, the clerk who kept the records. Logs what the fleet did each night into my LLM-Wiki.

What can I improve?

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@MatthewBerman Same here. I try to do as few clicks as possible. Crazy the difference between the way I use my computer and workflows while working on my corporate job than when I work in my own stufff. My corporate workflows and PC usage feels super old school now...its frustrating

[Rodrigo Díaz Jonguitud @rdiazjonguitud]

@aryanlabde I'm also feeling I'm getting addicted to AI. Would hate not being able to pay a sub...

I even prefer now using AI rather than playing videogames...wtf.