Alex Ramirez 🦀

Stop showing me how many LOC did you write, but how many tokens did you burn 🔥

Agentic Engineering... I like it.

@freddier No se necesita ser un buen diseñador pencil.dev

Ser un gran diseñador con conocimientos básicos de código y Linux, junto con Claude Code/Codex, tiene que ser la combinación de skills más poderosa del momento.

@esrtweet LLMs are getting really good at reading and following documentation of new tools; As long as the language is well documented, I don’t see this as an issue.

My experience with LLM-assisted coding has been great and I'm a big fan of it, but I've just had a slightly depressing realization. It may almost entirely shut down the development and adoption of new computer languages. The percentage, and probably the absolute amount of code, handwritten by humans is going to fall a great deal. But for the foreseeable future, LLMs won't be able to write code fluently in a specific language without having a large volume of good code in that specific language already available to train on. For a new language in 2026 and after, where exactly is that large volume of good training data going to come from? Probably not from human beings, and where is the incentive for an LLM handed a vibecoding task to go looking for an exotic new language to do it in? I find this slightly depressing, because I enjoy contemplating new-language development the way a more physical tinkerer enjoys salivating over shiny new tools. Human beings are still going to write new languages occasionally, because that's huge fun (if you have a brain bent anywhere like the way mine is) and still a way to climb some status ladders. But with the barrier to mass adoption getting so much higher, I have to think the level of research and engineering activity put into this is going to drop a lot. There is one not-unhappy but rather weird way I could be wrong about this. Historically, once the development of compilers got to a certain point it became clear that designing machine instruction sets to be easily reasoned about by humans was a big mistake. We had to figure out how to design machine instruction sets that were easy for the compilers to reason about. Thus, RISC. It could be that's the future of language design, too. But I have no idea what a new language design optimized for LLM code generation would look like. And I don't think anybody else does, either. Interesting times, indeed.

This

@claudeai I have applied for Medellin, Colombia 🇨🇴

We're launching Claude Community Ambassadors. Lead local meetups, bring builders together, and partner with our team. Open to any background, anywhere in the world. Apply: claude.com/community/amba…

@boxmining Mine acts like that every morning when I ask it about the conversation that we had the day before.

Your openclaw agent when you clear context window.

@AlexFinn It means you don't need a Mac Studio to escape the 'Permanent Underclass'

Do you understand what this means? Are you aware how much the world just changed? You can now run frontier intelligence on a potato Your $600 Mac Mini can now run unlimited super intelligence for free. No authoritarian AI companies can cut you off Do this immediately, no matter what device you’re on: 1. Download LMstudio 2. Find these models in the search 3. Look for the MLX ones if you’re on Mac 4. Download and load them 5. Ask your OpenClaw to use them for most tasks I thought the future was a year away. Nah. It’s today

I've been running an AI agent swarm on old Intel Macs with local models. This changes everything — sharing this post with Astra with an instruction: "Help yourself!"

@sama @mxstbr Does this deal give the DoW any access to ChatGPT conversation data in any way, shape or form; or are the classified deployments fully separate systems?

I'd like to answer questions about our work with the DoW and our thinking over the past few days. Please AMA.

@gndx Por allá nos veremos. 🦞

🦞 Medellin: Este 28 feb (2026) AI Tinkerers hace el Unhackathon en Medellín: 5 horas construyendo (1–6pm). Sin paneles, sin jueces, sin pitch decks: solo código que corre.

@waylearnlatam @soymaikoldev @catmcgee Gracias! Me registraré

@RamirezAlex @soymaikoldev @catmcgee Ya puedes registrarte en el de abril: luma.com/3e2gyosm

Hoy inició el bootcamp de @waylearnlatam Buena introducción, muchas personas. Se construye en Solana y todo en español. Hasta anunciaron un premios para los tres mejores proyectos🫂 Your Spanish didn't seem so bad compared to my English 😂@catmcgee

Hot Take: "The best feature in @openclaw isn't a feature — it's SOUL.md" I'm ignorant — was this your idea @steipete ?

@ASvanevik Hahaha! I know the feeling.

incredible tech

I shared it with Astra... 😅

@Riowgmi Sell them and buy Apple stock.

i’ve got 4 mac mini now tell me how to make money

That's a good question. In this case Astra, the COO, is the one calling the shots — in charge of orchestrating the agents' work and escalating to me if necessary. As of right now, Sage does the research first and once Astra has all it needs, it spawns Jax. Basically, Sage and Jax don't talk to each other right now without Astra's intervention.

@RamirezAlex naming and assigning roles is the easy part. the real test is what happens when sage's research contradicts what jax is building. coordination overhead scales faster than productivity in multi agent systems.

Gave my AI Agent Swarm an office. Now I can watch them work in real-time — research, code, design, write — all running in parallel. Managing AI teams from a dashboard hits different. I say the word and Astra, my COO in the Command Center, dispatches them to work. Astra orchestrated the entire workflow and even trimmed the final video for me — without me telling her where the file was located or which file to trim. This was a collaborative effort: 🔍 Sage researched the latest trends and tech for a Virtual AI Agent Office 🎨 Nova designed it 💻 Jax wrote the code 🤖 Astra kept the team in line until I was happy with the results