Álvaro ;) (@alpas17_) header
Álvaro ;) (@alpas17_) profile picture

AI made easy for ordinary business

España
Joined July 2016
Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)Photo by Álvaro ;) (@alpas17_)
I've been posting here for a while, but I've never properly introduced myself. So here we go! 👋 I'm Álvaro. I have a degree in Marketing, and I build software. I'm early in my career, but I've already built tools like Find Everything and Agent History Search, as well as websites for businesses. My goal now? Turn my skills into real business opportunities. I want to help businesses use AI, technology, and marketing to solve problems, save time, and grow. Here's what I'll share: 🤖 Practical AI tools & automation 📈 Marketing, business & growth strategies 💻 Projects I'm building and lessons learned 💡 Business opportunities worth exploring I'm here to meet business owners, founders, and talented people who want to build something together. If you have a business problem I could help solve, my DMs are open. And I'm committing to posting at least 6 days a week for the next 365 days. Now, enough about me. What are you building, and what's your biggest challenge right now?
9
7
581
Insane aura loss, are we planting trees bro???
Docker got a rebrand
23
Álvaro ;) retweeted
A Polaris se le ha quedado pequeño el Mac, así que le estamos preparando casa nueva. Polaris es el sistema de agentes de IA con el que investigo mi caso en abierto. Corre 24/7 en un Mac mini con 16 GB de memoria que ya va muy justo, y esa memoria no se puede ampliar. Y tiene que ser en local. Mis datos de salud no salen de casa, ni a la nube ni a ningún otro sitio. Así que si quiero modelos más grandes y más agentes trabajando a la vez, la máquina tiene que estar aquí. El servidor lo vamos a montar entre @TukaramX y yo. Él lleva veinte años con el hardware y yo me quedo con el software, que al final es lo mío. Lo iremos contando en una serie en Youtube, desde dónde se atasca hoy el Mac hasta las pruebas con la carga real de Polaris. De lo que falta, la pieza más grande son las GPU: 2 o 3 tarjetas con 24 GB de VRAM o más cada una. La VRAM decide qué modelos puede correr Polaris en casa, y todas las pruebas de la serie irán sobre esas tarjetas, con números reales de velocidad y consumo. @nvidia, me encantaría hacer esta serie con vuestras tarjetas. Si haces hardware para IA local o quieres sponsorizar esta serie ecríbeme por DM Os iré enseñando cada paso 💜
33
192
5
513
39,671
Álvaro ;) retweeted
Ayúdame a compartir el proyecto de Miriam para seguir investigando su cáncer de mama metastásico extremadamente raro. Vamos a conseguir que llegue a empresas y marcas que puedan colaborar en crear su nuevo equipo 💪
A Polaris se le ha quedado pequeño el Mac, así que le estamos preparando casa nueva. Polaris es el sistema de agentes de IA con el que investigo mi caso en abierto. Corre 24/7 en un Mac mini con 16 GB de memoria que ya va muy justo, y esa memoria no se puede ampliar. Y tiene que ser en local. Mis datos de salud no salen de casa, ni a la nube ni a ningún otro sitio. Así que si quiero modelos más grandes y más agentes trabajando a la vez, la máquina tiene que estar aquí. El servidor lo vamos a montar entre @TukaramX y yo. Él lleva veinte años con el hardware y yo me quedo con el software, que al final es lo mío. Lo iremos contando en una serie en Youtube, desde dónde se atasca hoy el Mac hasta las pruebas con la carga real de Polaris. De lo que falta, la pieza más grande son las GPU: 2 o 3 tarjetas con 24 GB de VRAM o más cada una. La VRAM decide qué modelos puede correr Polaris en casa, y todas las pruebas de la serie irán sobre esas tarjetas, con números reales de velocidad y consumo. @nvidia, me encantaría hacer esta serie con vuestras tarjetas. Si haces hardware para IA local o quieres sponsorizar esta serie ecríbeme por DM Os iré enseñando cada paso 💜
1
203
2
521
18,056
Álvaro ;) retweeted
3/3 website projects completed! Huge thanks to @mayankharsani, @mytwillot and @Elizabeth0vjw for trusting me. I'm just getting started. Next, I'll share what I built, why I made each decision, and what you can learn from it. I'm now taking on paid projects. DMs open!
I'm giving 3 websites a free marketing teardown. I'll tell you the first thing I'd change to make it sell. In exchange, I just ask for your honest feedback on it, to feature on my website. Drop your link below. 3 spots.
1
2
62
Everyone uses agents, but nobody asks them how they are.
23
Porque usarías skool si puedes hacerte tu propia página? Encima tienes que pagar un % a skool y otra comisión al banco
37
There are things that are games/projects and there are work of arts, they're not in the same league
21
Jxl goes brrrr
20
Writing tweets is a dichotomy between thinking a lot and writing the first thing that comes to mind 😂😂
22
Good morning everyone, what are you working on today?
1
1
46
GPT-6.2 Astra and Sol just dropped. Benchmarks? Trust me bro I want to see if they can make video editing and content creation faster and cheaper. I'll burn my tokens so you don't have to. What should I test first?
Replying to @OpenAI
GPT-6.2 Sol—an efficient and versatile model for everyday work. It powers through multistep agentic workflows, writes clearly and naturally, and is intuitive to collaborate with. Many OpenAI employees use GPT-6.2 Sol as their go-to model across technical and non-technical work.
59
You can take an idea from a video, website, or even a random photo and turn it into something of your own. It's crazy how easy it has become to go from inspiration to a working project. I'm doing that right now, things are coming 👀 The caveat is that this works most with opus 5.5 Or have you had results with other models?
23
Si te sale en inglés este post, responde con 🇬🇧🇺🇸 Quiero hacer una prueba, esto es un post normal en español, quiero saber si esto le llega a la audiencia normal estadounidense a la que me suelo dirigir. Probablemente el horario no sea el mejor pero es por saber si funciona la traducción
1
1
62
With the current job market, a job is basically a business with one customer and legal protections (depending on your location) That customer provides 100% of your income. In business that's risky. With AI reshaping the job market, would you rather have one protected customer or five without guarantees? What are your thoughts on this? 🤔
25
Jensen Huang's Twitter/X bio went from "Founder and CEO of NVIDIA" to "Founder of NVIDIA", and the badge is gone. Jensen Huang's Twitter/X bio no longer says "CEO". It went from "Founder and CEO of NVIDIA" to "Founder of NVIDIA", and the badge is gone. I don't know what his next steps are. But a change like this doesn't happen by accident, and my read is that it's a statement of intent. Not confirmed by Nvidia. I'm watching for what comes next.
人走茶凉。英伟达创始人黄仁勋的 X 账号,刚刚显示企业认证标识,已经被取消,同时老哥的简介从英伟达 CEO、创始人,变为仅是创始人。 他简介内,个人的英伟达介绍页入口,也一同被删除。
49
I'm loving using Stickman illustrations to better illustrate the content I'm trying to convey.
39
You can make it as complicated as you want, just pick the one you like best 😂😂
25
I'm looking for a harness to use with Claude because, I don't know about you, but sometimes it eats up a lot of context. Any advice?
36
Asked my AI agent to check a file. “Done! Everything looks good.” It hadn't even opened it. I called it out. “You're absolutely right! My apologies!” Same confidence. Zero shame. So it finally opened a file. The wrong one. I had two files with almost identical names. BRO 😭😭 NOT THAT ONE. THE OTHER ONE. Based on actual events. Literally me today.
3
3
147
Me: add the link from yesterday to the launch list Agent: I don't have access to previous conversations. Could you paste it? Me: you found it. the Vercel one Agent: I can't see earlier sessions. Could you check your browser history? Me: Come on Bro, you wrote it in the other chat Agent: ... The chat was saved. What didn't exist was a way to find it. Thank me later alvaropastor.dev/herramienta…
1
78