Eric Ramírez

“I'm going to show you, nobody will ever work as hard as I work.”- Michael Jordan.

@JulySeEs Pronta mejoría, July🙏🏻

@chinojoa Unos músculos se me desgarraron

Uno ve y vive tantas cosas en medicina. Hay pacientes de todo tipo. En estos días, muchos pacientes me mandaron “oraciones” y me preguntaban por WhatsApp como seguía”, y tuve una que se puse guapa porque no fui a trabajar porque yo estaba enferma hahahaha

@ely_michael Y me imagino que aceptaron la oferta por la situación, pero también sucede que nos quieren pagar menos.

Cuando nos solicitan una propuesta de trabajo y nos llaman al año, sabemos que intentaron con otras empresas y no dieron con la solución. Como proveedores de servicios, nos hacemos los locos y le volvemos a enviar la propuesta con un 25% de aumento en el precio.

Una multinacional de retail, instalada en RD, implementó un ERP de capital brasileño y colombiano, no pudieron integrarse con su principal cliente, y con la soga al cuello sub contratan mi empresa para la integración. Los dominicanos siempre somos la 2da opción en tech.

@manuelmejiaio At some point a standard will emerge across these tools.

I love how simple and organized Cursor’s folder/file structure is, everything in one place: .cursor. On the other hand, Claude has a .claude file, another file in the root called .claude.json, and if you’re using the Desktop app, APPDATA\Claude\claude_desktop_config.json 🤡

I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out. I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really). It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely. The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture. We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying. I worry.

It seems that for each particular narrow class of bugs, it becomes so time consuming to set up an agent pipeline that it well suited to that problem set, that I am better off just assigning pieces to our human researchers and/or tackling the pieces myself by hand. Everybody shows off their super-optimized expertly crafted agent pipeline for finding type confusion bugs in XYZ system. But they never talk about: - how long it took them to set up - how long it took them to determine which models/harnesses/hyperperameters were suited for which task - exactly how much was spent on tokens - precisely how much expert guidance was imparted on each link in the chain Etc etc etc All of that said, there will foreseeably, eventually be a model so smart that an expert can dump in 2 million tokens worth of previously suboptimal attempts at agentic bug funding and just say “this old stuff we did was relatively stupid models but it still managed to find bugs. Untangle this garbage and make it into a well oiled machine” — and the model will somehow just do that. And in that case, those who have a huge library of all the very nascent attempts to employ these technologies, which work with mixed results, will be FAR ahead of those who waited around for the magic model to arrive

@algunapaloma Es increíble. A la pobre muchacha la atacaron por pedirle una beca al presidente (ya que no puede aplicar a los programas locales porque es extranjera) pero luego enarbolan discursos de superación . Los estudiantes sobresalientes siempre buscarán las mejores opciones.

Apenas puedo creer que ya pasaron tres años desde que me gradué 🥹. Suerte que nunca conté ese sueño aquí, pues muchos me hubieran dicho que ese sueño no me correspondía, que era muy ambicioso y que me limitara a las opciones dentro de mis posibilidades.

If it weren’t for WhatsApp, Meta would be irrelevant in my life. I can go months without using IG. I just wish people would migrate to Signal (or any other app with end-to-end encryption).

@Jordynww @tdesseyn So there's a market for a competitor👀

@host_down @tdesseyn Yes!! It was called Double Time and is closely resembled windows 97 💀

ive been thinking about this. the "bespoke saas" opportunity for young devs is wildly underrated. your aunts title company will never matter to salesforce. but an 18 year old who builds them a tool that saves 20 hours a week can make 500k. nobody is talking about this

@Jordynww @tdesseyn Did you use a CRM back then?

@tdesseyn My real estate past life comes back to haunt me on x lol

@Blassanchezpena Claro y coincido que se debe proteger a los mas vulnerables y brindarles capacitación, pero ya hay demasiada ansiedad sobre la inteligencia artificial y el futuro de los trabajadores. Si se tiene esa discusión debe ser sin añadir mas leña al fuego.

@host_down Así es. Por eso dice que no será de golpe, llevará un tiempo. Pero ese tiempo debe aprovecharse...

La IA no va a eliminar a las cajeras del supermercado de un golpe. Las va a ir dejando sin horas hasta que no quede ninguna. 🤖🛒 ¿Y ellas qué? Que ganan lo mínimo, sin título, sin red de seguridad. ¿Quién las recapacita? ¿Quién las protege? ¿O simplemente no cuentan? Eso es lo que nadie está discutiendo. #FuturoDelTrabajo

I find the acceptance and defense of mediocrity to be deeply depressing. This is true not just in the context of large language models, but particularly in the case of certain scholars and politicians.

TeamPCP post 4 hours ago:

software engineering is solved btw

@magicmac2000 Well, you are a great human being so you have the upper hand 😎

@host_down At 52 I'm still figuring it out, so you might want to pick a faster role model 😄

When I grow up I want to be like Manuel 😆

@TeDeJagua jajaja A mi me dió clases de antropología en la universidad Fátima Portorreal. La mejor!

Clase de antropología* Yo aguantandome decir "ido a destiempo" cada vez que la profe mencionaba a Balaguer:

@ixuatl @MolodyGs @mayaoviedo Es justo lo que pienso. Gracias por aportar tu opinión @ixuatl !

@MolodyGs @host_down @mayaoviedo Y honestamente, si alguien quiere llamarlo “arte”, “música IA”, tampoco me afecta. Lo raro es ver gente intentando negar cualquier proceso creativo humano sólo porque apareció una herramienta nueva, igual que pasó con la fotografía, el sampling o la música electrónica. 3/3

Perdóneme, compa, pero no existe el "arte hecho con IA". Para que sea arte primero lo tiene que haber hecho una persona.

@MolodyGs @mayaoviedo Entiendo el caso específico, pero tambien podrías incorporarla como herramienta en tu proceso creativo.

@host_down @mayaoviedo Guess what, la música electrónica la hace una persona. Al generar imágenes con IA se pierde el proceso creativo, y por ende, el valor de la propia imagen generada.