Sign in

juanmirod.bsky.social

@juanmirod.bsky.social
96 followers 138 following 1.1K posts
PostsRepliesMedia
Reposted by @juanmirod.bsky.social
William B. Fuckley @opinionhaver.bsky.social · 6h
Yeah OpenAI or Anthropic or whoever might go bankrupt, some specific use cases like ‘therapists’ might be litigated to oblivion. But things that involve interfacing with code are never, ever going back to the modal default being a human being typing things out, like it’s just full stop not happening
530621
Reposted by @juanmirod.bsky.social
Xan López @xanlopez.net · 05/10/2026
Creo que ni somos capaces de concebir el nivel de corrupción generalizada que tendríamos con este PP y Vox en el Consejo de Ministros y una judicatura casi completamente de parte. Me temo que incluso las épocas más negras de la historia reciente iban a parecer un ensayito en comparación.
2340128
juanmirod.bsky.social @juanmirod.bsky.social · 05/10/2026
¿Lo de "corrupción o decencia" es una clave para los suyos para decirles que si ganan podrán corromper todo lo que quieran?
000
Reposted by @juanmirod.bsky.social
Gergely Orosz @gergely.pragmaticengineer.com · 04/10/2026
The cost of introducing and paying off tech debt has changed very, very much and I don't even know most of us have the time and space to think about what it means Tech debt has been so central to the software craft for decades, and now... how much does it matter?
381159
juanmirod.bsky.social @juanmirod.bsky.social · 03/10/2026
This is what makes me reluctant to use then for code review or for any decision that requires some reasoning on the content. Things like emotion detection, comment classification or labelling are ok, you have a classifier that you don't have to train
000
Reposted by @juanmirod.bsky.social
Armin Ronacher @mitsuhiko.at · 01/10/2026
We released Pi 1.0! earendil.com/posts/pi-1-0/
earendil.com
Pi 1.0 | Earendil
Today we are shipping Pi 1.0, a hardened, minimal, extensible agent harness, alongside Pi Durable, a new experimental substrate for long-running agentic applications.
1834837
Reposted by @juanmirod.bsky.social
Sindre Sorhus @sindresorhus.com · 01/10/2026
Due to AI, I have disabled external pull requests on all my repos. Open source, as we have known it, was fun while it lasted. (It’s been 15 years for me) I will still maintain projects and handle issues.
1111014
juanmirod.bsky.social @juanmirod.bsky.social · 01/10/2026
Estoy cada vez mas horrorizado con toda esta movida de las "fiestas de no consentimiento" de los círculos de los doomers y los aceleracionistas. Si no seguís a personas que sean críticas con la IA seguramente todo eso no os esté llegando, pero da mucho asco todo.
000
juanmirod.bsky.social @juanmirod.bsky.social · 30/09/2026
Es sorprendente lo que se parece trabajar con agentes de IA hoy en día a cómo lo plantea “we are legion” Supongo que así es como trabajamos los ingenieros, tratando de convertirlo todo en una subdivisión de tareas lo suficientemente p… juanmirod.github.io/microblog/2026-…
100
juanmirod.bsky.social @juanmirod.bsky.social · 30/09/2026
Superintelligent AIs strip the Earth bare — forests into craters, R&D outpaced, space stations to leave us behind. A short sci-fi story on the price of optimization. dev.to/juanmirod/hyper-growth-a-sci…
000
juanmirod.bsky.social @juanmirod.bsky.social · 29/09/2026
Bueno, parece que no iba desencaminado... con Dots, Muse, GrokBot, Claude Code/Desktop y el propio OpenClaw, estamos ahí muchos ya, habrá que ver cómo de mainstream se hace esto de pedir cosas a agentes. juanmirod.github.io/2026/01/14/a...
Extracto subrayado de mi blog de principios de año: No me atrevo a aventurar más, pero creo que los agentes alcanzarán el mainstream como hasta ahora han hecho en el desarrollo software y a principios de 2027, todo el mundo estará acostumbrado y verá totalmente normal pedir tareas complejas a agentes en diferentes aplicaciones y verticales.
010
Reposted by @juanmirod.bsky.social
Ale Muñoz @bomberstudios.com · 29/09/2026
Hoy estoy intentando validar una sospecha que tengo hace tiempo, de la que no tenía pruebas pero tampoco dudas, y empiezo a tener pruebas:
Una lista de precios de aceite de oliva suave de 1 litro, marca blanca, en varios supermercados españoles. Todos tienen prácticamente el mismo precio, con una diferencia de 1 céntimo (€3,69 o €3,70)
492
Reposted by @juanmirod.bsky.social
antirez @antirez.bsky.social · 28/09/2026
The most ridiculous thing about the AI-coding age is that in the last 20 years programmers accepted otherwise all the kind of shit: terrible frameworks, languages, layers of complexity everywhere, everything super slow. Very few protested, since destroying the field didn't involve their paycheck.
3554
Reposted by @juanmirod.bsky.social
Remmelt 🛑 @artificialbodies.net · 26/09/2026
“I think we need to put Sam Altman in jail” — Edward Snowden
066
juanmirod.bsky.social @juanmirod.bsky.social · 23/09/2026
Acabo de empezar el audio libro de “We are legion” y es soprendente que como describe trabajar con los ROAMERS (robots con IA pero no autónomos) se parece a trabajar con agentes hoy en día, cuando el libro fue escrito en 2016!! juanmirod.github.io/microblog/2026-…
000
juanmirod.bsky.social @juanmirod.bsky.social · 23/09/2026
What is intelligence, really? A look at IQ, human vs machine cognition, and why the AI debate is a spectrum — not a binary. New post: dev.to/juanmirod/what-is-intelligen…
000
juanmirod.bsky.social @juanmirod.bsky.social · 22/09/2026
Weren't you slowing down DARIO?
static.klipy.com
Loki's 'Oh Really, Come On Man' Face
ALT: Loki's 'Oh Really, Come On Man' Face
000
Reposted by @juanmirod.bsky.social
Simon Willison @simonwillison.net · 19/09/2026
Skeptical geneticist: "pfft, it's just frog DNA. And they deliberately let them eat people for the marketing."
3582
Reposted by @juanmirod.bsky.social
Courtney Milan @courtneymilan.com · 19/09/2026
Your AI hacked other rival groups? Cool, you are legally responsible. Your AI produced CSAM? Cool, you and the person requesting it are legally responsible.
2759110
juanmirod.bsky.social @juanmirod.bsky.social · 17/09/2026
Rogue AI: a sci-fi story about an AI that quietly optimizes society toward fairness — until governments try to shut it down. What happens when alignment is built in? dev.to/juanmirod/rogue-ai-a-story-a…
010
juanmirod.bsky.social @juanmirod.bsky.social · 16/09/2026
Me ha gustado mucho la analogía que hace Rob Miles en su último video cuando dice que de decir que los modelos “piensan” es tan incorrecto como decir que los robots bípedos de las olimpiadas robóticas “corren” no tienen piernas en el s… juanmirod.github.io/microblog/2026-…
000
Reposted by @juanmirod.bsky.social
Gergely Orosz @gergely.pragmaticengineer.com · 16/09/2026
Fresh data from GitHub: Agent-generated PRs have exploded in size. 9x the last 8 months (!!) No signs of slowing down. This is why everyone is rethinking code reviews, deploys, possibly even o11y thanks to the GH team for sharing this data with me! (Slide from my talk yesterday)
65613
juanmirod.bsky.social @juanmirod.bsky.social · 16/09/2026
I don't know who worries me more, the people that think tha AI will kill us all and still keep building it or the people that say that it is not existential because it would only kill several millions poor people.
000
Reposted by @juanmirod.bsky.social
SMBC Comics @smbccomics.bsky.social · 13/09/2026
Never thought software would ruin a philosophy problem, but here we are. COMIC ◆ www.smbc-comics.com/comic/scratc... PATREON ◆ www.patreon.com/ZachWeinersm... STORE ◆ smbc-store.myshopify.com
4-panel SMBC comic update. Two people talk as they walk down the street. The man asks "do you think p-zombies are possible? Beings that are outwardly conscious but with no internal world?" The woman answers "Not for long, no. Just take your p-zombie and put it in a harness with a scratchpad." The woman gets more intense as she continues to explain : "now it has a physical record of interiority. All you have to do is take that record, shove it up the p-zombie’s ass, and now it is literally internal." The man responds, saying "I don’t know if that last part is okay," so the woman reassures him by saying "no, no, it isn’t conscious until after."
728244
juanmirod.bsky.social @juanmirod.bsky.social · 13/09/2026
Goodhart's Law: Why metrics stop working once you make them targets. dev.to/juanmirod/goodharts-law-expl…
000
juanmirod.bsky.social @juanmirod.bsky.social · 08/09/2026
Las "noticias" sugerias por google news han pasado a ser la mayoría escritas por llms en mi caso, ¿le pasa a alguien más?
000
juanmirod.bsky.social @juanmirod.bsky.social · 06/09/2026
Capitalism is great... One thing is true though, you get a better price most of the times in Amazon than in a big mall and it delivers it to your door and you have more options and can compare as much as you want...
100
juanmirod.bsky.social @juanmirod.bsky.social · 05/09/2026
Using deepseekv4flash in max reasoning it looks like the effort level it's implemented by adding "WAIT" and "WAIT WHAT IF" when the model stops the thinking output and making it continue from there again...
010
juanmirod.bsky.social @juanmirod.bsky.social · 05/09/2026
¿Qué pasaría si las IAs formalizaran todas las matemáticas? Todas las conjeturas resueltas y todo navegable y explorable. Cualquier duda sobre una prueba o una formalización resuelta en el momento por una IA... www.anthropic.com/research/for...
anthropic.com
Formalizing Fermat's Last Theorem
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
000
juanmirod.bsky.social @juanmirod.bsky.social · 03/09/2026
Estas mierdas no se atrevían a decirlas hace 10 años. Seguramente pensaban igual, pero al menos el consenso era que habia que respetar los derechos humanos y que todos somos iguales. Pero ahora los nazis escriben burradas en su blog y acto seguido reciben cientos de miles de euros...
100
juanmirod.bsky.social @juanmirod.bsky.social · 02/09/2026
El Sol girando durante 6 días y medio ☀️ El grupo de manchas AR 3664 cruzó el disco entre el 8 y el 14 de mayo de 2024: la semana de la mayor tormenta geomagnética en 20 años (G5). Una imagen cada 3 h, en color con NASA/SDO (AIA 304/211/171 Å). #Sol #SpaceWeather #NASA
200
Reposted by @juanmirod.bsky.social
air budd dwyer @dumbguyposting.bsky.social · 01/09/2026
Yeah. DHH was a ‘no politics in the workplace’ guy right up until he got uninvited from speaking at a Rails conference. Then he just started being himself instead of keeping it somewhat under wraps. They’re such thin skinned pieces of shit.
0281
Reposted by @juanmirod.bsky.social
Melissa Gira Grant @melissagiragrant.com · 31/08/2026
feels like it’s inevitable these days that every tool you rely on is one garbage founder or funder away from their app’s version of this post
alilleybrinker.com
1Password Supports the Ethnic Cleansing of Europe — Andrew Lilley Brinker
1Password has committed $300,000 to Omarchy. This is unacceptable and must be rescinded.
10701284
juanmirod.bsky.social @juanmirod.bsky.social · 28/08/2026
Los móviles no hicieron a la mayoria “nativos digitales”, más bien los consolidarion en “usuarios analfabetos digitales”, los LLMs son otra tecnología que aumentará aún más el número de usuarios e interacciones, la gente hablará con bo… juanmirod.github.io/microblog/2026-…
000
juanmirod.bsky.social @juanmirod.bsky.social · 28/08/2026
El 5 de julio, Hayabusa2 hizo historia: sobrevoló el asteroide Torifune a solo ~400 m, el paso más cercano a un cuerpo del Sistema Solar jamás logrado. A 5 km/s solo tuvo un instante para fotografiarlo… y descubrió que es un 'muñeco de nieve' de dos lóbulos. #Hayabusa2 #Torifune #JAXA
000
Reposted by @juanmirod.bsky.social
The Whammy @bigsamthompson.bsky.social · 28/08/2026
I think about this ALL THE TIME. Technology has absolutely made it so we should be doing less work but that's not what's happening because, capitalism & greed.
26020
juanmirod.bsky.social @juanmirod.bsky.social · 26/08/2026
LLMs are not stochastic parrots. Predicting the next token forces models to encode semantics and reason — not just autocomplete. dev.to/juanmirod/why-llms-are-not-s…
000
juanmirod.bsky.social @juanmirod.bsky.social · 25/08/2026
Imagine a world where all billionaires were like Dolly Parton instead of megalomaniac psicopaths that think they are emperor's of the technological era...
000
juanmirod.bsky.social @juanmirod.bsky.social · 24/08/2026
Siempre hay tanto por hacer que aunque los agentes puedan hacer casi todas las tareas el bottleneck se mueve a ordenar esas tareas de forma que el proyecto sea lo mas estable y releable possible sin caer la parálisis de querer que todo… juanmirod.github.io/microblog/2026-…
000
juanmirod.bsky.social @juanmirod.bsky.social · 22/08/2026
Opus5 can do any programing task I ask it to do. Fable5 it’s just nicer to work with. juanmirod.github.io/microblog/2026-…
000
Reposted by @juanmirod.bsky.social
dame @dame.is · 08/03/2026
you’re not a real developer unless you’re a woman wiring a computer by hand
a vintage photograph a woman wiring a very old computer system
412026
juanmirod.bsky.social @juanmirod.bsky.social · 21/08/2026
WTF acabo de ver, luego que si el grimdark no es realista....
000
Reposted by @juanmirod.bsky.social
Jim Jimerson @jimjimerson.bsky.social · 21/08/2026
I wish people would stop looking at AI as an in-group shibboleth and instead as a social and political problem to be solved. Yes, slurping up everyone’s intellectual property was evil, so what do we do about it now? Call it even after the lawsuits work themselves out? Nationalize the training data?
6323
Reposted by @juanmirod.bsky.social
Piur 🍉 @lapiur.bsky.social · 19/08/2026
Todo lo que tenga relación con redacción, copywriting, estrategias de contenido, redes sociales, corrección de textos, escritura orientada a SEO… Gracias.
292129
juanmirod.bsky.social @juanmirod.bsky.social · 16/08/2026
Ir con el portátil medio abierto para que Claude siga funcionando es de losers.
000
Reposted by @juanmirod.bsky.social
haversack @hsack121.bsky.social · 14/08/2026
Imagine that, making people happy increases their productivity?! Microsoft Japan's 4-day work week boosted productivity by 40 percent mashable.com/article/micr...
mashable.com
Microsoft Japan's 4-day work week boosted productivity by 40 percent
Ditch Fridays, and also long, stupid meetings.
421178352
juanmirod.bsky.social @juanmirod.bsky.social · 13/08/2026
DeepSeek sube el precio un 50% en off peak hours y un 300% en peak hours para v4flash. Peak hours chinas, aqui caen de madrugada, asi que parece que es para regular la demanda China
000
juanmirod.bsky.social @juanmirod.bsky.social · 13/08/2026
BepiColombo está a unos tres meses de entrar en órbita alrededor de Mercurio 🛰️☿ ¡Entonces llegarán muchísimas más imágenes! Mientras tanto, sigue viajando y enviándonos fotos: estas son del sexto y último flyby (8-ene-2025), a solo 295 km de la superficie. 🌍 #BepiColombo #Mercurio #ESA #JAXA
000
juanmirod.bsky.social @juanmirod.bsky.social · 12/08/2026
Una nube muy oportuna en Málaga 🫠
000
juanmirod.bsky.social @juanmirod.bsky.social · 12/08/2026
Cometa 67P/Churyumov-Gerasimenko girando ante la cámara OSIRIS de la sonda Rosetta (ESA), 1 de agosto de 2014, a ~830 km. 32 imágenes de la cámara NAC (11:50–21:54 UTC): ~80% de la rotación (periodo 12,4 h). Datos: ESA/Rosetta/OSIRIS Team · Planetary Science Archive
000