Kahn

@delpynews Y Peña tardo 16.000 años en cambiarle. Vamos a estar mejor

▶️ DENUNCIAN PENALMENTE A JORGE BRÍTEZ POR PERJUICIO MONSTRUOSO | 💥 ⚠️Los abogados Juan José Bernis y María Muñoz presentaron una denuncia penal contra el expresidente del IPS, Jorge Magno Brítez, y los miembros de su consejo de administración. Los acusan de presuntas irregularidades que alcanzan un perjuicio total de G. 391.574.472.965. 😱La rosca denunciada, que incluye a exgerentes, directores y consejeros, está en la mira por graves irregularidades en al menos cinco licitaciones de gran envergadura. Entre los puntos más escandalosos saltaron sobrefacturaciones del 60% en lavado de textiles, falta de dictámenes técnicos en sistemas de recaudación y una superposición de contratos para fumigación y desinfección. 📌 La denuncia detalla una omisión deliberada de estudios de mercado y la multiplicación injustificada de contratos para una misma necesidad institucional. Vía: El Observador.

@SantiPenap Pero la puta madre!! Con que energia lo que queres construir datacenters si ya estamos al limite!!

Recibí a Cully Cavness, Presidente de Crusoe AI, una de las empresas más prestigiosas del mundo en construcción de infraestructura de procesamiento de datos para Inteligencia Artificial. Crusoe tiene interés en Paraguay, en nuestra energía y en complementar con nuevas fuentes de generación para potenciar sus operaciones. Una conversación que muestra que el trabajo que estamos haciendo para proyectar el país está generando atención en los lugares correctos. Paraguay está en el radar de quienes están construyendo el futuro.

@alex_prompter Sounds like a pile of crap to justify poor model performance.

Both OpenAI and Anthropic just released official prompting guides. Both say the same thing. Your old prompts don’t work anymore. But for opposite reasons. Claude Opus 4.7 stopped guessing what you meant. It does exactly what you type. Nothing more, nothing less. Vague instructions that worked on 4.6? They now produce narrow, literal, sometimes worse results. Not because the model got dumber. Because it stopped compensating for sloppy thinking. GPT-5.5 went the other direction. OpenAI’s guide literally says: “Don’t carry over instructions from older prompt stacks.” Legacy prompts over-specify the process because older models needed hand-holding. GPT-5.5 doesn’t. That extra detail now creates noise and produces mechanical output. Claude got more literal. GPT got more autonomous. Both now punish the same thing: prompts written without clear thinking behind them. One developer on Reddit captured it perfectly after analyzing hundreds of community posts. The complaints tracked almost perfectly with prompt specificity. Precise prompts got better results on 4.7. Vague prompts got worse. The model didn’t regress. The prompts did. OpenAI’s new framework is “outcome-first prompting.” Describe what good looks like. Define success criteria. Set constraints. Then get out of the way. The model picks the path. Anthropic’s framework is the inverse: be surgically specific about what you want, because the model won’t fill in your blanks anymore. Two different architectures. Two different philosophies. One identical conclusion: the person writing the prompt is now the bottleneck, not the model. Boris Cherny, the engineer who built Claude Code, posted on launch day that even he needed a few days to adjust. That post got 936 likes. Meanwhile, Anthropic increased rate limits for all subscribers because the new tokenizer uses up to 35% more tokens on the same input. The model is more expensive to run lazily. Cheaper to run precisely. The models are converging in capability. The gap between good and bad output is no longer about which model you pick. It’s about the 2 minutes of structured thinking you do before you type anything. That thinking system is the skill. The prompt is just what it produces.

Sucks for an AI agent to delete the prod DB - with no way to back it up - and risk the complete rental business. But the blame sits with the dev who decided to delegate decision making to the AI agent, and then not review actions, just YOLO it. Time for a blameful postmortem...

@AsadAljabane @icanvardar Sure, but that is a very big If...

@icanvardar Honestly it doesn’t matter. If a vibe coder solves a real problem better than a programmer the market already decided.

are vibecoders even programmers?

@chrysb @cursor_ai Unlimited?

i don't see how @cursor_ai survives - for $200/mo, you can essentially get unlimited usage on claude / codex that same usage would cost thousands on cursor even if cursor harness is better, the price difference is just wild

@treadon @Layton_Gott Good luck getting a job with that mentality

@Layton_Gott You didn't need to learn assembly before learning JavaScript... you don't need to learn coding anymore.

Learn to code first, then use AI? Or Start with AI, learn as you go? Which is better?

@Layton_Gott You have to do it yourself. You can learn from AI but you have to think as you write the code.

@milan_milanovic Yes, next question.

When Opus 5.0 with 10M context is released, do you still have a job as a developer ?

@hiarun02 You need to learn everything of those so you can spot AI bullshit.

Honest question: > AI can code. > AI can debug. > AI can review PRs. > AI can deploy. > AI can secure apps. What exactly are we learning in 4 year CS degrees?

@manuelnosoy Si es asi, tarde o temprano las suscripciones se van a eliminar y todo el mundo a pagar por api call

@juancristobalx @kiaran_ritchie Cmmon man, there is no parh from LLMs to AGI or ASI.

@kiaran_ritchie because they're shooting to get past that point. if they do, they get to ASI. no one knows what comes after that, that's why it's called the singularity, you can't look past the event horizon. if they don't get to ASI they get fully commoditized

I don't see how Anthropic, OpenAI or any of the model providers have any hope of defending their moats. And consequently, I think they're going to get wiped out. Right now, in early 2026 they have a meaningful advantage in terms of model capability. But far cheaper and open source models are not far behind. How long can they maintain a meaningful advantage? For the vast majority of use cases, we don't actually need much higher intelligence. It doesn't take 140 IQ to automate Turbotax or powerpoint. Eventually we will be saturated in cheap, local models that are "good enough". Of course some scientific labs and frontier research will always want the latest and greatest. But that market is orders of magnitude smaller than these company valuations can justify. What am I missing?

@mike49264897 @ibuildthecloud Can be used with an api key only, right?

@ibuildthecloud OPENCODE!!! It’s an open source claude code alternative.

It really would be great if Claude Code was open source. Just plain idiotic. Just don't get why they closed-sourced it. As a side note run Claude Code behind a proxy and see how many metrics it reports. Holy crap. It sends so much information.

@fedenunes @leoberniga Y es como pagar la suscripcion de netflix..

@leoberniga Nunca entendí lo del costo de mantenimiento por mes, por qué cobran eso? Cuál es la explicación?

Los bancos están DISPARANDO los costos de las TC 💳💸 ❌️ Suben intereses ❌️ Si antes "no cobraban" mantenimiento, ahora sí. ❌️ Te clavan un "seguro de vida" según tu nivel de endeudamiento. Cuanto más "cuotas sin intereses", más "compensación" vía seguro. ❌️ Menos promos

@burkeholland On Friday

When is the last time you wrote code?

@DiarioBitcoin Simple, porque queres que alguien entendido certifique y se haga responsable por el trabajo. Acaso Claude Code se va a jacer cargp si algo falla terriblemente?

"Dinos por que es mejor contratarte en vez de crear un agente con Claude Code?" Tu respuesta?

@_rabbi U fool 🤣

Just shipped a feature to client 30 minutes after he asked for it. Big mistake

@AryamanIyer3 @alexcooldev 4 seconds? Not true. It will thinl for 2 to 5 mins and if you are lucky it will one shot it.

I saw a young guy coding in a coffee shop today. No Codex. No Cursor. No OpenCode. No Claude Code. Just raw… manual typing. Line by line. From his own brain. No AI copilots. No autocomplete crutches. Lowkey terrifying. Respectfully… that’s villain origin story energy.

@AlexCVJ Pero si es una copia de mierda d las escenas de matrix!!

Los AI bros están bien lefios. Veo constantemente cosas como esta de “A Hollywood le costó 10 millones de dólares” a Seedance 2.0 solo 1 dólar. No, el costo de algo como esto son los 10M de la creación de la escena principal, más tu dolar. Sin la original no existe el AI slop.

@doctorwasif What is this copy paste bullshit!!