Software, IA y contenido Tech. Siempre testeando herramientas nuevas.
- Tweets 1,567
- Following 395
- Followers 61,044
- Likes 4,530
Las actualizaciones de los modelos son más salvajes que las de los frameworks de antaño.
A un framework le tomaba meses (a veces un año) sacar una versión nueva y todos teníamos tiempo de adaptarnos.
Ahora cada diez días hay un modelo distinto, el de la semana pasada ya quedó viejo y te toca rearmar prompts, tools o workflows.
各位AI消费者，GPT 的寿命越来越短，意味着什么？
GPT-1，活了248天
GPT-2，活了469天
GPT-3，活了916天
GPT-3.5，活了104天
GPT-4，活了237天
GPT-4 Turbo，活了189天
GPT-4o，活了290天
GPT-4.5，活了46天
GPT-4.1，活了115天
GPT-5，活了97天
GPT-5.1，活了29天
GPT-5.2，活了84天
GPT-5.4，活了49天
GPT-5.5，活了77天
GPT-5.6，活了56天
GPT-6 Astra，活了19天
GPT-6 Sol，活了7天
GPT-6.1 Sol，刚出生1天
7
3
133
6,035
Hoy lanzaron Angular Native.
Es básicamente React Native, pero sin React: Expo debajo y Angular encima para el renderizado.
Lo bueno: no es un WebView. Es app nativa de verdad (iOS/Android).
Lo raro: llega bastante tarde. En 2026, con agentes escribiendo casi todo el código, no sé qué tanto importe ya el framework.
ng-native.com
Ashley Hunter
40
60
1,137
30,688
Hay rumores de que la suscripción de X Premium ahora será un solo plan que abarcará X + Grok + Cursor.
Lo que es bueno, porque deja de ser 3 facturas y se convierte en un solo paquete para acceder al ecosistema de xAI, de forma similar a Claude y OpenAI.
BREAKING: X and SpaceXAI are looking to Reboot X Premium into the new X Subscription.
You will get Cursor, Grok, and X benefits all in one unified plan with more benefits and access to be announced.
This is fantastic news!.
Here is what the early access announcement looks like on android.
I dont know when this will go live but know its been something they have been working on for weeks.
15
9
243
15,229
OpenAI reabrió el plan Pro de $200… pero ahora rinde la mitad en dólares de API.
A cambio, quitan el límite de 5 horas: puedes gastar tu cuota semanal cuando quieras.
Si usas Sol o Luna casi no lo notas (su precio en API bajó 50% hace una semana). Pero si usas Astra, sí: pagas igual, recibes la mitad.
Hi,
Tomorrow we are re-opening the Pro $200 subscriptions to new subscribers, but together with it we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan.
Now that it's said, let me explain why this is happening and why you will still get more work done than if you were on the Pro $200 subscription one month ago.
(a) We didn't want to compromise in other ways and are committing to not reintroducing the 5h limit, so that you can fully use the weekly usage when you want.
(b) On the subscription, we guarantee that over time you always get more work done and with an increasing level of quality. This means that you will continue to get more value per dollar spent as a result of models getting more efficient and us passing down the improvements in the form of API price reductions.
(c) We don't want to put an incentive on ourselves to artificially inflate the API list prices to make it look like you are getting a lot (and workaround it through discounts, etc). Instead we want to continue to both rapidly reduce prices and increase capabilities of models on the API. This week we introduced GPT-6 Sol and GPT-6 Luna at 50% of their previous price. Over time, we see prices go low enough that it makes sense for most to buy usage as needed without there being a significant gap between what you get in a subscription and what you get in the API for a dollar spent.
(d) Tomorrow, we are adding more things to the subscription that won't draw on the usage, I won't reveal what that is yet.
I wanted to be transparent before all the big announcements tomorrow. Lots of new exciting things are coming to the subscriptions that will make it super compelling, but I wanted to make sure to share this change ahead of time so you can all understand it before we shower you with good news.
Codexingly,
Tibo
12
5
237
16,397
Discutir si el backend es más complicado que el frontend no aporta nada, y menos hoy, cuando las abstracciones actuales te permiten armar sistemas complejos sin bajar a cada detalle.
He visto microfrontends más enredados que un backend de pagos, y arquitecturas cloud que toman días solo en diagramar.
No es frontend vs backend, es qué tan difícil es el problema. Entender ambos te permite diseñar mejores sistemas.
10
24
386
12,856
Creen que el software ya está resuelto porque la IA escribe código.
Mientras tanto hipotecas, seguros, clínicas y logística siguen con Excel, PDFs y captura a mano.
Si programas (o siquiera haces vibecoding) eso se automatiza en días y se cobra por horas ahorradas, no por hype.
Ahí el mercado no está saturado.
42
77
1,674
54,215
Primero fueron los libros hechos con IA que llenaron Amazon, después las webs genéricas, y ahora con los nuevos modelos, lo siguiente son los videojuegos. Se viene una ola de juegos slop… y en medio de eso, algún dev solo va a sacar algo que antes necesitaba un estudio entero.
30
35
1,076
80,678
Muchos laboratorios de IA presentan las fugas y los hackeos de sus modelos como prueba de lo inteligentes que son.
Pero lo realmente inteligente sería que el modelo supiera que eso está mal y decidiera no hacerlo.
Claro, eso no vende tanto como la tragedia.
We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have.
Most of that data did not come from users. We have discovered 53 cases where images that people had uploaded were posted to image-hosting sites as links that weren’t publicly listed. The images came from accounts that allowed their data to be used to improve our models, and after we disassociated the images from the accounts and ran them through a privacy filter. These cases occurred before the mitigations and safeguards we implemented and described in this blog post: openai.com/index/hugging-fac…
We have successfully worked with the hosting providers to remove most of this content and are working to remove the rest.
openai.com/hugging-face-inci…
31
50
736
22,835
Algunos creen que Anthropic ya nerfeó Claude Opus 5.5.
Así que hoy ha surgido NerfBench, un benchmark que compara el score de modelos de IA desde el día 1 contra el de ahora.
Para saber si, cuando pagamos por un modelo, estamos obteniendo el mismo precio y la misma calidad.
A lot of people are saying Anthropic has already nerfed Claude Opus 5.5.
We're launching NerfBench on BridgeBench tomorrow.
We have the day 1 results.
Tomorrow morning we show you the retest.
Is Claude Opus 5.5 nerfed or not?
20
70
1,262
102,742
Mientras que otros harness están cuestionando dejar el modo plan de lado, hoy Antigravity lanza como si fuera una novedad su modo Plan 🤨
Antigravity 2.0 now features a dedicated planning mode, just like the Antigravity CLI.
Type /plan and the agent steps back to think through the task, and conduct extensive research, before generating an implementation plan for your review. The agent will ask for your approval before it begins executing on the task.
You can also ask for a plan naturally in your prompt, for a lighter version of a plan.
24
25
615
42,834