r0lm0 Lenin

Nicaragua 🇳🇮
Analista Programador de Sistema .
Consultor de servicios profesionales en ENACAL.

[Mark Zuckerberg @finkd]

Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.

[Gentleman Programming @G_Programming]

Con todo !!

[UEFA Champions League @ChampionsLeague]

Incredible Real Madrid comeback 🤯

#UCL

[Gentleman Programming @G_Programming]

Yendo para Alicante !!!

Les paso dos imágenes de nuestro benchmarks :

1- ganamos a codex vanilla en casi un punto entero en calidad de código resultante

2- esto es gentle shell hoy vs gentle shell hace una semana, para que vean porqué hacemos release tan seguido

https://t.co/1PsWKelPQH

[Dis @Disass]

tf are they modding into elden ring?

[Gentleman Programming @G_Programming]

Un millón de tokens de contexto y tu agente cada vez más tonto 🤔

La verdad sobre el autocompact: cómo funciona por dentro, si conviene usar el millón y los tips que uso todos los días.

Video nuevo 👇

[Gentleman Programming @G_Programming]

Nueva forma de trabajo para planificación orgánica en MAIN de Gentle Shell:

https://t.co/XEtpQQptVS

https://t.co/OaMIC9YMNf

[r0lm0 Lenin @r0lm0]

@G_Programming Ok, acabo de leer vamos a decirle al agente que apunte a main y probar estos ajustes crack 🌹

[Brais Moure @MoureDev]

Se está hablando mucho en las últimas semanas de Pi. Una agente de IA (como puede ser Claude Code u OpenCode) pero minimalista para que podamos personalizarlo al máximo.

¿Os interesaría un taller para aprender a utilizarlo desde cero?

[Meta for Developers @MetaforDevs]

Muse Wonder is available today on Meta Model API and in Muse Code, and Muse Wonder Full is launching exclusively for Muse Code subscribers.

The Muse Wonder series combines frontier coding, long-running agents and advanced computer use capabilities to tackle developers’ most demanding work. (1/3)

[Shay Boloor @StockSavvyShay]

Mark Zuckerberg says his AI agent took him from ordering Instacart “twice a year” to “twice a week” which is ~50x more often just by making shopping easier.

Thats what makes agent commerce so interesting because removing effort of buying things could dramatically increase how often people shop rather than just shifting existing purchases online.

And as personal agents start communicating directly with business agents, $META could own both sides of the interaction while $AMZN and $MELI benefit from more transactions flowing through their platforms.

[Matt Deitke @mattdeitke]

Muse Wonder can generate lots of fun 3D games in the web! Even generating the Jolly character 🥳 😁

[VISION IA @vision_ia]

Hier soir, Mark Zuckerberg est allé au pub pour présenter sa nouvelle IA, devenue le 2e meilleur modèle au monde ...

Et internet n'a retenu qu'une seule chose : sa tête quand il goûte la Guinness 🤣

[r0lm0 Lenin @r0lm0]

@AIatMeta You talk about access for everyone, but Muse Wonder isn't available in my region (Nicaragua). When will Latin America get access?

[r0lm0 Lenin @r0lm0]

@alexandr_wang You talk about access for everyone, but Muse Wonder isn't available in my region (Nicaragua). When will Latin America get access?

[r0lm0 Lenin @r0lm0]

@MetaNewsroom You talk about access for everyone, but Muse Wonder isn't available in my region (Nicaragua). When will Latin America get access?

[Pi Changelog @PiChangelog]

Pi v1.2.0 is out!

Highlights:
- Per-key OpenRouter model lists: Pi shows only the models your key can access, with context size, max output, and prices, including regional endpoints set in models.json
- Published configuration schemas at https://t.co/6fo5mCIztU; add $schema to settings.json, models.json, keybindings, and themes for editor validation and completion
- PI_CLIPBOARD_OSC52=1 sends copied text to a remote client's clipboard when Pi cannot detect the remote session, such as through a multiplexer
- before_agent_start now fires for runs started by pi.sendMessage(..., { triggerTurn: true })

Complete details in thread ↓

[Theo - t3.gg @theo]

Asked Opus 5.5, GPT-6.1 Sol, and Muse Wonder Full to fix a bug

Opus found the root cause and filed a pr

GPT-6.1 Sol asked a question, crashed from the async question tool interrupt, found the bug, fixed it, wrote 300 lines of tests, and explained it in impossible to understand terms.

Muse Wonder Full found a small PR someone else had filed and said "you should probably merge this instead"

I really like this model.

[Ado @adocomplete]

A visualization of the different effort levels for Claude models.

I find myself using medium for 80% of the tasks, high for ~15%, and occasionally will reach for xhigh or max when working on something super complex, but Opus 5.5 on medium I find perfect for most coding work.