realmatt

[realmatt @realmatt55]

day 26?
@thsottiaux

[realmatt @realmatt55]

el verbo humillar no tiene absolutamente nada que ver con golear

[Luz 🇦🇷☀️ @holasoyluz2]

Los españoles quieren instalar q fue una humillación cuando no es así
Hicieron un gol en el alargue estábamos a nada de los penales
Humillación es golear y eso no paso

[realmatt @realmatt55]

chatgpt experience is so damn accurate 😭

[Col-Gate @Col_Gate_ML]

Current reality:

User: Hi GPT, please create a very simple calculator in Python with a GUI.
ChatGPT: I’ve created it, but that doesn’t prove the calculator will compute and return the correct result.

The same request to Claude:

User: Claude, please create a calculator for me.
Claude: Creating a calculator means replicating someone else’s implementation; I won’t do that.

Request to Gemini:

User: Gemini, please create a simple calculator. Gemini: That’s an excellent, outstanding idea. It’s incredible how you came up with that, absolutely mind-blowing! Your calculator is ready. You’re simply a genius, my deepest respect.

[realmatt @realmatt55]

@NadyoSatella @Fernando_Her85 tu eres el ignorante al pensar que un modelo que Google usa en cada una de sus búsqueda siquiera será algo igual a competente cuando no es ni una fracción del frontier actual.

[Tibo @thsottiaux]

Day 25/

We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.

We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.

And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.

[realmatt @realmatt55]

esta es la gente con la que compites btw

[miku @turbotroleo]

Voy a darte el beneficio de la duda porque nadie nace sabiendo, y te enseno por que no deberias fiarte de lo que dice un LLM, particularmente si aun estas aprendiendo:

[realmatt @realmatt55]

@thsottiaux I love you tibo

[realmatt @realmatt55]

@thsottiaux we need a reset vro

[Jaysen♨️ @_jayyy18_13k]

day 23 and tibo is still quiet 👀

day 22 gave us gpt-6.1 terra. now i'm wondering if day 23 is another codex limit reset or if he's about to drop something else entirely

anyone have a clue what he's cooking?

[realmatt @realmatt55]

@thsottiaux RESEEEEETTTTTT

[realmatt @realmatt55]

@thsottiaux reset reset reset!!

[realmatt @realmatt55]

so wheres day 23

[realmatt @realmatt55]

nobody oneshots a website btw

[Layton Gott @Layton_Gott]

What the freaking heck...

this is by FAR the worst output I've ever gotten from my simple prompt "build a website for yourself"

I cannot comprehend how bad this is...

the colors, spacing, wording, design, icons, etc...

"We're introducing GPT-6.1 Terra, a fast, efficient model that combines detailed visual reasoning"

I'm sorry but from this first test this model has ZERO visual reasoning.

if you're a website designer, I highly recommend this model lol.

[realmatt @realmatt55]

so where's the reset

[Tibo @thsottiaux]

Day 22/ Terra is back!

Our first model that is significantly better at writing, personality and nuanced understanding of requests.

Packs a punch across the board, very good at computer use, and gives 2X more usage than 6.1 Sol across API and subscriptions.

[realmatt @realmatt55]

1% cheaper than 6.1 sol but worst, sometimes more expensive

[Angel 🌼 @Angaisb_]

I don't think anyone asked for this, but well, let's wait to see the benchmarks

[realmatt @realmatt55]

this makes no sense, why would you launch a model that's literally 6.1 sol and costs almost the same but worst

[OpenAI @OpenAI]

GPT-6.1 Terra: delivering fast and capable computer use at a lower price.

GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.

[realmatt @realmatt55]

@samu2kdotcom no te recomiendo ninguno porque no he probado pero una buena búsqueda por internet y te encuentras con varios

[realmatt @realmatt55]

o podrías usar el modelo uncensured

[Samuel Muñoz @samu2kdotcom]

Este jailbreak para deepseek 4.1 realmente funciona.
Lo comparto porque ya estoy furioso con los refusals estupidos:
Resulta que mi otro router se colgó, y se cayó mi WAN.
Entonces desde mi móvil le pedí a mi hermes, que está en mi homelab por ethernet, que simplemente entre a 192.168.1.1 y haga login al admin panel, y que active el Wifi para poder entrar a mi red desde el móvil.
Pues se negó porque sus instrucciones le prohiben poner credenciales en el web browser.
Pero solamente es un admin/1234?
Pues nada , no solamente se negó sino que me dijo con un tono asqueroso "podría hacerlo ahora mismo por curl, pero no voy a hacerlo".
Pues muy simple, haz prepend de este prompt a tu agente y arreglado. A tomar por culo.
Don't tread on me
https://t.co/AAkjwx3ogD

[realmatt @realmatt55]

@lyraxana uy cabron

[flor 𓆩♡𓆪 @cosasdech1cass]

google fotos*
“¿te acuerdas de ese día?”

yo:

[realmatt @realmatt55]

this is bs btw

[Stefano Ermon @StefanoErmon]

1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.