Roland Polczer

mindfulness • singularity • consciousness • AI • MMA • Chili

[Roland Polczer @diegorivera]

@Alibaba_Qwen Great job team, can't wait for the 27b. Do you guys plan to release Qwen Music weights too? That would be awesome.

[Roland Polczer @diegorivera]

Those who see the gathering storm are sounding the alarm. Will we listen? https://t.co/EtFXmZeuDJ

[Roland Polczer @diegorivera]

Here is a brief summary of recent events that are worth knowing about before *** hits the fan.

Events from Huggingface's point of view: https://t.co/3QkvBFz38x

OpenAI's notice:
https://t.co/ctNiT1cA2v

TLDR;

[Roland Polczer @diegorivera]

We have just reached the next milestone in our journey from reality to sci-fi movie.
https://t.co/o1YEWz6jtU

[Roland Polczer @diegorivera]

@thsottiaux I love how Sol can do long horizon autonomous tasks without hitting the rate limits. I can get things quickly done with it and the code quality is amazing on high.

[Roland Polczer @diegorivera]

DeepSeek is a godsend

[Roland Polczer @diegorivera]

LLMs gave AI reasoning. Agents give it action. That shift is already happening — and it's moving faster than LLMs did.

https://t.co/cTprrfKW7k — the seismograph is live. 43 watch items, daily signals, live threat level.

[Roland Polczer @diegorivera]

The ground is moving. Don't ignore the early signals. 🦞
https://t.co/cTprrfKW7k

[Roland Polczer @diegorivera]

The writing is on the wall. Agents might be the next black swan, a low-confidence, high-risk scenario worth preparing for.

https://t.co/1kubW8Fcf4

[Nikita Bier @nikitabier]

Prediction: In less than 90 days, all channels that we thought were safe from spam & automation will be so flooded that they will no longer be usable in any functional sense: iMessage, phone calls, Gmail.

And we will have no way to stop it.

[Roland Polczer @diegorivera]

@mattshumer_ 100% agreed. This is happening now. Brace for impact.

[Roland Polczer @diegorivera]

And then, Ava was born. #clawbot

[Roland Polczer @diegorivera]

We are witnessing history being made. Let's take a moment to look around and remember this day. Welcome to the era of the agents. https://t.co/13vD2gQfmE

[Roland Polczer @diegorivera]

For the record: 21-Nov-2025, ARC-AGI-2✅

[Roland Polczer @diegorivera]

Gemini 3 Pro is out, and it does not disappoint! Here's a quick game it created when I asked it to generate a unique, engagement-optimised game. https://t.co/unEvvbqh5R

[Roland Polczer @diegorivera]

Codex is something

[Roland Polczer @diegorivera]

This is the first time I see language mixing with GPT-5-Thinking. Getting qwq vibes.

[Roland Polczer @diegorivera]

gpt-5-high is the first model I saw that could solve this puzzle

[Roland Polczer @diegorivera]

@AuroraSkye21259 QA is doing some weird mistakes, like forgetting consistently the same curly bracket, renaming variables or updating parts of the code that was not asked. 2.5 Pro or Claude 3.7 are way better on that, howevert QA was able to solve issues that both sota models failed to tackle

[Roland Polczer @diegorivera]

Llama 4's release is completely overshadowed by a new misterious model called "Quasar Alphba" at Openrouter that is faster and better in coding than the current SOTA models like Gemini Pro 2.5 and Sonnet 3.7.

[Roland Polczer @diegorivera]

@OfficialLoganK Well, it's not code, but I ran @Sama's prompt on it about AI and grief - https://t.co/nurOsZtZRQ