Victor

Entusiasta programador

[Need More VRAM @needmorevram]

ISTA-DASLab just quietly dropped something very interesting:

Qwen3.8-Flash-Next GSQ-RCO Coder

A heavily compressed version of Qwen3.8 Flash Next focused on coding, agentic tool use, vision, and spatial reasoning.

https://t.co/KaulZY1sY1

[Beamsters @Beamsters1]

Sushi 🍣 quant focus much more on quality and size. This little Qwen3.8-Flash-Next-Sushi-2.6bpw made this pelican locally and it is the smallest sushi quant to date with only 43.8 GB weight, 64GB class enjoy!

Get it now! https://t.co/U0zFJh3xKV - spec below.

[jun @jun1228909]

今日地元の LLM ディスコで教わった、Qwen3.8 Flash Next を高速に動かすことに特化した謎の有志開発エンジン
https://t.co/uSrTfT0sfP

わいの 5090 + DDR5 128GB 環境 llama.cpp で 30tok/s くらいしか出てなかったのが
使うだけでいきなり 100tok/s 出て横転

llama.cpp 側の設定チューニングとか何もやってないから完全にはフェアではないかもしれんけど

しかし 100tok/s 出たらもう体感としてはオンラインモデルの応答くらいの速さで表示されてるようには見えるからなぁ
flash next 専用でオフライン開発環境を固めちゃうのもアリな気がしてきたな、、

[Victor @vLeocadiio]

@OrcaRouter E qual teremos um modelo red teaming ofensivo?

[Victor @vLeocadiio]

@OrcaRouter Where the link of download?

[Victor @vLeocadiio]

@Jossjsau32 @OpenAIDevs @OpenAI To com esse problema. Conseguiu resolver?

[Victor @vLeocadiio]

@OrcaRouter @phmt93126515 Me

[painn @painn_x]

stop burning your codex usage, do this asap

i switched to space bunny on Opencode Go and its Free for a week

it matches the performance of opus 5.5 for 1/10th price and almost 2x speed

1M context and multimodal, zero data retention

here's how to set it up in 3 mins:
→ install codex-router
→ add the openrouter key
→ pick space bunny in the model picker (or tell your agent to do it)

that's it. give this to your agent now

[Wildminder @wildmindai]

Qwen-Image-2.1 Object Remover Bbox LoRA

Turbo LoRA for bbox-guided object removal.
- strips selected objects while preserving textures, lighting, shadows, perspective
may artifact on complex backgrounds, overlaps, fine details, reflections.

https://t.co/2o92pUjGRd

[Victor @vLeocadiio]

@HuggingApps @ViggleAI Onde tá o link pra baixar??

[Victor @vLeocadiio]

@Oluwaphilemon1 Cadê o link para download do modelo?

[Dario Amodei @DarioAmodei]

Today we announced the Claude-led discovery of a molecular machine that we suspect could represent a new gene editing mechanism. Its precise function, biotechnological utility (if any), or level of significance is not yet clear, but at minimum it is work I would have been proud to do as a PhD student. The work was done mostly, though not entirely, by Claude: our life sciences team suggested a broad area of research, Claude read through the literature and a bunch of genome data and discovered something interesting, then Claude proposed experiments to verify the discovery and our team carried them out.

It’s easy to dismiss this as a one-off or curiosity, but we’ve repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years. In 2023 models struggled to do math at the level of an average high-school student. In 2024 they started to do well on math competitions for the best high-schoolers in the country, in 2025 they started to solve minor open problems, in early 2026 more significant open problems, and in late 2026 they are beginning to solve the top few open problems in all of mathematics. We believe AI for biology is on a similar exponential trend.

The main difference between biology and mathematics, of course, is that math can be done purely theoretically, while biology requires experimentation. Some have used this to draw the conclusion that AI’s utility in biology will be limited. We think this is wrong. As we’ve demonstrated today, humans can collaborate with AI to perform the experiments, validate key results in a few weeks and, if necessary, work with the AI to iterate on what they find. Eventually it may even be possible for Claude itself to safely perform the experiments by autonomously controlling lab equipment, with appropriate safeguards in place, but we aren’t doing that today (our lab is also a BSL1/BSL2 facility that doesn't handle materials dangerous to humans).

More broadly, biomedical advancement has many stages — from fundamental biology discoveries, to translational research, to drug discovery, clinical trials, and finally the actual delivery of medicines and health care to patients. We are also interested in these later stages, but even simply accelerating the first stage of fundamental biological discoveries has the potential to speed up and broaden the entire pipeline. Improving our understanding of biology and sharpening biologists’ tools can drive forward all of the later stages, for example by identifying new drug targets, finding new therapeutic modalities, allowing for more precise measurement, and speeding up the experimental loop which itself further accelerates our understanding of biology. This will not in itself speed up clinical trial times, but if it succeeds it could greatly increase the number of promising candidates that go into the pipeline — an increase in throughput even though latency remains.

In Machines of Loving Grace, I wrote about AI’s potential to “cure most diseases in 5-10 years” — a goal that sounds impossible, but one I believe is just barely possible if AI is applied to every stage of the pipeline. The first step is showing that AI can first help with, and then drive, biological discoveries.

Claude’s discovery is the latest in a line of related prior work that goes back decades, beginning with systems like CRISPR, and continuing with discoveries like the bridge recombinase and VIPR in the past few years. Recently, there has been heightened interest in systems based on reverse transcriptase (RT) enzymes, the enzyme underlying the system Claude identified. And most recently, a Stanford team working independently described a novel RT system with an associated non-coding array that is in some ways similar to the one Claude found, though they are distinct systems that evolved independently from each other. I believe that we’re at the very beginning of finding such systems and developing them into powerful tools for biotechnology.

I’m proud of the resources Anthropic has invested in accelerating the public benefits of AI through the life sciences, and we’re aiming both to grow our life sciences team and to work with other scientists to extend this approach to a broad range of problems. If you have a proposal for a research collaboration or are interested in joining our life sciences team, please reach out.

[Mirage @trymirage]

We rebuilt After Effects and Premiere for GPT-6 Astra and Sol.

Introducing Tesseract.

The video creative suite for your AI agent.

Give your agent direct access to our creative engine.
No more clicking through software built for people.

→ Edit footage, motion, and sound
→ Make precise changes with full creative control
→ Revise and render directly through your agent

And yes, this video was made with Tesseract.

Get the ChatGPT plugin.

[Victor @vLeocadiio]

@DanielHaupt @LabMike3D @aisearchio Onde acho esse Modelsamplingflux?

[Victor @vLeocadiio]

@elder_plinius What I can do with it?

[nihui @nihui]

Qwen-Image-2.1 full BF16 running with 2GB VRAM
ncnn/Vulkan for NVIDIA/AMD/Intel/Apple

No CUDA/PyTorch/Python
Portable executable
Text-to-image & Image editing
Up to 10 reference images
Transparent RGBA generation
Dynamic output resolution
Batch generation
https://t.co/mTiok6Qija

[Victor @vLeocadiio]

@sekizonyedi_ @mr_r0b0t Tem para 2x rtx 5060 ti 32gb?

[Guillermo Casaus @_guillecasaus]

Cloudflare ha liberado su propia skill para hacer auditorías de seguridad con IA.

Es la que utilizan internamente en la empresa para detectar vulnerabilidades y ahora cualquiera puede utilizarla con su agente.

Se llama Security Audit Skill y organiza cada auditoría en 6 fases:

✓ Reconocimiento de la arquitectura y superficies de ataque
✓ Búsqueda de vulnerabilidades por cobertura
✓ Validación independiente de cada hallazgo
✓ Generación de resultados estructurados
✓ Verificación final con agentes distintos
✓ Creación automática del informe de seguridad

Lo más interesante es que el agente que valida una vulnerabilidad nunca es el mismo que la encontró, reduciendo bastante los falsos positivos.

Es completamente gratuita, open-source y tiene licencia MIT.

Repo 👇

[王小波 @wanxubo]

想搞 “那种🥳🥳🥳” 盆友福利来了

昨天Qwen-Image-2.1 刚登顶开源之王的宝座

民间大神的 “去审查版” 已经来了😂

这次主要的改造是：
他们 直接对 Qwen-Image-2.1 的 Qwen3-VL-8B 文本��码器做了 Abliteration（拒绝方向消融）。

作者测试结果：

原版：100/100 次拒绝
去审查版：5/100 次拒绝
另一组独立测试：0/20 次拒绝

而且模型主体没被乱改：只修改 文本 Encoder，其他都没有变😁 尽量减少对正常生成能力的影响

目前已经提供 4 个版本：

BF16：约 17GB
NVFP4：5.87GB，面向 RTX 50/Blackwell
W4A8：5.88GB，适合 NVIDIA + ComfyUI
GGUF Q4_K_M：4.68GB，Mac / Apple Silicon 也能跑

主仓库（Heretic 去拒绝版 Text Encoder）
地址： https://t.co/yK3J73K8Fw
此外，GGUF 如果需要参考图编辑，还需额外约 1.08GB 的视觉��

大家快去试一试🥹，

‼️‼️‼️个人建议：装备之前让AI审查一下你的电脑适合哪个，不要装了带不动🥹🥹🥹😂

[toyxyz @toyxyz3]

https://t.co/YozDzzlqg4