Wire Cat

@h4x0r_dz @C2IRIS Bianry ninja license gives you updates for a year and once it’s expired you can renew or keep using the same version

@C2IRIS Why not ida ?!

Finally bought Binary Ninja today Absolutely tremendous product Obviously built by total craftsmen Well worth the $3,000

@TeksEdge MiniMax m2.7 is not open weight it’s proprietary

📊 Check out the latest PinchBench, Clawdbot Benchmark, Leaderboard. Open Source models dominate. New model MiniMax M2.7 is in the top 10. Clawdbots 🦞 rejoice. Low cost 🧠

@Ex0byt @MichaelDell 90k is a hefty sum

Take my money @MichaelDell

@init1security > OPSEC > 8 years ago

227 tips dedicated to red teaming, OPSEC, infrastructure, payloads, etc. github.com/vysecurity/Red…

@DarkWebInformer @grok does payment through Apple Pay prevent this? Pretty sure you can’t reuse the same “relay” more than once of the handshake

‼️ A threat actor is selling a NFCRipper tool which allegedly enables NFC relay, card capture, session cloning, and POS/ATM CVM bypass functionalities. They claim it can capture, replay, and manage card data through a centralized web panel for research and testing purposes. nfcripper[.]su

@CattardSlim @grok is this true, if not what is an estimate

97% of Elon's predictions have been wrong.

@sama What’s your favourite colour and what makes you think the DoW won’t make the same request of openai like they did with anthropic (considering they had no issues in the beginning)

I'd like to answer questions about our work with the DoW and our thinking over the past few days. Please AMA.

@ZoomerXavier @PathofTheDollar @KobeissiLetter Remove the high IQ from your bio. Thanks

BREAKING: President Trump orders ALL Federal agencies in the US Government to immediately stop using Anthropic's technology.

@yagilb Hopefully it’s not some bullshit like ollamas cloud thing

Something we've been working on for a long time is coming today ✨

@royjossfolk @ivanfioravanti Larger open source models like kimi 2.5, glm5 are good to use for software dev if you have the hardware. If not proprietary models are the way to go for now.

I know it is just a benchmark but don't you think these models have this type of application memorized by now? I use local models every day and while they can pop out snake games like nothing, they struggle with building real production software without supervision of a cloud model. If you try to build something complex or intricate, you can see there is still a ways to go. Likely 1-3 more iterations. Wondering your thoughts and what is the most complex software you've built with a local model?

MLX + OpenCode + Qwen3.5-122B-A10B-4bit on M3 Ultra created a great snake game! Work zero-shot. Video clearly in super fast mode during generation. I generated the prompt using Grok 4.20, it's in the article.

@Ominousind @SaiNemani1 Influencer status

@SaiNemani1 u just prompjacked me TF!!!

One-shotted a flight simulator using Gemini 3.1 Pro (Preview) Genuinely impressive! Prompt: "create a website of HTML, CSS and JS all in on code page for a complete flight simulator!" More tests otw, follow me!

@_RastaMouse Ai is so productive it’s made a bunch of people unemployed reducing their working weeks to 0 days!!

If AI is so productive, why are we not just working 1-2 day weeks?

@Ominousind There’ll probably be another model release this week, given the rate of open source and proprietary models coming out so far this year.

@wire_catt I wish I had seen this idea two hours ago hahaha good test to keep in mind

Qwen 3.5 SVG Test Results are looking good

@Alibaba_Qwen @Ominousind time to cook 🧑‍🍳

🚀 Qwen3.5-397B-A17B is here: The first open-weight model in the Qwen3.5 series. 🖼️Native multimodal. Trained for real-world agents. ✨Powered by hybrid linear attention + sparse MoE and large-scale RL environment scaling. ⚡8.6x–19.0x decoding throughput vs Qwen3-Max 🌍201 languages & dialects 📜Apache2.0 licensed 🔗Dive in: GitHub: github.com/QwenLM/Qwen3.5 Chat: chat.qwen.ai API：modelstudio.console.alibabacloud.com/ap-southeast-1… Qwen Code: github.com/QwenLM/qwen-co… Hugging Face: huggingface.co/collections/Qw… ModelScope: modelscope.cn/collections/Qw… blog: qwen.ai/blog?id=qwen3.5

@ivanfioravanti Wow can I get a copy of the html for the far right one?

Eulerian Fluid simulation test! Zero-shot! Opus 4.6 vs GPT-5.3 vs Gemini 3 Deep Think! My personal preference: 🥇 Gemini 3 Deep Think (really strong!) 🥈 Opus 4.6 🥉 GPT 5.3 High

@5mukx Not really, only works for exe’s in disk. And requires you to ctrl+click so this is nothing crazy. Unless the CVE is something else

This is nuts ... x.com/vxunderground/…

@Ominousind Will you make a video about this tiiny box? I’m interested in how it actually performs

@TheAhmadOsman @LLMJunky @ivanfioravanti Yes, however considering pricing makes a studio more attractive. In addition you can also do batching on the studio. youtu.be/E-8KJpUFalM?si… 9:10 in he sends off two prompts. But there is no replacing GPUs with m chips and unified memory any time soon.

@wire_catt @LLMJunky @ivanfioravanti It’s worth noting that with Unified Memory you’re effectively limited to a single request at ~25 tokens per second. By contrast, GPUs handle hundreds of parallel requests at thousands of tokens per second. Memory is not everything.

No M5 Pro/Max today 😢 We have to wait.

@LLMJunky @TheAhmadOsman @ivanfioravanti Yeah seems like gpus might be better in terms of resell. Main thing was that you have unified memory so it’d be more accurate to say it’s more versatile in the studio. Rtx6000 pro is 96gb would still require 11 gpus around $10k each, studio is same price at 512gb.

Thanks! a few counterpoints though: i dont think you can argue better resell. have you seen the price and availability of GPUs? 3090s are still selling for over $1000... Multiuse is not as great of an argument either, because GPUs require computers just like your unified memory is useless on its own. GPUs live in multi use compuers. It's a wash. BUT Mac is a great dev ecosystem. M5 Ultra could cost boo-koo bucks. RAM prices are skyrocketing. Too early to tell. 4x performance seems unlikely but who knows. That would be special if true. MaxQ RTX6000s are 300w and can be run at 250. Not bad. But fair point. Plus...CUDA. That said...I still want as many Mac studios as I can get my hands on. But I dont think I'm SUPER interested in 20-24 toks/s right now. M5 will have to be great. Thanks for the comment. Happy to hear any counter counterpoints.