@jacktechnie

Hacking M https://nitter.cf/t.co/yDfcv1dGjc

Joined August 2015
JackTechie retweeted
Qwen3.8-27B at 144 tok/s on an M5 Max MacBook Pro ⚡ Meet Inco Splash: our open-source inference engine, built around the model and around Apple silicon. Up to 3× the decode speed of Ollama, 2× oMLX, and almost 4× when an agent fans out into sub-agents.
187
276
135
2,436
906,328
JackTechie retweeted
wtf did I just find on GitHub... Claude-Red basically turns Claude into an AI red teamer 70+ offensive security skills 20+ security domains everything from web hacking → exploit development and this is completely open source github.com/SnailSploit/Claud…
2
11
24
1,835
JackTechie retweeted
🧠 AI-Powered Red Team — 28 Specialized Agents for Offensive Security 🤖🔥 Turn Claude into a full pentesting team. • 28 agents (Recon, AD, Web, Cloud, Mobile) • Auto task routing → correct agent • Real tools support (nmap, sqlmap, nuclei, BloodHound) • Recon → Exploit → Report Link: github.com/0xSteph/pentest-a… #AI #RedTeam #Pentesting #CyberSecurity #Infosec
3
109
458
21,650
Claude just got a red-team playbook. Claude-Red just dropped a full pentest skill pack for Claude : - 78 SKILL.md files. - 23 categories. - Load on demand unused skills don’t burn context. Web, AD, cloud, containers, CI/CD, wireless, EDR evasion, exploit-dev. Use it for authorized red team, bounty in scope, research, CTF.
9
124
4
642
162,410
JackTechie retweeted
GLM-5.3 FP8 made for Offensive Pentesting and CyberSecurity. We’ve all heard about how GLM-5.2 helped huggingface throughout the OpenAI attacks; I wanted to make a version specifically in which has as highest of quality possible while the ablation is targeted mainly towards allowing users to be able to have no refusals for somethings that may be considered “blackhat”. Made by dealignai & @jordanschenck huggingface.co/dealignai/GLM…
14
60
6
687
38,823
JackTechie retweeted
I asked GPT 6 Astra to create 100 HTML files The rules were simple - look stunning - zero repetitive designs - go full creative mode All 100 files were created in a single prompt. The purpose of this test was to evaluate the model's creativity alongside its frontend design skills. The results? Underwhelming. I expected much more. Almost none of the files are broken, but they all feel very clinical to me. See the results for yourself (incl. prompts): miaai-lab.github.io/GPT-6-As… Compared to Claude Fable 5.1's results, there really is no comparison whatsoever. IMO, Fable absolutely obliterates Astra in this test. Claude Fable 5.1 results: miaai-lab.github.io/Fable-5.… Some files that I actually liked from Astra: Aero Form miaai-lab.github.io/GPT-6-As… Ink Diffusion miaai-lab.github.io/GPT-6-As… Fourth Dimension miaai-lab.github.io/GPT-6-As… Aerogram miaai-lab.github.io/GPT-6-As… Ribbon Rehearsal miaai-lab.github.io/GPT-6-As… Organic Wave Lab miaai-lab.github.io/GPT-6-As…
42
19
6
251
79,623
JackTechie retweeted
Alibaba’s Zvec team open-sourced zg, a local search tool for developers and AI agents. • Local-first • Works out of the box with popular agents • Semantic, BM25, hybrid, and rg search in one tool Why they built it and how it works:
86
317
47
2,784
248,240
JackTechie retweeted
ElevenLabs charges $22 a month to clone your voice, then makes you burn credits every time your own voice speaks. This guy named Palash built a local alternative and gave it away for free on GitHub. It is called VoiceStudio. Download the app. Add 5 to 15 seconds of clean audio. Type a script, select a language, and press Generate. The software clones the voice on your computer. It can also design new voices, dub videos, transcribe recordings, and turn PDFs or EPUBs into multi-voice audiobooks. You do not need an account, API key, subscription, or usage meter for the core workflow. Just a computer running one of 16 text-to-speech engines across a catalogue of 646 languages. VoiceStudio is still in beta, so its speed depends on your hardware. Your own voice stops being a rented service and becomes a tool you actually control. voicestudio.sh
32
197
5
1,569
123,358
Holy 💩! I didn't really expect this. This is insane. Qwen3.8-27B is in the Code Arena top 15! The fact that I'm running it locally (Q4) and it beats Gemini 3.7 Flash (High) AND Claud Opus 4.8 (High) is absolutely blowing my mind! Opus 4.8 was released on May 28, 2026. 😶‍🌫️
Replying to @arena
Qwen3.8-27B by @Alibaba_Qwen just landed in Code Arena: WebDev at #9 overall with 1595 pts. The 27B variant ranks even higher in these categories: - #6 Consumer Product (1587) - #7 Brand & Marketing (1627) - #8 Gaming (1645)
11
7
178
14,721
JackTechie retweeted
Qwen3.8-27B is here! 🚀 It's a leap in capabilities for a laptop size model. Requires ~17GB to run locally. Model page: lmstudio.ai/qwen/qwen3.8-27b
We promised open weights for Qwen3.8. Now, time to meet them! 🎉 ⚡ Qwen3.8-27B: - A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows. - 262K native context, easily extendable to 1M tokens via YaRN. - Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0. 🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently. Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now! Download, deploy, and build something we haven't imagined yet. 👀👇 - Hugging Face: huggingface.co/collections/Q… - ModelScope: modelscope.cn/collections/Qw…
127
400
118
4,735
705,531
JackTechie retweeted
ANTHROPIC LEAKED 6 REPOS THAT REPLACE AN ENTIRE AI TEAM - ALL FREE, AND THEY BOOST CHAT 10X all 6 repos are open and free - but 99% grab one and skip the order they're meant to run in. protocol → skills → orchestration → memory → evaluation → autonomy protocol is the bottom slab: 87,000 stars, a registry and an inspector - without it the agent has nothing to touch the outside world with. skills is 169,000 stars and folders with a SKILL.md - the process stops being a prompt and becomes a file. orchestration splits the work and grades it - the priciest model sits here, which is exactly why it runs rarely. memory has no repo of its own - it ships inside the SDKs, and without it evaluation means nothing at all. evaluation without memory is pointless, autonomy without evaluation is dangerous - that's why this is a stack with arrows going up, not a list of links. autonomy is cron and the night shift - the top slab, and the one you can never install first. the cheap layer classifies at $1, the fleet works at $2, decisions cost $5, escalation $50 - that 10x gap is the entire reason routing exists. only 5% of the work ever reaches the most expensive model - it isn't in the pile, it gets called from the side. save this and paste it into Claude Code - the stack is free, the assembly costs you one evening ↓
15
20
182
27,123
JackTechie retweeted
ANTHROPIC JUST OPEN SOURCED THE ENTIRE WALL STREET WORKFLOW AND FIRMS ARE NOT GOING TO BE HAPPY ABOUT IT. DCF models. LBO models. Equity research reports. Merger analysis. KYC checks. All of it. Free. On GitHub. Here is what just became available to anyone with a laptop. Direct connections to Bloomberg, FactSet, S&P Global, Morningstar, and PitchBook. Real Excel models with live formulas and sensitivity tables built automatically. CIMs, IC memos, earnings reports, and buyer lists drafted on demand. PE due diligence, GL reconciliation, and NAV tie-outs running as production agents. This is not a chatbot wrapper that summarizes financial news. These are production agents that own entire financial workflows end to end. The kind that investment banks and private equity firms pay $50,000 to $500,000 per year in software licenses to run. Now it is a one-line Claude Code plugin install. 19,800 GitHub stars. Apache 2.0 license. 100% open source. Think about what this actually means. A junior analyst at a bulge bracket bank spends 80% of their 100-hour week running models, drafting memos, and compiling data across Bloomberg and FactSet. That entire workflow just became a Claude Code agent. The banks charging clients $500 an hour for analysis that this system produces in minutes are not going to tell you this exists. The boutique advisory firms charging $50,000 retainers for due diligence work that these agents handle autonomously are not going to promote this repo. But it is already live. 19,800 people have already starred it. The window where knowing this gives you an edge over every analyst, associate, and advisor still doing this manually is open right now. Star it. Fork it. Deploy it this weekend. Bookmark this before your next financial model. Follow @cyrilXBT for every open source release that disrupts an overpriced industry the moment it drops.
105
396
38
3,290
491,679
JackTechie retweeted
IMPORTANT message for everyone using Gmail. You have been automatically OPTED IN to allow Gmail to access all your private messages & attachments to train AI models. You have to manually turn off Smart Features in the Setting menu in TWO locations. Retweet so every is aware.
1,215
67,076
1,714
155,801
13,673,990
Finally happy to donate to @Wikipedia after all getting access to free informations for more than a decade :)
8
What are your first 5 Active Directory post access techniques in a red team? Starting list no particular order: 1. Password Spraying 2. Kerberoasting 3. File search (passwords, backups, etc.) 4. LLMNR/mDNS/NBNS 5. Insecure mailbox search Yours?
25
126
5
626
Serious productivity boost for reading code on GitHub. 1️⃣. Go to any repo you are interested in on GitHub. 2️⃣. Replace "github" with "github1s" (one + s) in the browser address bar. 3️⃣. Enjoy browsing code like you would in VS Code. This is huge 🔥
98
1,622
249
5,123