내가 좋아할 만한 콘텐츠
This was such a wonderful talk from @MillionInt that I’ll listen to it a second time during my morning walk in nature tomorrow. I need my chain of thought fully focused on these insights. ☺️
How GPT-5 thinks, with @OpenAI VP of Research @MillionInt 00:00 - Intro 01:01 - What Reasoning Actually Means in AI 02:32 - Chain of Thought: Models Thinking in Words 05:25 - How Models Decide How Long to Think 07:24 - Evolution from o1 to o3 to GPT-5 11:00 - The Road…
Excited to share two new #fuzzing technical reports from my group! 🔹 InsightQL: Advancing Human-Assisted Fuzzing with a Unified Code Database and Parameterized Query Interface 🔹 DynamiQ: Unlocking the Potential of Dynamic Task Allocation in Parallel Fuzzing (tool released!) 1/n
Oct 24, learn how SEC535 teaches red teamers to integrate AI without losing judgment with recon, phishing, malware & more. 🔗 sans.org/u/1CVK #AI #RedTeam #SEC535 #SecurityAutomation
Can we eliminate the C2 server entirely and create truly autonomous malware? On the Dreadnode blog, Principal Security Researcher @0xdab0 details how we developed an entirely local, C2-less malware that can autonomously discover and exploit one type of privilege escalation…
How to Build an AI Agent Step 1: Define Goal & Environment Set the agent's objective and operational context AI Agent Core Architecture: Input Layer: Perception Module 👁️: Receives input from sensors (camera, mic, etc.) Processing: Cognition & Reasoning Module 🧠⚙️: Analyzes…
The new version of my #BinaryNinja plugin Obfuscation Analysis (v1.2) adds recursive function inlining in the decompiler. It collapses call-heavy code into a single function; analysis, constant propagation, DCE and other analyses work across boundaries. github.com/mrphrazer/obfu…
Ilya Sustkever just posted this and deleted.. what did he see?
You just injected it as part of your prompt in the very beginning. Nothing fancy, you literally just use the word "Ultrathink" and it does it.
If you're in the Claude Code ecosystem, make sure you're using these two things: 1. Plan mode 2. Ultrathink Ultrathink is almost like Deep Research, but for thinking. Uses tons of tokens, but it gives extraordinary results. Plan mode with Sonnet 4.5 is nearly as good as…
If you want to take it one step further
This ChatGPT prompt replaced all AI newsletters for me
Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,…
If you've ever wondered how Large Language Models like ChatGPT work, this course is for you. It covers the data preparation, model training, and fine-tuning that has to happen before an LLM is ready to go. You'll learn about the whole training process from tokenizing raw text…
gpt-5 pro for verifying scientific literature:
GPT 5 Pro is extremely good in identifying serious gaps in published papers.
Steal this nano banan prompt ”Create a photo-style line drawing / ink sketch of the faces identical to the uploaded reference image — keep every facial feature, proportion, and expression exactly the same. Use blue and white ink tones with intricate, fine line detailing, drawn…
I've shared the full transcript of every agentic coding session from implementing the unobtrusive Ghostty updates and provided commentary alongside about my thinking and process. Total cost: $15.98 over 16 sessions. "Vibing a Non-Trivial Ghostty Feature" mitchellh.com/writing/non-tr…
Slides for the keynote I gave at the AI Security Forum in July proposing a framework for thinking about how to secure AI agents, and where my thesis is that we need to hybridize two currently separate technical disciplines; AI alignment and cybersecurity: docs.google.com/presentation/d…
Most security practitioners will tell you that the traditional network perimeter between internal trusted systems and the external untrusted internet disappeared with the emergence of client side exploits. It was replaced by coarse grained network isolation roughly between what…
Slides for the keynote I gave at the AI Security Forum in July proposing a framework for thinking about how to secure AI agents, and where my thesis is that we need to hybridize two currently separate technical disciplines; AI alignment and cybersecurity: docs.google.com/presentation/d…
updated the Lucid README and repo with new build instructions/details. you can now build the custom musl, --static-pie Bochs, and all the binaries you need with Docker and it copies them out of the container for you. github.com/h0mbre/Lucid
United States 트렌드
- 1. Ace Frehley 18.1K posts
- 2. John Bolton 91.4K posts
- 3. #NationalBreadDay 1,828 posts
- 4. Ashby 1,688 posts
- 5. Asheville 7,091 posts
- 6. RIP Spaceman N/A
- 7. Glasnow 2,213 posts
- 8. Steelers 27.9K posts
- 9. Space Ace N/A
- 10. Putin 190K posts
- 11. New York Groove N/A
- 12. Mitch 55.6K posts
- 13. Jake Bauers N/A
- 14. #KonamiWorldSeriesSweepstakes 2,124 posts
- 15. Smartmatic 6,162 posts
- 16. Cam Taylor N/A
- 17. Curt Cignetti 6,468 posts
- 18. Term 196K posts
- 19. Muncy 2,373 posts
- 20. Andrade 12.8K posts
Something went wrong.
Something went wrong.