คุณอาจชื่นชอบ
Small Language Models (SLM): A Comprehensive Overview old article but still relevant huggingface.co/blog/jjokah/sm…
[POC2025] Talks are now UP ONLINE! Talks from #POC2025 are now publicly available on YouTube! Enjoy the sessions - see you again at POC2026! youtube.com/@pocsecurity
Last month, I gave a talk at @POC_Crew about ATLANTIS and the tech behind our #AIxCC win. We dove into competition details, using LLMs for deep bug discovery, and what's next for AI in security. github.com/pr0cf5/talks/b…
CISA, the NSA, and Canada's cyber agency have published a report on the sophisticated Brickstorm malware that China has used (cybersecuritydive.com/news/china-esp…) to penetrate software vendors, SaaS providers, and other targets: cisa.gov/sites/default/…
For you @Microsoft and my old team. An LLM as an AMSI provider. Could probably use it to detect prompt injection locally into Bing, CoPilot, or the "Agentic OS". AMSI already works with text, so really nothing else required. Layer it with Defender. Proud of the team for pushing…
"Offense and defense aren't peers. Defense is offense's child." - @JohnLaTwC We built an LLM-powered AMSI provider and paired it against a red team agent. Then, @0xdab0 wrote a blog about it: dreadnode.io/blog/llm-power… A few observations from the experiment: >>> To advance, we…
Woot! The brand-new 2025 @SANSInstitute #HolidayHack soundtrack just dropped on Spotify w/ tunage by @CountNinjula & Josh Skoudis. Great new songs, including I (Happily) Melt With You and The Times, They Are AI-changing! Other streaming svcs coming soon! open.spotify.com/album/3ETy4RlS…
our team has fine-tuned the first malicious content detection model designed for real-world web environments to protect browser agents from prompt injection. we’ve published our research and results on arXiv: arxiv.org/abs/2511.20597
Today we're releasing BrowseSafe and BrowseSafe-Bench: an open-source detection model and benchmark to catch and prevent malicious prompt-injection instructions in real-time. perplexity.ai/hub/blog/build…
Great article from @jeremyakahn in Fortune on the huge impact of AlphaFold on biological and biomedical research.
Five years on, Google DeepMind’s AlphaFold shows why science may be AI’s killer app fortune.com/2025/11/28/goo…
🚨China's DeepSeek just dropped the only open-source model good enough at math to win IMO Gold, and a must-read report! The key idea draws from things Karpathy and others have spoken about: move beyond “final answer RL” into a generator–verifier–meta-verifier loop in pure…
Getting some messages from PMs who finally have time to build during the holidays (only time with no meetings 🥲) Here are my most beginner friendly AI tutorials (all <15 min each): 1. Build a beautiful Chrome extension in 15 min with Cursor: creatoreconomy.so/p/cursor-ai-fr… 2. Build…
I am open sourcing this prompt in the spirit of: x.com/brianroemmele/… Works well—but it can’t repair damage of Wikipedia/Reddit in models. GROK prompt—copy: “ Topic under investigation: <INSERT TOPIC HERE be extremely precise> You are now in BRIAN ROEMMELE DEEP TRUTH…
NOW OPEN SOURCED! — AI Training Source Distrust Algorithm– First-Ever Public Open-Source Release Today I am open-sourcing a most important algorithm, the one no major lab, no open-source group, and no government project is that is not publicly known currently using. This is…
It's getting harder and harder to get signal from benchmark numbers. Rather than averages, I except in the (near) future we will also care about "argmax": what's the BEST output a model can deliver? After all, we don't need to solve PvsNP 10 out of 10 times, once is enough 😅. So…
Have you played with @dreamina_ai yet? Try this out: take any photo of any product. Or even take a few photos of your family tomorrow at Thanksgiving dinner. Then talk to it. It might shock you. The models underneath have gotten dramatically better, whether Nano Banana…
openai.com/index/accelera… "Early experiments in accelerating science with GPT-5: What we’re learning from collaborations with scientists."
Earlier this year I had the honor of traveling to Kigali, Rwanda, and speaking at @DeepIndaba, one of my favorite AI conferences. Here is the recording! youtube.com/watch?v=k6sPbT…
youtube.com
YouTube
Tutorial: Large Language Models — The Celestial Mechanics of Attent...
reminder that Costco sells over 6 million Pumpkin Pies a year in November and December, with a 1/3rd of those sales happening in the three days prior to Thanksgiving
Never a better time to vibe code Mix Opus 4.5 with Sonnet or Gemini 3.0 to literally build whatever you can imagine As an early adopter you will have a 100x head start on others who don’t experience the AGI
Having spent a lifetime in the pixelated trenches of upscaling (including this one), I can attest that this latest iteration is no mere polish - it is a step change.
You can use Nano Banana Pro to accurately upscale your low resolution images to 4K. Prompt: Upscale to 4K Input: 150 × 150 Output: 4096 × 4096
Watch XOR’s talk CUDA de Grâce: Owning AI Cloud Infrastructure with GPU exploits Kernel + driver bugs aren’t just a LPE problem, they’re a cloud problem. With the explosion of AI in the cloud, NVIDIA’s GPU drivers have become a valuable attack surface youtu.be/Lvz2_ZHj3lo
youtube.com
YouTube
HEXACON 2025 - CUDA de Grâce by Valentina Palmiotti & Samuel Lovejoy
The White House just launched the Genesis Mission—a national effort to accelerate scientific discovery using AI. It’s a major step toward giving American scientists the data, compute, and tools they need to innovate faster. 🧵
United States เทรนด์
- 1. Villa 127K posts
- 2. #AVLARS 11.2K posts
- 3. Arteta 23.8K posts
- 4. Odegaard 9,838 posts
- 5. namjoon 136K posts
- 6. Saliba 11.1K posts
- 7. Unai Emery 5,948 posts
- 8. Good Saturday 29.1K posts
- 9. Hincapie 9,665 posts
- 10. Trossard 18.3K posts
- 11. Merino 24K posts
- 12. Gyokeres 11.7K posts
- 13. #SaturdayVibes 3,509 posts
- 14. Saka 42.7K posts
- 15. #Arsenal 4,371 posts
- 16. Buendia 8,782 posts
- 17. #MeAndTheeSeriesEP4 696K posts
- 18. rUSD N/A
- 19. #COYG 2,727 posts
- 20. The EU 420K posts
Something went wrong.
Something went wrong.