Joe Lucas (@josephtlucas) 's Twitter Profile
Joe Lucas

@josephtlucas

developer.nvidia.com/blog/tag/ai-re…

ID: 1171003245222932483

linkhttp://josephtlucas.com calendar_today09-09-2019 10:11:58

2,2K Tweet

480 Followers

934 Following

Vik Paruchuri (@vikparuchuri) 's Twitter Profile Photo

Parsing PDFs has slowly driven me insane over the last year. Here are 8 weird edge cases to show you why PDF parsing isn't an easy problem. 🧵

Parsing PDFs has slowly driven me insane over the last year.  Here are 8 weird edge cases to show you why PDF parsing isn't an easy problem. 🧵
Wil Gibbs (@cl4sm) 's Twitter Profile Photo

While playing DEF CON CTF Finals with Shellphish I managed to solve the ICO challenge using LLMs (GPT5 + Cursor) and almost no human intervention. You can read how I did it here! wilgibbs.com/blog/defcon-fi…

LaurieWired (@lauriewired) 's Twitter Profile Photo

The West has a blindspot when it comes to alternative CPU designs. We’re so entrenched in the usual x86, ARM, RISC-V world, that most people have no idea what’s happening over in China. LoongArch is a fully independent ISA that’s sorta MIPS…sorta RISC-V…and sorta x87!

The West has a blindspot when it comes to alternative CPU designs.
 
We’re so entrenched in the usual x86, ARM, RISC-V world, that most people have no idea what’s happening over in China.
 
LoongArch is a fully independent ISA that’s sorta MIPS…sorta RISC-V…and sorta x87!
Matt Mazur (@mhmazur) 's Twitter Profile Photo

Sebastien Bubeck I'll throw this out there: I have a site, Emergent Mind, where I've built a platform for analyzing arXiv papers at scale. I also have over $100K in Azure credits and can use those credits on GPT-5. It would not be hard for me to run thousands of papers in math, AI/ML, etc

Suha (@suhackerr) 's Twitter Profile Photo

New post and tool! Attackers can break production AI systems by using image scaling to hide multi-modal prompt injections from users. 🧵for more info on what broke, how this works, and our new tool to try this out yourself

Joe Lucas (@josephtlucas) 's Twitter Profile Photo

You must augment LLM systems with info retrieval; they won’t excel as the IR system itself (context limitations, poisoning, hallucination, etc). Especially when you start engineering for cost, throughput, etc. AI apps make search API calls. Feels like Google/search is safe.

Joe Lucas (@josephtlucas) 's Twitter Profile Photo

"Building internal-only tools! Where you don't need to worry about security, scalability, malicious usage." Red teams everywhere: 🤪

Chaowei Xiao (@chaoweix) 's Twitter Profile Photo

We are looking for interns to work with us on VLA, reasoning model and security related topics. Please apply via the following link

Joe Lucas (@josephtlucas) 's Twitter Profile Photo

Jensen didn’t mention all of our published AI security research?! I guess we need to get into his script generation prompt. Gotta be a RAG system somewhere… Seriously though: it’s great how much we’re encouraged to share with the ecosystem at NV.

Robin (@digininja) 's Twitter Profile Photo

Can anyone point me at a good example of a multi-arch docker image that is built through GitHub actions? The image description has gone missing since it was changed to multi-arch and I want to find another repo who has it working so I can see what fix is needed.