Teddy
@teddyntran
Followers
1K
Following
77K
Media
525
Statuses
11K
AI Red Team ~~~ Informatics @sjsu Data Architecture @northropgrumman All opinions and views are my own.
Sunnyvale, CA
Joined April 2013
I’ve been telling people this a lot today: I enjoy so much working with people who care about what they are building and craftsmanship. It is a privilege to have a chance to work on something I’m passionate about, beyond making a living. I cherish it and don’t take it for
52
53
1K
I'm really excited to share my first research article related to hacking Google Gemini! https://t.co/e7GcJuGLCb
#bugSWAT #GoogleVRP
9
103
488
Agency > Intelligence I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce. Are you hiring for agency? Are
2K
9K
50K
This guy literally dropped the best life advice you’ll ever hear
673
21K
135K
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,
702
3K
24K
I gave a talk on Wednesday at the Bay Area AI Security Meetup about prompt injection, MCP security and the lethal trifecta. Here are the annotated slides from my presentation, including notes on my weird hobby of trying to coin or amplify new terms of art
simonwillison.net
I gave a talk on Wednesday at the Bay Area AI Security Meetup about prompt injection, the lethal trifecta and the challenges of securing systems that use MCP. It wasn’t …
23
57
409
just contributed my first "no, nothing from my end" in a work meeting
104
5K
92K
IT employees when you submit a ticket after you messaged them for help and they told you to submit a ticket
459
13K
235K
New Anthropic Research: A new set of evaluations for sabotage capabilities. As models gain more agentic abilities, we need to get smarter in how we monitor them. We’re publishing a new set of complex evaluations that test for sabotage—and sabotage-monitoring—capabilities.
59
228
2K
RT to help Simon raise awareness of prompt injection attacks in LLMs. Feels a bit like the wild west of early computing, with computer viruses (now = malicious prompts hiding in web data/tools), and not well developed defenses (antivirus, or a lot more developed kernel/user
If you use "AI agents" (LLMs that call tools) you need to be aware of the Lethal Trifecta Any time you combine access to private data with exposure to untrusted content and the ability to externally communicate an attacker can trick the system into stealing your data!
100
527
3K
> be apple > richest company in the world, every advantage imaginable > go all in on AI, make countless promises > get immediately lapped by everyone > 2 years into the race, nothing to show for it > give up, write a paper about how it's all fake and gay and doesn't matter anyway
396
2K
33K
this ending felt so inevitable that it almost ruins the enjoyment of the plot for being too predictable
3
2
178
The hustle bros & the harmony bros are both wrong. This is the right mindset: “You mentioned sacrifices, about all the things that you lost to become professional. I never had the feeling that I did a lot of sacrifices. I just was not able to do all the things that my friends
13
83
1K