Tim Yardley (@timyardley) header
Tim Yardley (@timyardley) profile picture

Tim Yardley

@timyardley

husband, father, security researcher. All thoughts and opinions are my own.

 Champaign, Illinois
teapawt.org/search?q=%20from:@timyardley/exclude:replies
Joined February 2009
  • Tweets15K
  • Following1.6K
  • Followers1.4K
111Media
Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)Photo by Tim Yardley (@timyardley)
Stolen Thoughts AI Trace Work
48
MatrAIx: Simulating the World with 8.3B Persona Agents - YouTube youtube.com/watch?v=cNFkz9Wo…
MatrAIx: Simulating the World with 8.3B Persona Agents

MatrAIx: Simulating the World with 8.3B Persona Agents

MatrAIx is the simulated-user evaluation infrastructure for digital...

youtube.com
63
Oracle to Make Gemini Models Available to Thousands of Enterprise Applications Customers oracle.com/news/announcement…
Oracle to Make Gemini Models Available to Thousands of Enterprise Applications Customers

Oracle to Make Gemini Models Available to Thousands of Enterprise Applications Customers

Integrations can help Oracle Fusion Cloud Applications and Oracle NetSuite customers automate processes, accelerate decisions, and securely execute mission-critical work.

oracle.com
39
Critical OpenWrt DHCPv6 Flaw Could Let Unauthenticated Attackers Run Code as Root thehackernews.com/2026/07/cr…
Critical OpenWrt DHCPv6 Flaw Could Let Unauthenticated Attackers Run Code as Root

Critical OpenWrt DHCPv6 Flaw Could Let Unauthenticated Attackers Run Code as Root

OpenWrt 24.10.8 fixes CVE-2026-53921, a critical odhcpd stack overflow triggered by crafted DHCPv6 requests that could enable code execution.

thehackernews.com
32
IonStack part II: GhostLock, a stack-UAF that has existed in ALL Linux distributions for 15 years | Nebula Security nebusec.ai/research/ionstack…
IonStack part II: GhostLock, a stack-UAF that has existed in ALL Linux distributions for 15 years

IonStack part II: GhostLock, a stack-UAF that has existed in ALL Linux distributions for 15 years

GhostLock (CVE-2026-43499) is a Linux kernel vulnerability found by Nebu that exists in every major distribution since 2011. Triggering the bug does not require any special kernel config or privile...

nebusec.ai
97
Interesting read. krebsonsecurity.com/2026/07/…
21
GitHub Copilot: Sorry Dave, I can't do that harmful thing - unless you ask me in code theregister.com/security/202…
GitHub Copilot: Sorry Dave, I can't do that harmful thing - unless you ask me in code

GitHub Copilot: Sorry Dave, I can't do that harmful thing - unless you ask me in code

More fun with AI jailbreaks, this time at the workflow level

theregister.com
37
30 papers · The reading list Ilya Sutskever gave John Carmack 30papers.com/

30 papers · The reading list Ilya Sutskever gave John Carmack

A curated reading list of foundational AI and deep learning papers, hosted in full with plain-language explanations of difficult terms.

30papers.com
53
Prompt Injection as Role Confusion role-confusion.github.io/
Prompt Injection as Role Confusion

Prompt Injection as Role Confusion

LLMs can't tell who's speaking. We show they identify roles by writing style, not tags, and exploit this with CoT Forgery, injecting fake reasoning that models mistake for their own thoughts.

role-confusion.github.io
26
China-Linked Hackers Backdoored Linux Login Software to Hide for Nearly a Decade thehackernews.com/2026/06/ch…
China-Linked Hackers Backdoored Linux Login Software to Hide for Nearly a Decade

China-Linked Hackers Backdoored Linux Login Software to Hide for Nearly a Decade

Sygnia says Velvet Ant modified Linux PAM and OpenSSH components to steal credentials and maintain stealthy access since 2016.

thehackernews.com
42
You can’t afford to lead agentic engineering from the sidelines blog.kilo.ai/p/you-cant-affo…
You can’t afford to lead agentic engineering from the sidelines

You can’t afford to lead agentic engineering from the sidelines

AI is changing the work faster than most operating models can keep up. Leaders need firsthand experience before they start redesigning the process.

blog.kilo.ai
15
Cheap code means formal verification is reasonable now — Antfly Blog antfly.io/blog/agent-formal-…
32
DeepSeek-V4-Flash means LLM steering is interesting again seangoedecke.com/steering-ve…

DeepSeek-V4-Flash means LLM steering is interesting again

Ever since Golden Gate Claude I’ve been fascinated with “steering”: the idea that you can guide LLM outputs by directly manipulating the activations of the…

seangoedecke.com
53
Interaction Models: A Scalable Approach to Human-AI Collaboration - Thinking Machines Lab thinkingmachines.ai/blog/int…
Interaction Models: A Scalable Approach to Human-AI Collaboration

Interaction Models: A Scalable Approach to Human-AI Collaboration

Interaction models move beyond turn-based AI interfaces by handling multimodal, real-time collaboration natively across audio, video, and text.

thinkingmachines.ai
1
56
Humanity faces an uncertain fate as experts brace for superintelligent AI telegraph.co.uk/business/202…
Humanity faces an uncertain fate as experts brace for superintelligent AI

Humanity faces an uncertain fate as experts brace for superintelligent AI

Artificial intelligence industry claims looming ‘singularity’ will change everything

telegraph.co.uk
21
The AWS MCP Server is now generally available | AWS News Blog aws.amazon.com/blogs/aws/the…
The AWS MCP Server is now generally available | Amazon Web Services

The AWS MCP Server is now generally available | Amazon Web Services

AWS announces the general availability of the AWS MCP Server, a managed remote Model Context Protocol (MCP) server that gives AI agents and coding assistants secure, authenticated access to all AWS...

aws.amazon.com
33
Behind the Scenes Hardening Firefox with Claude Mythos Preview - Mozilla Hacks - the Web developer blog -- hacks.mozilla.org/2026/05/be…
Behind the Scenes Hardening Firefox with Claude Mythos Preview – Mozilla Hacks - the Web developer...

Behind the Scenes Hardening Firefox with Claude Mythos Preview – Mozilla Hacks - the Web developer...

New details about what we found, and how agentic harnesses are now able to reproduce real bugs and dismiss false positives.

hacks.mozilla.org
1
49
A tale of hope and positivity not despair -- Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150 - Ars Technica arstechnica.com/ai/2026/04/m…
Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150

Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150

CTO says new AI model is "every bit as capable" as world's best security researchers.

arstechnica.com
37
MAD Bugs: Even "cat readme.txt" is not safe - Calif -- blog.calif.io/p/mad-bugs-eve…
MAD Bugs: Even "cat readme.txt" is not safe

MAD Bugs: Even "cat readme.txt" is not safe

Turning "cat readme.txt" into arbitrary code execution in iTerm2.

blog.calif.io
1
50
Assessing Claude Mythos Preview’s cybersecurity capabilities | red.anthropic.com red.anthropic.com/2026/mytho…
Frontier Red Team

Frontier Red Team

Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

anthropic.com
26
Load more