Recordings from [un]prompted.au 2026 at ILUMINA, Sydney, 18 to 19 September. Every session is listed in program order; recordings appear here as they are published to YouTube.
21 of 24 sessions published. Press play to load a video (nothing from YouTube loads until you do).
Get notified when the next talks drop. One email when new recordings land, and first word on the next [un]prompted.
The 45-minute headline sessions that opened and anchored each half-day.
Paul McMillan · OpenAI
You must budget tokens for model safety. The major model providers do it, so should you. This talk will discuss a variety of approaches, including their downsides, and give you… Read more →
Valentina Palmiotti · IBM X-Force
A technical look into BinChomp, an autonomous N-Day pipeline focused on the Windows Kernel. Read more →
Shane Huntley · Google Threat Intelligence Group
Succeeding as a threat analyst in the Google Threat Intelligence Group requires building deep domain-specific intuition to navigate our massive internal systems and the unique… Read more →
Dion Blazakis · Calif Research
A first-hand account of working with AI agents to turn two macOS vulnerabilities into controlled kernel read/write and root, and a look at where the models helped, where they did… Read more →
Thirty-minute deep dives: original research, tooling, and results from the field.
Saurabh Kumar · Microsoft Edge
Zenshin, an LLM auditor that only hands over reproduced WebKit bugs: call-graph reachability, cross-model debate, and an ASAN binary as the final judge. Eight WebKit CVEs so far… Read more →
Tristan Steele
Letting an LLM agent loose on a live CAN bus, using a consumer portable power station as the target. The harness, permission gates, instrumentation and physical isolation it took… Read more →
John McIntosh · Clearseclabs
Two Windows bugs found with a frontier model, the first in just over an hour, and the human direction it took to get there. Then the same hunt replayed on open-weight local models… Read more →
Jasper van Woudenberg · Keysight Device Security Testing
Voltage-glitching a commercial NPU to make an edge-AI camera see phantoms, or nothing at all, with LLM agents left to search the fault space overnight. Why neural accelerators are… Read more →
Dominic Marks · Slack
Inside the event-driven agentic detection and response system Slack's SecOps team has run in production since April 2026, told through a no-notice red team exercise it caught… Read more →
Sylvia Xu & Songjiu Liu · TikTok
Traditional Static Application Security Testing (SAST) is highly effective at identifying syntax-based vulnerabilities like SQL injection or Buffer Overflows, but completely blind… Read more →
Anthony Lai & Alan Ho · VXRL / Ken Wong · HKUST
A three-month study adjudicating 630 agent-generated vulnerability findings against the code at the reported line. In their data, precision depended more on method than model, the… Read more →
Rich Smith · Mindgard
Prompt injection gets treated as a text problem, something that arrives over an API. But more and more of what an LLM reads into context starts in the real world: a swiped badge… Read more →
Seunghyun Lee · Carnegie Mellon University
Exploitation is a ladder, not a coin flip. ExploitBench grades LLM agents across 16 capability flags against real V8, from first crash to arbitrary code execution, and shows where… Read more →
Wai Kin Wong, Dongwei Xiao & Shuai Wang · HKUST
Fuzzers reach the vulnerable code and still miss the bug because coverage tracks lines, not state. StateLens has an LLM identify security-relevant internal states offline and lets… Read more →
Ash Fox · Google
We have seen an incredible increase in AI's offensive capabilities. Presumably, that is why we are all here! AI is putting sophisticated hacking capabilities into more hands… Read more →
Ben Gittins · Zepto / Cole Cornford · Galah Cyber
The Australi.ai co-founders on how agentic AI is commoditising the artisanal skills cyber careers were built on, and why Australia should build that capability domestically rather… Read more →
Thomas Roccia · SecurityBreak
AI is transforming how we work, how we build software, and how we interact with technology. But it is also creating an entirely new ecosystem that few organizations understand or… Read more →
Ariel Herbert-Voss · RunSybil
When an AI-generated result looks convincing at first glance but falls apart under closer inspection, it can signal a collision in the model’s latent space, where concepts… Read more →
Dmitrijs Trizna · AISLE
Once a security fix becomes public, the diff becomes a roadmap to the vulnerability. An attacker can reconstruct the missing invariant and start building a trigger before a CVE or… Read more →
Danila Sashchenko · TikTok
This talk presents an AI-driven system for discovering and exploiting vulnerabilities in ElectronJS applications. By combining source–sink analysis with real-time harness… Read more →
Fifteen-minute dispatches: one idea, sharply made.
Pete Werner · Penumbra AI
AI can dramatically reduce the time it takes to complete a task, but verifying the result often takes just as long as doing it yourself. This talk explores how to build AI… Read more →
Rahat Masood · UNSW Sydney
Capture the Narrative is a CTF-style massively multiplayer wargame in which 288 Australian university students ran LLM-powered bot campaigns to swing a simulated election. The… Read more →
Casey Ellis · Tall Poppy Group / disclose.io
Applied AI, self-improving loops, and a security tool that gets better every time it fails. I keep telling people to build the tooling instead of being the tooling — use AI to… Read more →
Brendan Dolan-Gavitt · XBOW
Two case studies in LLM agents ripping through obscurity: reverse-engineering the currency detectors in photocopiers and image editors to build an unphotocopiable cat, and… Read more →
Stay up to date on what unprompted.au is up to next
This list is where the next [un]prompted will be announced first: dates, the call for papers and new recordings, straight to your inbox. Low volume. Unsubscribe anytime.