Went through this and wrote a rough timeline of this video in case it helps anyone: 0:00-1:33 About Me 1:33-3:15 Fuzzing Overview 3:15-3:58 Fuzzing Explained w Diagrams 3:58-6:08 Limitations of Fuzzing 6:08-6:48 Brief Explanation of Harnesses 6:48-12:25 Explanation of example fuzzing programs available today. 12:25-21:20 Coverage Guided Fuzzing and how this can improve code coverage by 20-30% 21:23-26:44 Crash Amplification using PageHeap/ASAN, adding hooks to program to find interesting cases. 26:44-27:47 Performance and why linear scaling is important, but why good input and targets are more important 27:47-30:24 Snapshot fuzzing and how it helps with deterministic fuzzing and performance. 30:24-32:11 Discussion of Brandon's career in security, explanation of symbolic fuzzing. 32:11-41:01 Explanation of Brandon's research into Chrome sandbox from a fuzzing and debugging perspective. Discussion of fruitsnack tool (custom allocator for Chrome). 41:01-45:47 Moving on to skeletor, VT-x hypervisor development, moving on to fuzzing entire systems in a system (Falkervisor/Brownie). 45:47-49:12 Explanation of Grilled Cheese improvements. 49:12-51:10 Experiments with QEMU and why it wasn't great for fuzzing. 51:10-52:31 Creating of Salad, a MIPS32 emulator. Adding of register coverage, why it wasn't useful. 52:31-52:48 Moving tools over to Rust 52:48-58:50 Discussion of work on Vectorized Emulation starting from 2017 onwards. 58:50-1:00:06 Discussion of modern day work that Brandon is working on such as wafflecone. 1:00:06-1:07:07 Questions
Went through this and wrote a rough timeline of this video in case it helps anyone:
0:00-1:33 About Me
1:33-3:15 Fuzzing Overview
3:15-3:58 Fuzzing Explained w Diagrams
3:58-6:08 Limitations of Fuzzing
6:08-6:48 Brief Explanation of Harnesses
6:48-12:25 Explanation of example fuzzing programs available today.
12:25-21:20 Coverage Guided Fuzzing and how this can improve code coverage by 20-30%
21:23-26:44 Crash Amplification using PageHeap/ASAN, adding hooks to program to find interesting cases.
26:44-27:47 Performance and why linear scaling is important, but why good input and targets are more important
27:47-30:24 Snapshot fuzzing and how it helps with deterministic fuzzing and performance.
30:24-32:11 Discussion of Brandon's career in security, explanation of symbolic fuzzing.
32:11-41:01 Explanation of Brandon's research into Chrome sandbox from a fuzzing and debugging perspective. Discussion of fruitsnack tool (custom allocator for Chrome).
41:01-45:47 Moving on to skeletor, VT-x hypervisor development, moving on to fuzzing entire systems in a system (Falkervisor/Brownie).
45:47-49:12 Explanation of Grilled Cheese improvements.
49:12-51:10 Experiments with QEMU and why it wasn't great for fuzzing.
51:10-52:31 Creating of Salad, a MIPS32 emulator. Adding of register coverage, why it wasn't useful.
52:31-52:48 Moving tools over to Rust
52:48-58:50 Discussion of work on Vectorized Emulation starting from 2017 onwards.
58:50-1:00:06 Discussion of modern day work that Brandon is working on such as wafflecone.
1:00:06-1:07:07 Questions
Thanks
Seriously, how did those kids just walk away from one of the most interesting presentations ever before it ends?
GOD DAMN YOU ARE AMAZING
LIVEOVERFLOW WAS RIGHT YOU ARE A GOD