
interested in building trustworthy systems. CTO and co-founder of @gapfruit
Joined September 2019
- Tweets154
- Following69
- Followers87
- Likes1.4K
Pinned Tweet
My @wasm_io talk in Barcelona is now online. It shows WASM in OT environments and how a "capabilities all the way down" systems design enables resilient and scalable industrial infrastructure. Witnessing how WASM is gaining massive momentum at all SW stacks is very exciting.
🧩 Capabilities all the way down - an OS for WASM by @sidhussmann @ Wasm I/O 2024
▶️ Video: youtu.be/Jl67bMS1848
🔗 Slides: 2024.wasmio.tech/slides/capa… #wasmio24
Sid Hussmann retweeted
Announcing the Boot Security Mastery Conference Review Board:
• Piotr Król (@pietrushnic) - Chair
• Xeno Kovah (@DarkMentorLLC) - Member
• Sid Hussmann (@SidHussmann) - Member
They'll review and rank your BSMConf CfP submissions.
Sid Hussmann retweeted
In 20 years, vibe coders will look at the Linux kernel repo the way we look at the pyramids. In awe, unable to imagine how they managed to drag all those giant stones and pile them up in the middle of the desert.
Sid Hussmann retweeted
The "AI vs. AI" Fallacy.
The narrative that you need "AI defense" to stop "AI attackers" is marketing.
It frames security as a reactive arms race rather than a structural discipline.
If your posture relies on a tool guessing what the bad guys will do next, you’ve already lost the architectural battle.
That being said. Speed in response is important.
So using AI tools in your processes that can help speed up your team or having an agent collect context and even help react faster is a win.
There is a myriad of other ways AI tools can help a security team, but saying that the hackers are using AI so we MUST do so too isn't actually a thing I subscribe to for that reason.
Sid Hussmann retweeted
There will not be the time nor resources to fix all of the vulnerabilities uncovered in the near future as more AI bug finders get to work. What should defenders do now?
The irony to me is that they should do what they should have already been doing: assuming those bugs existed.
Sid Hussmann retweeted
If the idea is that somehow, at ecosystem-of-ecosystems scale, we are ever going to reliably get vulnerabilities in important infrastructure patched everywhere or almost everywhere before they can be exploited... well, I have great trouble finding that even remotely plausible.
Sid Hussmann retweeted
fixedAiSlop redd.it/1mopapc
Quite on point:
> ALL hypervisors are software. There is no such thing as a "hardware hypervisor", despite arrogant people spreading this marketing myth very hard.
As a rare and precious hypervisor VM escaper authority, this RETbleed-on-VM research from Google is cool.
Few realize: most Speculative Execution & Side-Channel exploits affect hypervisors *automatically* – even if it isn't mentioned in the paper.
Why?
They cross software privilege boundaries *from below*. That’s every boundary – including the hypervisor’s.
Today’s Confidential Compute / Trusted Virtualization technology adds *more* privilege boundaries → more attack surface.
And yes, ALL hypervisors are software. There is no such thing as a "hardware hypervisor", despite arrogant people spreading this marketing myth very hard.
Every modern hypervisor rides hardware virtualization frameworks to speed up its code – which adds *zero* security, only easier issue management, and a slightly reduced attack surface. Which clearly isn't helping to stop the exploits.
This work doesn't add anything new – it simply demonstrates a real impact of hardware side channel exploits on virtualization technology, which the original paper missed.
bughunters.google.com/blog/6…
Sid Hussmann retweeted
🚨Yesterday marked a notable moment as CISA added, for the first time, an AMI MegaRAC BMC vulnerability (CVE-2024-54085) to its Known Exploited Vulnerabilities (KEV) Catalog. Baseboard Management Controllers (BMCs) have long been attractive targets for attackers, yet unfortunately, security practices surrounding these critical components have evolved little over the years.
Vulnerabilities like CVE-2024-54085 remain widespread and alarmingly easy to exploit, echoing the early days of application security in the 90s when secure development lifecycle practices were virtually non-existent. Additionally, the choice of Lua as a scripting language raises concerns, as it lacks robust support from mainstream security tools, creating significant blind spots.
AMI MegaRAC BMC technology underpins critical infrastructure used by numerous enterprise vendors, forming the backbone of today's data center and cloud computing environments. CVE-2024-54085 is impacting the entire ecosystem rather than just a single vendor.
The @Binarly_io team has consistently highlighted these hidden dangers within BMC security, urging the industry multiple times:
✸ CVE-2024-36435 Deep-Dive: The Year’s Most Critical BMC Security Flaw
binarly.io/blog/cve-2024-364…
✸ Old But Gold: The Underestimated Potency of Decades-Old Attacks on BMC Security
binarly.io/blog/old-but-gold…
✸ Binarly REsearch Uncovers Major Vulnerabilities in Supermicro BMCs
binarly.io/blog/binarly-rese…
🛡️ We added Fortinet FortiOS, D-Link DIR-859 Router, & AMI MegaRAC SPx vulnerabilities CVE-2019-6693, CVE-2024-0769, & CVE-2024-54085 to our Known Exploited Vulnerabilities Catalog. Visit go.dhs.gov/Z3Q & apply mitigations to protect against cyberattacks. #Cybersecurity
Sid Hussmann retweeted
Paper #7. This paper is on the reading list of our PhD qual. L4 is among the most successful microkernel family; it keeps pushing the microkernel idea forward, with seL4 being considered the first formally verified OS kernel. This paper, however, is not all about the construction of L4; it is a great example of how to do system performance analysis.
In preparation of CS 523, I'm reading through many classic OS papers (just did Dijkstra's "THE" OS paper and Hansen's Nucleus paper). I plan to post my notes, as a forcing function to read papers carefully before going to class (inspired by @qianl_cs and @petereliaskraft). Bear with me if my posts are too boring or pedantic.
Sid Hussmann retweeted
Many people have the misconception that security is about patching vulnerabilities. That's the bare minimum. Security should to be part of the design and implementation from the beginning. Linux kernel is an example where that wasn't the case at all and it's hard to provide it.
Linux kernel is a large monolithic kernel, meaning it has no internal isolation. All of the code including obscure drivers enabled in the build have access to everything it does. It's almost entirely written in C, a memory unsafe language where many tiny mistakes are code execution vulnerabilities.
Linux kernel is currently a major weak point for Android's security. It's the easiest way out of both the app sandbox and the more constrained sandboxes used for Android media processing and Chromium renderer processes. The Linux kernel is also a major physical and remote attack vector itself.
Android is moving towards writing new Linux kernel device drivers in Rust to prevent most of these vulnerabilities. We'll leave that to them and will be focusing on deploying better exploit protections and more heavily using hardware-based virtualization for better sandboxing than Linux can provide.
Sid Hussmann retweeted
The beauty of technical problems - both in youth and adulthood - is that while they feature trade-offs, they usually have reasonably satisfying solutions.
The annoying part of adulthood is that you encounter problems where the trade-offs imply there is no good solution.
Sid Hussmann retweeted
The WebAssembly devroom will be held at the second day of FOSDEM 2025!
I would be glad if this could contribute to the growth of the Wasm community.
FOSDEM 2025: fosdem.org/2025
WebAssembly devroom CFP: lists.fosdem.org/pipermail/f…
Sid Hussmann retweeted
You asked for it, so here it is.
Visualizing CPU cache speeds relative to RAM.
Cache optimization is important too!
Sid Hussmann retweeted
I'm trying to start a positive security trend of #ResearchRespect where we give shoutouts to researchers whose work we really respect and describe why. I'll go first, focusing on the terms "evil maid attack" and "ring -1/2/3" by Joanna Rutkowska @rootkovska
Sid Hussmann retweeted
if you want to try out my port of Clang to WebAssembly, here are instructions: discourse.llvm.org/t/rfc-bui…

































