
dragosr
@dragosr
Followers
24K
Following
3K
Media
496
Statuses
16K
Stop, Think, Pwn! (see also @[email protected], https://t.co/BjclXYWQ9R for alternate)
Vortex of Cool
Joined December 2008
Someone needs to remind Canada Post SMS phishing scammers that mail services are on strike. "Canada Post: We tried to deliver your package but there seems to be an error..." No you didn't, you assholes, there's a strike... https://t.co/jVN7LkYu3F These scammers are so
1
0
8
What's happening in Romania. An analysis for the rest of the world: On 11/24/2024 we had the first round of the presidential election. In Romania, the voting system states that the top two contenders qualify for the second round, where whoever wins 50%+1 of the votes becomes
156
2K
7K
How to make sure your writing is irrelevant for future generations and any insights from it lost to our knowledge. I would probably stop reading after reading that attempt at a legal limit. Mostly because I don't see much difference between training my brain with it and
1
2
12
Having a conversation with a friend going through his own later-in-life diagnosis for ASD and ADHD, mostly after his kids were diagnosed to help deal with their challenges in school, I had a realization of why ND folks are generally more successful in getting useful results from
2
1
9
...and important details about how fast you can run those tools: https://t.co/SzMyAW7bSB
2
0
0
I'd been avoiding it because someone else had snagged dragosr and seems to have parked it but: https://t.co/YkX8UypGEZ
3
0
3
I'm delighted to share with you a #Rust project we've been working on for a while called #Hyperlight. The Hyperlight project enables devs to use KVM or Hyper-V as a library for running hypervisor isolated funcs with extremely low-latency cold starts.
opensource.microsoft.com
The Microsoft Azure Core Upstream team is excited to announce the Hyperlight project, an open-source Rust library you can use to create very small VMs for embedded functions. Learn more.
7
70
195
Your emotional support Canadian checking in on folks. Hope all'y'all are OK down there.
4
0
13
I heartily endorse the conclusions of this article. Very useful advice, that you should consider carefully.
https://t.co/H680f6K6c8 Hacking the Edges of Knowledge: LLMs, Vulnerabilities, and the Quest for Understanding
0
0
3
To all the folks having a cow about the Mac Mini design: What kind of a wierdo turns computers OFF anytime except to force a reboot ?
5
0
10
RIGGED ELECTIONS IN GEORGIA 🇬🇪 I've spent the last 48 hours proving fraud in the Georgian elections. Let me show you the evidence in 2 charts. Chart 1: Russian tail...
426
4K
18K
This is nifty. A nice solution to a Real world problem Dupont cable mess! https://t.co/9VFR3IyD41
nerd-corner.com
Universal Dupont connector for Dupont and Jumper cables. Very strong holding force and 3D printed at home!
2
2
19
It's morally wrong to be reinforcing learning patterns that DON'T use AI. Our educational systems need to be teaching students how to more effectively apply AI to _every_ problem and task to achieve good results - regardless of subject - not putting up barriers to their use.
It is morally wrong to use AI detectors when they produce false positives that smear students in ways that hurt them and where they can never prove their innocence. Do not use them. https://t.co/cSBZRHgif5
2
1
6
Here is the promised code and data of our binary analysis technique ReSym https://t.co/VsH4CCAgZK Honoured to received the #CCS2024 Distinguished Paper Award! @acm_ccs @cerias @PurdueCS #binaryanalysis #llm #llm4code
github.com
For our CCS24 paper 🏆 "ReSym: Harnessing LLMs to Recover Variable and Data Structure Symbols from Stripped Binaries" by Danning Xie, Zhuo Zhang, Nan Jiang, Xiangzhe Xu, Lin Tan, a...
Introducing ReSym (CCS'24): our binary analysis technique, an LLM+static analysis solution that recovers names, types, and layouts of variables and data structures from binaries https://t.co/7ln6lFfYnJ
@danning_x, @i2huer, @NanJiang719, @XiangzheX XiangyuZhang #LLM4code #CCS #LLM
3
23
77
Your stochastic parrot has a three dimensional world in it. https://t.co/wExbY6sEU1
lesswrong.com
Gurnee & Tegmark (2023) trained linear probes to take an LLM's internal activation on a landmark's name (e.g. "The London Eye"), and predict the land…
0
0
3