@mruef profile picture
Marc Ruef š–¢„ @mruef

infosec researcher @scipag Ā· lead architect @vuldb Ā· lecturer @eth @fhhwz @hslu @ikf_lucerne Ā· keynote speaker Ā· book author šŸ˜ infosec.exchange/@mruef

Joined June 2009
  • Tweets 25,654
  • Following 273
  • Followers 5,205
1,875 Photos and videos
Pinned Tweet
Marc Ruef š–¢„@mruef
12 Apr 2018
Ɯber 20 Jahre lang habe ich mit computec.ch ein Community Projekt betreut. Die letzten Jahre konnte ich dafĆ¼r eher weniger Zeit aufwenden, weshalb ich nun zum Schluss gekommen bin, offiziell das Ende bekannt zu geben. Es war eine grossartige Zeit, vielen Dank dafĆ¼r!

computec.ch 1997-2018

Portal zum Thema Cybersecurity

computec.ch
5
19
AMD denies researcher a $10,000 bug bounty after fixing critical auto-updater vulnerability security flaw took 124 days to patch #vulnerability #bug #flaw #law tomshardware.com/tech-industā€¦

AMD denies researcher a $10,000 bug bounty after fixing critical auto-updater vulnerability ā€”...

Quis renovatores renovat ā€” who updates the updater?

tomshardware.com
129
Google sues alleged Chinese cybercrime operation that used AI to send scam texts | TechCrunch #google #ai techcrunch.com/2026/06/12/goā€¦

Google sues alleged Chinese cybercrime operation that used AI to send scam texts | TechCrunch

The tech giant said a group called "Outsider Enterprise" used AI to scam hundreds of thousands of victims, sending 2.5 million text messages over a span of two weeks.

techcrunch.com
71
A narrowing window to understand AI science.org/doi/10.1126/scieā€¦

A narrowing window to understand AI

As capabilities of artificial intelligence (AI) advance rapidly, human understanding of these systems is increasingly falling behind. Several trends are converging to make AI systems harder to...

science.org
47
Pokemon Go Scans Quietly Trained The Navigation Tech Now Headed Into Military Drones dronexl.co/2026/06/09/pokemoā€¦

PokƩmon Go Scans Quietly Trained The Navigation Tech Now Headed Into Military Drones

PokƩmon Go players' 30 billion scans trained navigation tech now bound for military drones, via Niantic Spatial's defense partnership with Vantor.

dronexl.co
70
What you need to know about CISA BOD 26-04 tenable.com/blog/cisa-bod-26ā€¦

What you need to know about CISA BOD 26-04

CISAā€™s BOD 26-04 now requires federal agencies to patch the most dangerous vulnerabilities in as few as three days. Read our FAQ to understand the new four-variable prioritization model.

tenable.com
63
Obfuscation Pipeline - An Approach to automating Malware Obfuscation scip.ch/en/?labs.20260611
1
116
The U.S. Military Quietly Turned GPS Into a Global 'Numbers Station,' Evidence Suggests 404media.co/the-u-s-militaryā€¦

The U.S. Military Quietly Turned GPS Into a Global ā€˜Numbers Station,ā€™ Evidence Suggests

A random sequence in an innocuous GPS message field is likely encrypted traffic from the U.S. military's system for remotely updating cryptographic keys around the world.

404media.co
80
Mythos Grading Mythos: Got Patches Yet? | flyingpenguin flyingpenguin.com/mythos-graā€¦
69
Remote work -- not AI -- has sidelined recent college graduates, research finds #ai npr.org/2026/06/01/nx-s1-584ā€¦

Remote work -- not AI -- has sidelined recent college graduates, research finds

Research from the New York Fed finds that younger college graduates have been sidelined by remote work in recent years, as companies may be reluctant to hire those needing more training and mentoring.

npr.org
78
AI Can Now Design and Run Thousands of Experiments Without Human Hands. We Aren't Ready for the Risk to Biosecurity. singularityhub.com/2026/06/0ā€¦

AI Can Now Design and Run Thousands of Experiments Without Human Hands. We Arenā€™t Ready for the...

The gap between what AI can do in biology and what governance systems are prepared to handle is growing.

singularityhub.com
1
1
96
AI Agents Enable Adaptive Computer Worms #malware #worm arxiv.org/abs/2606.03811

AI Agents Enable Adaptive Computer Worms

A computer worm is malware that spreads on a network by replicating itself from one machine to another. Traditional worms, like WannaCry, exploited predetermined vulnerabilities, and their spread...

arxiv.org
1
118
Hackers Used Meta's AI Support Bot to Seize Instagram Accounts #hacker #breach #hack #instagram krebsonsecurity.com/2026/06/ā€¦

Hackers Used Metaā€™s AI Support Bot to Seize Instagram Accounts

The Instagram accounts for the Obama White House and the Chief Master Sergeant of the U.S. Space Force were briefly defaced with pro-Iranian images and messages over the weekend, after instructions...

krebsonsecurity.com
128
Day 2: Anatomy of an LLM Inference Request on DGX Spark blog.kubesimplify.com/day-2-ā€¦

Day 2: Anatomy of an LLM Inference Request on DGX Spark

A beginner-friendly walkthrough of tokenization, prefill, KV cache, decode, batching, TTFT, and why memory bandwidth shapes local LLM performance on NVIDIA DGX Spark.

blog.kubesimplify.com
1
1
105
LLM reasoning, automated: tokens drop 69.5% venturebeat.com/orchestratioā€¦

LLM reasoning, automated: tokens drop 69.5%

Researchers from Meta and Google built AutoTTS to automatically discover optimal LLM reasoning strategies, cutting token usage by 69.5% without accuracy loss.

venturebeat.com
1
102
Why Tesla's AI trainers don't trust its self-driving tech - or its safety stats #ai reuters.com/investigations/wā€¦

Why Teslaā€™s AI trainers donā€™t trust its self-driving tech ā€“ or its safety stats

A Reuters examination found the automaker isnā€™t close to safely delivering self-driving vehicles at scale ā€“ a central promise underpinning Tesla's $1.6 trillion stock-market value.

reuters.com
49
Websites have a new way to spy on visitors: analyzing their SSD activity - Ars Technica arstechnica.com/security/202ā€¦

Websites have a new way to spy on visitors: Analyzing their SSD activity

Telltale SSD activity can be measured in the browser using simple JavaScript.

arstechnica.com
72
LLM agent memory at 0.12% of model parameters venturebeat.com/orchestratioā€¦

LLM agent memory at 0.12% of model parameters

Researchers built delta-mem to give AI agents working memory at 0.12% parameter overhead, outperforming RAG and context expansion on memory-heavy agent tasks.

venturebeat.com
1
91
We Stopped Choosing LLMs by Vibe. Here's the Eval Harness We Built Instead magicsword.io/blog/llm-eval-ā€¦

We Stopped Choosing LLMs by Vibe. Here's the Eval Harness We Built Instead

Picking an AI model by feel isn't a strategy, it's a gut check. MagicSword built an open-source evaluation harness to compare models on what actually matters: accuracy, cost, and speed against real...

magicsword.io
55
Best Small Language Models on Hugging Face Right Now! kdnuggets.com/best-small-lanā€¦

Best Small Language Models on Hugging Face Right Now!

Take a curated look at the best small language models currently available on Hugging Face, what each one is actually good at, the benchmark numbers that back those claims up, and the code to get...

kdnuggets.com
1
59
Mythos finds a curl vulnerability #vulnerability daniel.haxx.se/blog/2026/05/ā€¦

Mythos finds a curl vulnerability

yes, as in singular one. Back in April 2026 Anthropic caused a lot of media noise when they concluded that their new AI model Mythos is dangerously good at finding security flaws in source code....

daniel.haxx.se
101
Load more