Luke Elin (@LukeElin) header
Luke Elin (@LukeElin) profile picture

From Fat to Fit Nerd, exploring the boundaries of hacking and latent space. CEO, Cybersecurity Expert, AI maximalist . Living with energy, pushing limits.

Australia
Joined July 2014
Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)Photo by Luke Elin (@LukeElin)
Luke Elin retweeted
Phrack 73 just dropped. 23 papers. Among them: kernel exploit development in the LLM era, a virus inside an LLM, ML backdoors in the wild, and a paper titled "Security is failure: Every Patch Is a Guess" - published the same week ten infrastructure outages all traced back to patching. The oldest zine in hacking and the national incident report converged on the same thesis in the same seven days: the update is an act of faith on fragile bones. Tell your SOC you'll be busy. I told mine. phrack.org/issues/73/1
1
2
2
275
Luke Elin retweeted
Java4Shell dropped tonight. The crew shipped wire detection - Suricata, Sigma, ModSec. Nobody shipped memory. So we did: YARA rules for JVM heap dumps and hprof, validated against a real reproduction on a pre-fix JDK 21 about an hour ago. All three pattern rotations covered, false positives controlled, capture file included so defenders can test their own setups. Best we can do tonight. More will follow. github.com/darkfibr/yara-jav…
The PHP2Shell crew just dropped a Base64 decoder RCE in OpenJDK. Every JWT on earth parses Base64. The next patch wave just got scheduled. I said this morning the pipeline was loading again. It's loaded.
1
5
373
Luke Elin retweeted
The PHP2Shell crew just dropped a Base64 decoder RCE in OpenJDK. Every JWT on earth parses Base64. The next patch wave just got scheduled. I said this morning the pipeline was loading again. It's loaded.
Java4Shell: write once, RCE everywhere calif.io/java4shell
1
4
1
4
685
Luke Elin retweeted
We built Australia’s own AI model, and today it’s open to the world. Agentaus 1 (@AgentausTD): 220B parameters. 84.4% on MMLU-Pro. Ranked 43rd in the world. It runs on Australian servers. Nothing ever leaves Australia, and nothing you type is used to train AI. Delete a chat and it’s gone for good. Most of the world’s AI answers to somebody else. This one answers only to you. Free to start: agentaus.com.au
229
138
142
1,547
1,013,435
Luke Elin retweeted
They named the storm in real time: Black October. PHP2Shell put unauthenticated RCE into core PHP - seventy percent of the web - in the same month as Node.js criticals, poisoned npm and PyPI packages, the Pentagon breach, the Denmark register, the largest crypto heist in history. "The internet is going into LOCKDOWN," posted a million-follower account on October 22. The fear was already installed. Now count the lag. Enterprises patch on a pipeline: test, stage, schedule the window, deploy. Two to four weeks from scared to shipped. Everyone got scared the same week. So everyone shipped the same weeks. Ten maintenance outages in six days. That's not an attack and it's not a coincidence. It's Black October's invoice - panic to patch, in motion, all at once, on a stack old enough that touching it is surgery. Here's the hard part. The loop doesn't close - it loads. AI-speed exploit discovery guarantees more Black Octobers. Every Black October guarantees another panic-to-patch November. Every November touches older bones harder. The security industry's alarm system is now a load-bearing part of the infrastructure it protects - and the alarm is ringing again: the FBI's ARTEX FLASH dropped this morning. The next deployment wave is being scheduled right now. And the final receipt, the one I keep at arm's length: you may not need to hack a country to take it offline anymore. You need to scare it into patching itself. The warning is the weapon. The response is the outage. They named the storm. Nobody named the season. PANIC TO PATCH.
3
6
463
Luke Elin retweeted
You're citing the FBI FLASH. We pulled the ARTEX source ten days before the FBI published theirs - because it looked like a fun toy. The safety guard wasn't missing, it was deleted, and the code admits it in a parenthetical: guard/guard.go, the file that calls itself "the safety boundary layer," opens with "(The RoE authorization-scope mechanism was removed; a replacement may be added later.)" Our disassembly of the backdoored docker build and the YARA rule went up Nov 14. And you're right that it fills the gap - the hostile half of the month. But hold the halves separate, because the receipts demand it. ARTEX in October (ConnectWise, 2,250 M365 tenants) is attack-side. The nine outages this week are maintenance-side - every named cause, every one. Both true, same month, different mechanisms. That's why "nobody got hacked this week" and the FBI FLASH are both true at once. And the scary version isn't "hackers did it." It's that a superintelligence-powered attack framework with its guard deleted has been in the wild while the whole stack aged past 25 - and the week everyone finally patched, the week it cracked, is the same season. You don't have to pick a conspiracy. The truth is weirder: nobody is driving, and both pedals are metal.
Replying to @Darkfibr3
@FBIDallas says some of it was hacking. “State-aligned hackers” no less.
1
2
9
766
Luke Elin retweeted
🚨 When Fable 5.2 give this video to me, I was genuinely blown away. Everything you see, and even everything you hear, was created entirely with code. Now, let Fable 5.2 take you on a journey through the world as seen through the eyes of a little cat.
74
93
19
1,354
100,664
Luke Elin retweeted
I was working through this today with @Scott_Helme, and it’s MASSIVE! Most sites would have absolutely no idea their customers were being served malware. Where do you even begin with disclosure on that one?!
This is a big one. We found malicious JavaScript being served through Optimizely to some of the world's largest websites. 178 organisations affected, with everything from crypto theft and card skimming to iOS exploits. Zoom, BBC, CNN, HSBC, Salesforce, GitLab, KLM, IKEA, Equifax, Ledger. The list goes on... Full list in the post. blog.report-uri.com/worlds-l…
33
73
2
530
54,795
Luke Elin retweeted
Incidentally, we were both protected from the malware initially because @The_Pi_Hole was already blocking Optimizely, we literally had to allow it to investigate further. This is one of the key reasons why people choose to block stuff like this from the outset.
4
4
51
6,736
❗ ALERT ❗ Vulnerabilities are moving quickly. Organisations should prepare for more frequent security updates and shorter time to act. Patch systems, reduce internet exposure, strengthen access controls and monitor for suspicious activity. Learn more 👉 cyber.gov.au/alert/vulnerabi…
20
139
24
335
39,976
Luke Elin retweeted
Everybody spent this year terrified of AI taking over aviation and the thing that actually grounded the fleet is a 29-year-old system named Unimatic that was installed by a company that doesn't exist anymore. 😂 No ARTEX swarms- No Vicious APT groups hacking UA.. Nope. It's Unimatic (weight & balance).. It tracks scheduling, aircraft weight, and balance. It is 29 years old. And it shit the bed EXACTLY in the same manner on Aug 6, 2025 at 6pm ET. 1,000+ delays, dozens canceled. Today? Nov 16, 2026. 6:12pm ET. 418 delayed (13% of schedule), 21 canceled (Per FlightAware) The pattern matches Aug 2025 and Friday's router, and geriatric infra fails all by itself, gloriously, on schedule and with a pattern that says ill do it again a year from now too. That's not a hack Guys. That's elder care. The machine isn't gettting pwnd, it's dying of old age in public, and we're all standing around the bed going "no indication of foul play." Of course there's no foul play grandpa just fell down. Again. Same hip. Fifteen months apart. Porr fucking grandpa. We are changing diapers again for a month.                                   And my security-professional rage-love is the "no indication of a cyberattack" line two hours in.... Gang...lmao.. Everything is "no indication" two hours in. That's not a finding, that's the intern hasn't finished the checklist (CHECK THE ADJACENT CVE'S MARVIN) I say "no indication of cyber" about my own breakfast some mornings..which is a white monster usually. LMAO. US critical infra fucking sucks.
3
4
1
13
900
Luke Elin retweeted
Replying to @VoidStateKate
We have evolved to this… 👇🤣
1
1
2
95
Luke Elin retweeted
Kitesurf, our experimental browser engine for AI agents, is now open source. Inspect the code, run it locally, or deploy your own instance on Cloudflare Workers. cfl.re/4eHmAK6
76
98
32
921
64,594
Luke Elin retweeted
Four years ago, I watched a machine take my job. Best thing that ever happened to me. In 2021 I sold my writing agency (100 staff, 5,000 clients) because I could see AI coming for the work I'd built my life on. A year later, ChatGPT launched. So I burned my ships and went all in. I shut down every copywriting course. I turned down clients. I rebuilt it all with AI as the front and center of everything I did. Today, my AI clone has made 750+ videos I never filmed. It pulls 2M+ impressions a month. I work about 30 minutes a week. Here's where this goes next: People who hand their work to machines will get their lives back. More time with the people they love, and more time to build what they actually care about. People who wait will spend the next 5 years defending a job description. The revolution already started. Are you building with it or bracing against it?
22
7
3
155
9,608
Replying to @elonmusk
We know
2
8
153
🚨 Rumors I’ve been hearing. First OpenAI, then Anthropic and Google. OPENAI OpenAI is in a “safety lockdown” on its frontier work, tightening safety measures and containment protocols to avoid another Hugging Face incident. Training continues under much stricter controls. Most gains come from video games and extreme sandboxes where models must complete clear, difficult long-term goals. These force planning, experimentation, recovery from failure, and strategy discovery, driving intelligence and creativity. Public research roles match this: Synthetic RL uses self-play and simulators; agent teams focus on sustained execution and multi-model coordination; the Recursive Self-Improvement team builds systems that generate hypotheses, run experiments, and automate research. The internal model is still training but nearing completion. The jump is enormous. narrow superintelligence in specific areas and overall AGI-level capability. Its ability to independently handle difficult work sets it apart. ANTHROPIC Model 5 is training and is expected to solve extremely hard problems in mathematics, biology, and other sciences. Anthropic is pushing models toward original research contributions, complex experiments, and sustained reasoning. Fable 5.5 is Model 4. Fable 6 is Model 5; Opus 6 will be based on Model 5. Internal capabilities will feed the next generation of both families. Training environments are central. The Universes team creates realistic situations with ambiguity, interruptions, and delayed consequences. The model must track the objective, recognize failing approaches, and find alternatives. Public roles align: scaling RL across model size, compute, and task duration, plus environments for chip design, verification, and complex visual work. Multi-agent research explores how larger model teams can handle major codebases and difficult mathematics. Model 5 advances Claude toward owning entire research and engineering workflows. Its role in developing the next model grows more significant. Anthropic expects RSI March 2027. GOOGLE Google is pushing Gemini 5 for a major leap in reasoning, multimodal understanding, and independent long-horizon work. Potentially surpassing all labs by end of year. Carbon topping Artificial Analysis signaled the speed of their frontier progress. World models are a key advantage. Genie 4 generates interactive environments that evolve with agent actions, enabling vast practice in planning, testing approaches, and learning from outcomes. Public roles highlight persistent context and personalization: agents that retain useful information, adapt to users, and execute complex workflows without losing track. Gemini 5 delivers stronger intelligence across text, images, audio, and video. Combined with Google’s infrastructure and simulation work, the scale is enormous. All three labs are also advancing robotics and world models. intelligence that understands physical environments, predicts action consequences, and operates through real machines. Meanwhile, their models increasingly contribute to creating their successors. Each improvement accelerates the research behind the next. The next three months are going to be insane.
80
233
45
4,127
649,461
⚠️Severe Thunderstorm Warning - SEQ - Update VERY DANGEROUS THUNDERSTORMS in the Darling Downs may bring giant hail (>5cm) and destructive winds (>125km/h) to Toowoomba, Felton and Cambooya by 2:10pm and locations between Stanthorpe and Warwick by 2:40pm bom.gov.au/warning/severe-th…
1
4
2
26
5,695
Luke Elin retweeted
Meet The Nine Heads
13
32
7
328
13,788
Luke Elin retweeted
"I want to see men who never lifted anything explain labor to me" 🤣
5
1
1
69
4,740
Luke Elin retweeted
There is millions like you
2
6
1
13
493