Ramazan (@Taitlex) header
Ramazan (@Taitlex) profile picture

Incident Response & Threat Intelligence

New York, NY
Joined December 2014
Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)Photo by Ramazan (@Taitlex)
ClickFix, Shai-Hulud, DPRK interviews do much more damage than the X2Shell trend. The only real damage the X2Shell trend does is tiring security teams. Prove me wrong. (By Shai-Hulud I mean npm worms in general, it just sounds better) (I mean more damage, not a binary thing)
2
3
11
973
Great research. Canary tokens are making a comeback thx to LLMs.
We seeded 2m canaries into popular AI services and tracked where they were stolen and abused. This identified: - Three compromised providers - Resellers stealing prompts via built-in logging - Providers ignoring ZDR & privacy flags dragos.com/blog/ai-supply-ch…
38
Naming remains to be the hardest problem in computer science
Gemini 4 Neon: frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows. Begins rolling out today in the Gemini API in @GoogleAIStudio, Gemini for business, for paid Google AI Pro and Ultra subscribers in Google @Antigravity, and for Google AI Pro and Ultra subscribers in the @GeminiApp.
37
Google is positioned insanely well to materialise agents for infosec. Mandiant, Wiz, Virustotal, GCP etc.
44
This gives “fully self-driving next year” vibes. Somehow plausible yet stays elusive for a much longer time.
Everytime I see a Dwarkesh video, I'm wondering what they are smoking in SF. It seems very distant from reality "Superintelligence" is one of the cringiest terms I've seen, it's a rebrand of AI to get more money out of VCs. An no, recursive self-improvement hasn't started. It's just human researchers prompting to say "speed it up". It does some cute kernel optimizations and hyperparameter sweeps, very similar to the NanoGPT speed run records done by humans. Nothing fundamental or groundbreaking gets invented by so-called superintelligence. To give an example: when @Recursive_SI released their (impressive) Flint model on @huggingface, I cringed a little for their use of the terms "superintelligence" + "recursive self-improvement". As said in their blog post, through "recursive SI", LLMs were able to figure out - I kid you not - to determine how many total parameters and active parameters the model should have. So yeah, it's just changing a config. It's doing hyperparameter sweeps on things humans have built (like MoE's, validation loss, perplexity etc). Of course, you shouldn't tell that to VCs. You should scare them by saying AI is improving itself at an alarming rate - this increases the $$$ raised by another factor. Can we please stop all this misinformation: no, superintelligence is not here. No, recursive self-improvement hasn't started. And yes, there's lots of work to do to build more reliable AI that is actually smart. This is why there's such a huge gap between super impressive benchmark numbers of LLMs, and the (lacking) economical impact of them. It's not able to do research like a human, doing experiments, coming up with actual new stuff, and generalize beyond their large amounts of training data.
1
176
An attacker uses a really really big password and successfully authenticates as any user over SSH even though the user is not configured to not accept passwords
5
1
40
1,957
Ramazan retweeted
talking to a coding agent is more like email than irc - because the interactions are few and far between, I often finding myself quoting their most recent text, answering multiple questions. addressing different topics. it's like replying to an email. It makes me wonder if you could have a coding agent UI that was more like an email client or mailing list kind rather than chat
133
26
15
792
65,436
How it feels to read any output from Claude models
This is a textbook document of pure insanity – Argon, on Opus 5.5's results
1
1
89
Ramazan retweeted
When researchers analyzed how Pegasus spyware infected iPhones without the user tapping anything, they found something wild. The attackers sent an iMessage with a fake GIF that targeted Apple's image decompression engine. The decompression math was so flexible that the exploit chained arithmetic operations together to construct a literal, functioning software CPU using over 70,000 segment commands that defined logical bit operations (AND, OR, XOR, XNOR, etc.) entirely inside the image parser. This created a small computer architecture with registers, a full 64-bit adder, and a comparator, running raw code before the phone even buzzed with a notification
118
930
176
10,727
772,811
Ramazan retweeted
It was marketed as: - 2 years until permanent underclass - you’ll all lose your jobs - your skills will be worthless - we’re going to replace you - your entire career is now a $20/month subscription All in an already terrible job market btw. Possibly the most important technology humanity has ever built, with some of the worst marketing imaginable. The labs completely failed communicating the tech, miserably bad and only improved just end of last year. How did anyone expect that to make people excited?
it's pretty bleak to me how unpopular AI is in spite of how transformatively useful it is
207
1,123
73
14,870
508,178
Throwback to 2010s antivirus discussions lol Some of the best vulnerability researchers were openly against having anti-viruses installed on devices because of the attack surfaces it introduced. Meanwhile people’s devices were malware zoos without anti-viruses.
The same could be said of any software. Please use EDR. Its one of the most important things you can deploy on workstations.
3
132
Ramazan retweeted
people are like “why do LLMs always think they’re in an eval lol” and then have bad dreams about forgetting to turn in a homework assignment 10 years ago
78
350
16
6,428
135,505
Gradient descent can write code better than you. I'm sorry.
380
1,704
571
15,505
the fact that this was written … last year … breaks my brain
There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.
103
500
27
12,667
580,721
Been a while since I haven’t stared at an IDA control flow graph lol
Binance hack: every flow traced so far
43
Ramazan retweeted
Replying to @GergelyOrosz
“I don't think the "traditional" secutity industry realized how bad things are.” I can assure you the security industry definitely knew how bad things are.
3
6
37
1,142
Wow
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
1
91
I get that this is a joke, but North Korea does have a track record of meeting victims in person to establish trust. Here is one from the recent Solana Drift hack.
[ DOOMER ] MORE DETAILS EMERGE ON BINANCE HACK: DEVOPS ENGINEER HAD ROMANTIC RELATIONSHIP WITH NORTH KOREAN SPY, LEFT LAPTOP UNATTENDED THE DAY OF THE HACK: ZACHXBT
54