Joined March 2015
- Tweets 9,446
- Following 1,253
- Followers 159,938
- Likes 6,163
2,051 Photos and videos
Pinned Tweet
7 Jul 2021
Where to find me:
🔴 Hacking Videos: youtube.com/LiveOverflow
📜 Articles: liveoverflow.com
💻 Start Hacking? liveoverflow.com/start-hacki…
🥰 Support: liveoverflow.com/support
📹 2nd Channel: youtube.com/LiveUnderflow
🤳 Instagram: instagram.com/liveoverflow
14
101
635
Was wondering why Fable was suddenly disabled in Cursor.
If the EU doesn’t get their asses together we gonna have big problems in the future…
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
15
7
127
15,978
LiveOverflow 🔴 retweeted
Jun 12
I cancelled my $10/mo Calendly subscription and vibe coded my own with Fable for $12,000
466
612
19,256
963,694
Jun 11
Anybody into model interpretability research. Do we know if certain tokens get “abused” in the layers as a “summary” token for the context before it?
I don’t know how to describe it properly 😅 but when you have multiple rounds of sliding attention, summaries of meaning might be carried from the front to the end? But this kinda “muddies?” what the vector should mean?
So I was wondering if a special “snapshot” token could be used that is somehow trained that (sliding) attention uses it to “summarize” prompt before. Maybe if it is ignored during training for next token prediction, and neural network or attention can more “freely” use it.
Or am I basically describing the idea of linear attention?
2
7
4,095
Jun 11
Do we know the attention layers used by the large proprietary models?
It’s great that we have huge contexts. But if there is eg. sliding attention of a certain limit, it would be good to know. Then you can plan prompt layout or define tool output token limits around it.
6
3,489
Jun 3
Reminder 👇 we are giving some really fun hardware hacking trainings at BlackHat USA. Learn how fault injection (security) research works.
May 22
Last chance: Early-bird pricing for @LiveOverflow's and mine "Applied Fault Injection" training at @BlackHatEvents USA expires today
2
3
17
7,233
LiveOverflow 🔴 retweeted
Jun 2
this is insane, if you have a money-printing machine and can max out tokens, then sure, it is worth keeping an agent in the loop and feeding it every single file.
a better harness, cheaper models, and a strong group of security researchers can produce the same results, or even better ones, at a fraction of the cost.
Mythos at Palo Alto Networks "found more than two dozen critical vulnerabilities in around three weeks, roughly five times what the company would typically find using existing tools"
But the company "burned through more than $1 million worth of tokens using Mythos"
2
11
146
20,476
May 31
We could theoretically automate entire factories with robots. But it doesn’t always make sense economically, so we don’t do it.
The future of AI will look similar. And there will be a point where you cannot justify 10x cost for agent when cheaper harness exists.
7
4
99
8,671
LiveOverflow 🔴 retweeted
May 22
Ok, confession time: I use agentic coding *all the time* and *every day*. And have been doing so for many months.
I am *terrified* of skill deterioration on my side. I see the studies, I can feel it myself. The agents make me much more productive, but I feel I need to force...
109
61
1,450
175,104
LiveOverflow 🔴 retweeted
May 22
Last chance: Early-bird pricing for @LiveOverflow's and mine "Applied Fault Injection" training at @BlackHatEvents USA expires today
1
1
18
13,787
LiveOverflow 🔴 retweeted
May 20
When Your VPN Opens Your Private Network to the Public!
An auth bypass in Palo Alto PAN-OS CAS Auth (CVE-2026-0265) that lets an attacker connect to the company's GlobalProtect VPN.
Blog - hacktron.ai/blog/cve-2026-02…
4
75
260
118,247
May 19
Giving my first training at BlackHat USA!
Well, actually it's @ghidraninja's training, and I am just hist mascot and hand out the hardware :P
6
17
172
14,015
May 19
Early pricing ends Friday
August 1-2: blackhat.com/us-26/training/…
August 3-4: blackhat.com/us-26/training/…
2
2,569
May 19
this is fine 🥲
May 19
Codex (GPT-5.5) solved it with 0 guidance. Just gave it access to the Chrome MCP and let it run on background!
7
5
158
19,882
May 18
😢
I've been seeing posts all over about the state of CTFs post-LLM. I've seen many attempts to explain why this is just a new evolution of CTFs, but I fundamentally disagree. I believe the original spirit is gone and I've written why in my blog.
kabir.au/blog/the-ctf-scene-…
3
4
116
24,441
May 18
All I know is, we cannot stop the progress of technology and it needs to be embraced. I am hoping some creative people figure out new ways for CTFs that keep the spirit alive. But we should also accept that there is a reality where CTFs are truly dead.
3
45
7,697
May 18
So when I say "adapt", I basically just mean "acceptance". We can all see how AI is changing our field and the world, and nobody ever can bring back the good old days ever.
3
11
2,642
May 18
anybody else getting older and noticing that you increasingly talk to yourself????
48
10
194
15,153
LiveOverflow 🔴 retweeted
6 Firefox entries at pwn2own.
5 withdrawals due to our 150.0.3 security release.
1 failed attempt.
0 Exploits.
No incidents.
Time to party :)
4
21
283
167,193