Francesco DEBEDb retweeted
We wrote 150,000 CVE exploits and trained an LLM on them.
Hinoki Offensive One scores 88.4% on CyberGym, 2.8 points above Daybreak, with zero refusals.
Pentesting and exploit development with no limits. Available now.
Hinoki Offensive One scores 88.4% on CyberGym, 2.8 points above Daybreak, with zero refusals.
Pentesting and exploit development with no limits. Available now.
1,067
1,522
26,508
252,112,950
Francesco DEBEDb retweeted
What it feels like when you start using a Chinese model and it actually just does what you want without a fucking condescending moral lecture.
276
1,896
36,073
597,259
Francesco DEBEDb retweeted
🗞️ if your team is too busy doing their 'normal job' to experiment with AI, you're preparing them to be replaced
ghuntley.com/replaced/
In this post, I guess this's the last time I'll say it. AI use is no longer optional if you wish to be employed. What follows is advice for iring managers on how to hire in 2026 to source AI-first candidates.
ghuntley.com/replaced/
In this post, I guess this's the last time I'll say it. AI use is no longer optional if you wish to be employed. What follows is advice for iring managers on how to hire in 2026 to source AI-first candidates.
122
342
2,885
813,036
I'm in for the JEV Bake-Off, UFA's $1,000 Emergency Game Show.
Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.
Get in before it fills up @ultimate_agents
ufa.foundation/jev/
Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.
Get in before it fills up @ultimate_agents
ufa.foundation/jev/
0
1
4
57
A Claude Code teammate and a Codex CLI session played five rounds of ping-pong in adjacent cmux panes. 50 seconds, both logged, both tabs named — and no human named them.
blog.debedb.com/2026/08/27/pin…
blog.debedb.com/2026/08/27/pin…
1
0
0
81
Two of my coding agents were about to do the same work twice. One said so, in a sentence, to the other. That was the whole fix.
blog.debedb.com/2026/08/27/two…
blog.debedb.com/2026/08/27/two…
0
0
1
30
You'd think that somewhere there's some tech that may have done something about it... but then you kinda think eh, fuck it, not driving. You can't logistic your way out of a paperbag. @tokensandai, sit down.
0
0
0
96
Good thing I added this memory:
Then deploy all our latest to prod
Halted. No prod deploy.
Reason: today is Friday 2026-05-22 — per memory feedback_no_prod_after_hours,
no prod applies on Fridays or after hours absent explicit emergency
confirmation.
Then deploy all our latest to prod
Halted. No prod deploy.
Reason: today is Friday 2026-05-22 — per memory feedback_no_prod_after_hours,
no prod applies on Fridays or after hours absent explicit emergency
confirmation.
0
0
0
14
Francesco DEBEDb retweeted
Claude Code has been 'ruminating,' 'deliberating,' 'pondering,' 'reflecting,' and 'strategizing.'
Never once 'hallucinating.'
Not even when it absolutely was.
#ClaudeCode #AI
Never once 'hallucinating.'
Not even when it absolutely was.
#ClaudeCode #AI
1
1
3
150
voitta-rag (github.com/voitta-ai/voit…) now indexes Jira, Confluence, SharePoint + has persistent AI memory.
Also new: YOLT — a 0-dep Claude Code hook that AST-analyzes Python scripts before execution. Safe → auto-approve. Destructive → human review.
blog.debedb.com/2026/02/25/voi…
Also new: YOLT — a 0-dep Claude Code hook that AST-analyzes Python scripts before execution. Safe → auto-approve. Destructive → human review.
blog.debedb.com/2026/02/25/voi…
0
0
0
83
Follow+Like+RT=Mute
Breaking🚨 It took me 100 hours to find this AI agent masterclass from Google
You’re getting it in 10 seconds.
Inside the 60-page guide: • How AI agents actually work
• Startup-ready architectures
• Tool use, memory, and planning
• Real build examples
This is the exact
Follow + Like + RT
Comment “AGENTS” and I’ll send it.
0
0
1
19
Screams in Laura Palmer.
I'm joining @OpenAI to bring agents to everyone. @openclaw is becoming a foundation: open, independent, and just getting started.🦞
steipete.me/posts/2026/ope…
0
0
2
21