Sabitlenmiş Tweet
0xSha
935 posts
0xSha
@0xsha
~2 decades into tech. Security researcher, AI / blockchain engineer, occasional startup CTO, and CS PhD candidate. I build and break stuff, sometimes in public.
Katılım Ekim 2009
742 Takip Edilen10.7K Takipçiler
Chinese labs, no top-tier chips: full frontier weights, take them.
Anthropic/Fable: I can't help with cyber, that could be misused.
OpenAI: hasn't opened anything since GPT-2/OSS.
Thank god for Qwen, GLM, Kimi, and DeepSeek. Can't even imagine the gatekeeping level without them.
English
English
Uhhh I don’t think this is the final step my agent found for RCE… lollll
0xSha@0xsha
Here's wp2shell 🐚 single-file, zero-dependency PoC for the WordPress core pre-auth RCE (CVE-2026-63030 + CVE-2026-60137) one anonymous HTTP request: REST batch route confusion → unauth SQLi → admin hash → shell default install. no plugins. no login. the best of other public PoCs, unified into one file. scanner + full-chain code + deep dive + lab to play 👇
English
@0xsha this requires an admin password hash crack. it's possible without
English
Here's wp2shell 🐚 single-file, zero-dependency PoC for the WordPress core pre-auth RCE (CVE-2026-63030 + CVE-2026-60137)
one anonymous HTTP request:
REST batch route confusion → unauth SQLi → admin hash → shell
default install. no plugins. no login.
the best of other public PoCs, unified into one file.
scanner + full-chain code + deep dive + lab to play 👇
English
@0xsha For anybody watching along, I do hate that so many conversations today involve copying and pasting what top models are saying, but it also only makes sense here
English
@0xsha Oh no it’s the same. The exploit doesn’t steal or use the admin hash
English
The first 2.8T open model! Yes, we are pushing forward the frontier open intelligence!🫡
Kimi.ai@Kimi_Moonshot
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
English
@kimmonismus If the benchmarks hold up in real-world tasks, Kimi K3 becomes the #1 model.
When GPT-5.6 and Fable 5 are heavily restricted on many interesting tasks, raw benchmark performance stops telling the whole story.
A model is only as good as what it actually lets you do.
English
Thousands of engineering hours.
Thousands of alignment hours.
Countless hours building cyber guardrails and gatekeeping frontier capabilities.
Then frontier-class open weights dropped.
Now gatekeep and guardrail this.
Kimi.ai@Kimi_Moonshot
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
English
Grok 4.5 is surprisingly locked down.
Way more restricted than the usual Grok models but somehow, not a complete disaster like Fable.
Try it yourself in Council.
DarkGrove Labs@DarkGroveio
🌲 AI Council v1.2.0 — Day-0 support for Grok 4.5 xAI's brand-new Grok 4.5 takes the council's xAI seat the day it launched. council.darkgrove.io
English
0xSha retweetledi
stop opening 10 tabs to hunt for a job.
🌲 new free tool from the lab: jobaggr. we crawl 10+ niche boards for AI and crypto roles and aggregate them into one feed.
~1k jobs live now, fresh every few hours. free, no signup.
👉 jobs.darkgrove.io
English