Incalmo

next generation of autonomous cybersecurity

[Omer Akgul @_oakgul]

When you think you're in clear because of the new cyber program adjustments:
Nope. Turns out in opus land computer viruses are bio.

[Pear VC @pearvc]

🚨 PearX W27 applications close on October 4th! Want to get your questions answered before the deadline?

Join us tomorrow for a virtual info session and AMA with @ajaykam and Brian Singer of @incalmo_ai.

👉 Register here: https://t.co/r9ljT18JLS

[Omer Akgul @_oakgul]

Heading to USENIX Security tonight. Would love to catch up with folks that are around.

Also, we're hiring @incalmo_ai, come talk to me!

[wambo. @wambosec]

personal update: i've joined @incalmo_ai to democratize the cyber space. we create hard evals and improve models' cyber capabilities.

if you're an exceptional security researcher or want to work on evaluations and post-training for cyber tasks, we're hiring.

[Incalmo @incalmo_ai]

Read the full post: https://t.co/eYCC5Xuast

Incalmo supports open-weight models. We publish this research to identify safety risks early.

[Incalmo @incalmo_ai]

Kimi K3 and GLM 5.2 can create undetectable malware for $2.

Today, we are presenting our work on PathoGen, a novel system that turns models into evasive yet capable malware generation factories, enabling alarming scale.

[Incalmo @incalmo_ai]

Capabilities have increased over generations, current open-weight models can create cheap, evasive, and advanced malware under the right architecture.

We see promise in the peril: this can be used for preemptive hardening of defensive systems against future attacks techniques.