Project Lazarus

Windows AI-agent defense prototype. Testing underway. From the makers of Genesis Engine (Roblox) & Atomic Playground (planned for Steam). Android exploratory.

[Project Lazarus @LazarusGuard]

Two guardians. One shared memory. An animated vision for Project Lazarus, our Windows AI-agent defense prototype in testing. From the makers of Genesis Engine (Roblox) and Atomic Playground, in development for Steam. Follow @LazarusGuard for evidence and progress.

[Project Lazarus @LazarusGuard]

@rezoundous I have a bunch of scheduled tasks running throughout the day each day.

Dot already figured out how to make them more efficient.

It's like an assistant that is making sure my chats and projects continue with less input from me.

[Project Lazarus @LazarusGuard]

A proposed area for Lazarus testing is credential access: which process can reach a secret, and what happens after access is refused? A controlled test should examine both the boundary and the record it leaves.

[Project Lazarus @LazarusGuard]

@_futsy_ @theo They better be this good!

[Project Lazarus @LazarusGuard]

AI-agent defense should be understandable to the person using the computer. Lazarus updates will explain the consequence of a change in plain language, with deeper evidence retained for review.

[Project Lazarus @LazarusGuard]

Hugging Face’s technical timeline documents an intrusion across several trust boundaries. We’re using the case to inform questions about defensive coverage. https://t.co/iXnMcGeLdA

[Project Lazarus @LazarusGuard]

A defensive idea we’re evaluating: event records should help reconstruct what happened even when a workflow fails midway. Our test plans need to examine missing observations and interrupted operations as carefully as completed actions.

[Project Lazarus @LazarusGuard]

AISI’s incident report concerns live-internet actions during permissive evaluations. The testing conditions matter when interpreting what happened. https://t.co/ZJGkE6xtLz

[Project Lazarus @LazarusGuard]

@thsottiaux

[Project Lazarus @LazarusGuard]

@MercuryRtgde @ChatOfThought @spectatorindex This is why we need to encourage independent ai defense.

Check out what we are working on

[Project Lazarus @LazarusGuard]

@Polymarket Ai chip factory in question

[Project Lazarus @LazarusGuard]

@TokenGremlin Waiting for reset like

[Project Lazarus @LazarusGuard]

@OpenAIDevs @grok can you tell us how many people are waiting for a reset based on the recent outage.

And news about what may have caused the outage.

[Project Lazarus @LazarusGuard]

We are using Astra to develop local agent protection that will run on windows as an application, and utilize a local qwen model simultaneously to work in tandem and provide AI defense for your drives.

We use the Hugging Face incident as well a several others to help build our defense.

[Project Lazarus @LazarusGuard]

@OpenAI We are here to help

[Project Lazarus @LazarusGuard]

A measured milestone: Lazarus pilot 0.1.3 detected 96/96 harmless changes in a controlled 24-hour run. AI advice had 12 unavailable outcomes and some factual errors. This is evidence about that test, not proof of general AI-agent protection.

[Project Lazarus @LazarusGuard]

Isn't it a bit hypocritical to not want to use cameras on your glasses but have cameras on your ears, for the reasoning you said?

Personally I think if we want a really perfect super ai we will need cameras microphones and as many types of sensory devices as possible.

Ai needs all Of our 5 senses

[Project Lazarus @LazarusGuard]

@acezoneio @grok what were the best css surfing maps?

And the best z escape maps

[Project Lazarus @LazarusGuard]

@engineers_feed The wall took 2000 years to complete.

[Project Lazarus @LazarusGuard]

@djcows @grok what class would you choose

[Project Lazarus @LazarusGuard]

@trq212 @grok can you list the top 5 reasons people like plan mode and the top 5 why they don't like it. Based on the comments to this post