About
I'm Aliasgar Khimani.
I research AI alignment, recursive self-improvement, and emergence: the stuff that happens when systems start building themselves.
Also interested in epistemics: how we know what's true when generation is free and verification still costs what it always did.
Started in offensive security (CTFs, malware dev, the chaos era), now trying to understand things before they get too hard to understand. Hardware tinkering on weekends, mostly because I'm tired of everything being software.
01Operating Principles
- meaningful design first
- less is more
- ship fast, iterate faster
- build for the problem you have
- YAGNI
02Building Now
all projects██████████
████████ ████████ ██████
ØCLOAK隠
Crowdsourced privacy infrastructure against mass surveillance: at-cost RF / WiFi-sensing hardware plus a P2P threat-intel network. Offense is shipping; defense is empty.
docket決
Decision tracking for complex agent reasoning. Keep settled choices, rejected paths, and open questions available to later steps.
stoat貂
Local QEMU VMs on Linux from a TUI and a CLI. No libvirt, no daemon, one binary.
03Open Source
Open to interesting problems - AI infrastructure, distributed systems, or anything that makes computers have opinions.
@