The longer argument
Why I can see this blind spot
The short version is on the front page. This is the record it rests on — every claim a link, because the point isn’t to be believed, it’s to be checked.
Biological AI safety has a structural blind spot: a vaccine and a weapon emit the same signal. Defend by blocking the signal and you block legitimate work while the real threat — the credentialed insider who understands the supply chain and the regulatory surface better than the monitor does — walks straight through.
The discriminator was never molecular. It is contextual and human, and it exists in every case. The work is teaching the system to read that layer — and I can see it because I have lived it from every side at once: I built the capability, I run the facility, and I am myself the case the broken filter rejected and the right one saw through.
I work the attack side, in public
In 2017 I gave the solo BioHacking Village talk on what bioweapons can now be made in home and community labs, and how regulation might catch them without killing legitimate work. In 2024 I returned to argue that the field misreads the threat: AI-plus-biotech is an old danger accelerated, not a new one. The vector is as old as the plague.
2017 · DIYBioweapons and Regulation 2024 · Past, Present & Future of Bioweapons
I built and certified the facility they said couldn’t be
I founded BioFoundry, Australia’s first open-access molecular biology lab — built from salvaged and self-made equipment, then taken through certification the system did not believe a community lab could pass. I deliberately run a containment level above what the law requires. The point of the attack-side work is defensive: you cannot design detection for a capability you have never built.
I am the blind spot’s own case study
Every institution that could have given me an IBC pathway read me the same way: high public profile, therefore liability for no gain — and locked me out. The regulator was the one node that saw what they missed: that the gain was me — that what you want is exactly the person who gives others a clear path to doing the right thing. So they invited me to speak at the regulatory conference with no IBC engagement at all, because my having no possible path to one was itself the problem worth airing. The institutions applied the broken filter. The regulator applied the correct one. I have spent a decade on that exact distinction — which is the distinction this role exists to build into a model.
I am about to run the whole toolchain in the open
My next project is a three-month, fully public, livestreamed antimicrobial-peptide design-and-validation run across the modern generative-biology stack — every sequence, protocol, and result released to the public domain. The molecule is the demo; the point is that the toolchain generalises. It is built, in part, as a live and recorded demonstration of exactly where the dual-use signal becomes indistinguishable — the precise thing I want to help you learn to read.
The profile that filters me out is the qualification
The cyborg-rights case, the name, the public history: a normal process reads these as risk and drops me before anyone looks. But the attacker’s mindset and the “credible” institutional position are mutually exclusive — anyone academic enough to sit comfortably inside the system cannot think like the threat. That is why the system has no answer here, and why this has to reach you sideways instead of through a form.
I am not making the case that I am hireable. I am making the case that the blind spot is real, that it is load-bearing for everything this role is meant to protect, and that the combination needed to correct it — attack-side fluency, a certified facility, and a decade of regulatory trust, held by one person — is close to unrepeatable.
Whether that person is hired matters less to me than whether the blind spot gets seen. That is why this is a page, not a cover letter.