The Identity Thread by Entrust

Gharibian delivers several genuinely useful insights - qubit reduction from 1M to 10K, AI agents reverse-engineering unpublished quantum circuits, and a concrete 6-month AI capability leap on an in-house benchmark - but Kleiman's contributions are largely framework-level repetition and the host generates filler throughout.

“going from million qubits, requiring only 10,000 qubits, that is, has been quite a big deal in the media. But we don't have really quantum computer even with 10,000 qubits. But uh, but those are computers we can actually see how to build in the next few years”

“Group of researchers with uh, AI agents actually reverse engineered the circuit and then published it. And in fact they kept running the AI workflow and improved the circuit even more”

The AI-accelerating-quantum-algorithms angle and the specific anecdote of AI agents improving an unpublished Google circuit are genuinely fresh; however, the cybersecurity half of the episode recycles standard zero-trust framing and well-worn 'cryptographic agility' talking points without adding new angles.

“Group of researchers with uh, AI agents actually reverse engineered the circuit and then published it. And in fact they kept running the AI workflow and improved the circuit even more so like the depth of the circuit resource requirements, they were even further improved”

“AI workflow. I think one of the technical terms is auto research coined by Kurt Pathy”

Gharibian is a legitimate practitioner - Stanford quantum PhD, active Blue Cubit research published alongside Google and Quantinuum teams, working hands-on with the best available hardware - but Kleiman is an internal Entrust employee functioning as a vendor spokesperson rather than an independent practitioner, dragging the overall caliber down.

“a theoretical physicist by training who earned his PhD from Stanford where he specialized in quantum computing and black hole physics”

“we published an article last year along with teams from Google, Continuum and others of this like class of solutions called scientific quantum advantage claims”

Gharibian provides concrete numbers (1M vs 10K qubits, IBM's $10B over 5 years, NIST 2030 vs Google's 2029 deadline, named algorithms ML-DSA/ML-KEM/SLH-DSA) and a verifiable AI circuit-reversal anecdote; Kleiman's portions are less evidenced, relying on illustrative analogies rather than data.

“going from million qubits, requiring only 10,000 qubits”

“the NIST standardization of the first three post quantum algorithms, mldsa, ML, Chem and slhtsa, that happened um almost two Years ago”

The host asks broad, open-ended questions and adds little intellectual pressure; there is no meaningful pushback on any claim, no probing of timelines or uncertainty estimates, and the wrap-up is purely affirmatory - Kleiman as an internal guest makes the format closer to a branded explainer than an interview.

“Can you talk a little bit, just a little bit about what your company's doing”

“Well, yeah, it seems like, you know, maybe it sounds crass in this conversation, but it's a budget problem, right?”