Sebastian Buhai
banner
sbuhai.bsky.social
Sebastian Buhai
@sbuhai.bsky.social
Academic economist. Bon vivant. "Enfant terrible".

Website: https://sebastianbuhai.com

Twitter: https://twitter.com/sbuhai
Mastodon: https://mastodon.social/@sbuhai
Threads: https://www.threads.net/@sebastianbuhai
You can do it, Paul!! We’ve run this route many times: survival rate 100% (baby included; she even earned Frequent Baby Flyer status). P.S. As a pro tip, gift-wrap a few mini toys: boom, mystery-toy reveal goes live. 📦✈️🧸✨👼
November 6, 2025 at 1:06 PM
Policy teaser: allow short, posted, transparent quiet windows, and certify reversibility. Selection gets you clean thresholds; reversibility shrinks financing wedges.

If you care about AI product cadence, disclosure, or real options with financing, I’d love your feedback!

7/7
November 4, 2025 at 8:25 PM
Pressed for time? Prefer a considerably shorter version? Take a look at these recent presentation slides (≈30 min talk):
www.sebastianbuhai.c...

(And here's that "IO-in-Chile" workshop tweet, where I announced the world about my new research paper: x.com/sbuhai/status/...)

6/7
November 4, 2025 at 8:25 PM
Sparked by (typically wine-infused) inciting conversations with friends in AI labs / IT from all over the world, and very puzzling real-world release/ patch rhythms. The hereby model aims to rationalize that evidence and make it testable via new, sharp empirical predictions.

5/7
November 4, 2025 at 8:25 PM
Measurement blueprint: one could use firm-authored telemetry (blogs, release notes, advisories, OSS tags etc.). Model predictions:
• pre-release dip in cadence & dispersion (silence),
• post-reset patch cascades,
• debt-insensitive patch timing when reversibility is high.

4/7
November 4, 2025 at 8:25 PM
Math: boundary-value verification (value matching + smooth pasting at triggers; target optimality at jump points).
Adoption: simple belief cutoff.
Finance: leverage bites only via irreversibility—patches debt-neutral; distortion at pivots bounded by takeover switching costs.

3/7
November 4, 2025 at 8:25 PM
Setup: the firm controls two resets—patch (reversible) and pivot (costly)—plus a public release clock.
Key move: short posted “clock-off” windows (local silence). Set clock=0 → kills belief martingale ⇒ pure S–s thresholds: two-rung release ladder (patch/pivot), no mixing.

2/7
November 4, 2025 at 8:25 PM
Merci de interes! Probabil săptămâna viitoare cândva voi pune un prim draft online (adresând/ incluzând și unele comentarii pe care le-am avut la prezentarea lucrării în cadrul workshop-ului acesta).
October 31, 2025 at 11:44 AM