letao, on the menu:
- less disclosure is always better (for the venture).
- this paper's main goal isn't to publish something groundbreaking. it is a very intricate, easy to replicate for any computationally technical competitor teams, and therefore serves as internal advantage. the goal for this for internal use (not licensing) is to establish the venture's maximum sole control over the software pipeline to win on the reproducibility and quality war against engram and the like.
- if we are going for going deep on biology, she can be primary author instead of me to writet he parts taht are computationally vague, but biologically specific. i can tell that this paper would be bad though. if we are set about doing disclosure of any level (vague or complete) but temporally delayed, cannot be too late (academically useless) and cannot be too early (commercially dangerous). but again, going back to the previous point, it makes the most sense to not do the paper at all, not because doing paper will bring any benefits. it won't. a good example can be the fact that uc berkeley doesn't do large scale qc, nor does upenn. no one in academia will logically try to cite/replicate/peer-review our software protocols. ONLY the commerical competitors will do. and that is because the entire QC thesis has no academic value, but all the commercialization value.
another risk about the paper is such: the competitors will always try to design around no matter how little I disclose in the patenting. so information gap is crucial to buy time, above all else...
letao's intuition on beatriz: she just wants a bonus paper for the sake of academia recognition.
the hardware would've already been fall under partial if not all control under umass, from sun because of employment. then this would be a introduction drawn toward the software rationale: if the software falls under umass control, even with exclusive licensing, the company will have no direct control over the technology that the people within the company invented. the only way around it is to evade the TTO, and eventually assign everything back to the control of all the core personnel of the company.
about patent rejection + litigation risks:
- [potential uspto rejection reason after submission] innovative but abstract ideas.
- [potential uspto rejection reason after submission] not inventive enough and references previous inventions.
- [competitors can file lawsuit on] you are not disclosing the best mode
1st and 3rd will be more challenging than the 2nd.
to counter 1 and 3, i can defend about the general purpose aim, by nature (a speciifc best biomarker itself is a flawed demand to ask about because there's no best biomarker for this general-purposed pipeline specific to brain organoids, you swap whatever biomarker, and this implementation is for all types of morphogenetic/patterned brain organoids anyway, and we are being specific enough about brain. try do this for non-brain organoids), on the biology end and go more in-depth about the "best mode" of the computational implementation, which is actually sitting at the core of everything. this has to involve careful crafting, definitely with help of IP attorney to make all patenting claims litigation proof.
additionally, litigation is inherently expensive (around millions). so competitors often just try to find better/worse loophole alternatives.
summary:
- paper is bad or net-zero on all fronts.
- the patent should avoid TTO (so under me, and later transition to the venture through formal contracts etc), should avoid USPTO rejection, and litigation-proof in the first place. this is of utmost importance, above all else.
- even with IP secured, it's not enough; therefore it requires additional layers of protections such as trade secrets, NDAs to avoid informational leakage, and the company's own operation speed (locking revenue, generating more data, better use of funded runway, etc.) to not let competitors catch up assuming they can crack the IP or engineer loopholes to match your speed.