recently!

sweet sweet november!

what i've been doing:

• barely surviving academic workload
• built an ai debate arena (vallina js/html and api calls to frontier models)
• following prev. update, changed my x acct username to @organoidboy.
• telora... more on that later.
• bringing some catharsis in my life by writing [untitled]. more on that later.
• maybe starting a business w/ my high school friends to make smart audio glasses. more on that later.
• building a local canvas UI interface w/ memory and search functionalities and prompt engineer qwen 2.5 14b instruct to run directly on my mac. it also needs to help analyze my deep research reports downloaded from gemini. this will be a long process of vibe coding, and i need to prob implement google's programmable search engine and web scraping abilities. would be a side project that i don't really care about for now.
• finally, getting one month free access to whoop 5.0 and peak tier subscription and comparing it to my current apple watch s10. more on that also later.

first thing first, what is going on with telora?

momentum has slowed down because of a lot of assessments, prelab work, and just hw in general. can't be everywhere all at once, and aarjav is very occupied with exams as well. our academic workload is similar, our interests in digital twins and ai x biotech is similar, and our desire to start a company is also very similar.

what's at stake is industry level outreach. and here is what's going on on a macro scale:

• team formation [pretty much done, with prof. sun and aarjav, he is good enough to be both an advisor and down-the-line builder]
• tech validation [original hypothesis validated, requires trajectory-level proof POST market validation]
• data partnerships [all informal. currently have salk, upenn, u colorado, and umass... most of them are all generating data except for salk. stanford might be the be-all-end-all if we successfully reach out to them and partnership deal is made. but before we do that... we need TRAJECTORY-LEVEL PROOF of MARKET VALIDATION (yes... again...)]
• ok here is the big deal: market validation: we have saturated market validation from academia. all stars are pointing toward industry. and it's time to find all the contacts we can find in mid-size and large-size BO manufacturing firms to gauge demand & needs, interests, understand their workflow, and understand if they are using cerebral or cortical. this industry outreach would determine whether we pivot or stay in the lane to gather more data, enough to build a sustainable prototype. we are delaying the build process as much as possible because we need conviction. and conviction comes from effective communication w/ potential industry level customers, as well as a few mid-sized academia labs ofc.
• what follows would be prototyping, g&a on parallel with incorporation and legal documents, and finally... raise.

i'm going to start learning mom's test today AND doing agentic searches (using gemini and chatgpt) to fill up second page of my potential customer/partner google sheet. meeting with aarjav tmr to discuss outreach plans and see if he did his hw.

i need to pick up my pace.

[untitled]

this novel/novella would be moving slowly. it follows victor gao trying to juggle his past experiences, social and academic pressure, and also his ambitious life journey.

Metis Glasses

Metis is the first wife of Zeus, mother of Athena, and was the embodiment of wisdom.

I tried the mijia glasses by xiaomi and it's actually useful. it made siri useful. and with potential software updates from apple next year in 2026, we can ride on this wave of technological growth, whether apple outsource their ai capabilities is another story.

I've been trying to set up so many automations on my phone to max out the capbility of the glasses. for example, a shortcut automation that whenever i put on the glasses, it connects to my phone and thru bluetooth it starts reading me my health data from the last 24 hours extracted from apple watch, as well as weather forecast, my notion tasks, my upcoming events on calendar, and finally extracting from an open sourced news API to read me what i care about each morning. the model that would be interpreting and reading all these info would probably be apple's PCC.

However, what i found out is that the rss is very tedious to set up, the weather is very accurate, the health data is a mixed bag (that's what i cared about the most though), and i haven't figured out how to connect notion yet and im skeptical about its capbilities since im not sure the way i operate notion is how most ppl do. and my potentially unique method of using notion might diminish what i can do to automate things. for example, notion AI is terrible at calling tasks that are due today or in the near future. poke can't properly read all of these information. and im not sure if apple's PCC model can do better.

anyways, if these can all be set up, then i think the glasses can be extremely powerful. it can already call your voice assistants by pressing on right stem and control volume/media-playback. i also wear it when it's dead because it still works with the prescription lenses. it's fantastic. and it's one of my favorite gadgets so far. and also it's really cheap to manufacture.

so my idea is that, if the lenses can be PMMA by default to workaround the FDA hurdle, plus the lenses can be detached and the frame can be hooked up by opticians later on, then it would provide so much life-time value. my dad knows manufacturers in china and this could easily be a good b2c model, if the distribution is done right. there is no technicality in this product. just pure demand and distribution. if the connection is solid between my dad and the manufacturers, william and i will jump straight into market validation and see if this can be a viable business. if so, Metis Glasses would be born. the greatest utility-level bluetooth wearable!

the design needs some work but it has great potential. i see potential in things i use myself. it's just a matter of how many people i can reach, and among the ppl i reach, how many are just like me who want the same thing in their lives.

apple watch s10 vs. whoop 5.0 (peak)

i got a 1-month free trial with whoop (peak tier) and i would try it out and see if it's better than apple watch. if yes, i will use that over my current watch.

there is nothing wrong with apple watch besides battery life being less than 7 days. i don't like to charge my watch constantly. i would like to have it on 24/7. but that's not possible with the current progress of incremental battery performance boost from apple's end. with lower power battery mode on, even the ultra 3 can only go up to 72 hours. which is only 3 days. its 18 hour battery life increase (which is less than 1 day if you think about it) is not that crazy.

the sunk cost of my current watch is 0 since it was as gift, not a purchase on my end. the battery capability after 10 months of use is excellent, sitting at 98%, and the storage is barely used since i mostly download music on my phone and not my watch.

the main problem is still battery life. i dont want to charge it every day or 1.5 days. whoop 5.0m, on the other hand, can garantee at least 7 days of constant metric tracking, and the official site claims it can go beyond 14 days. this is wildly interesting and it does take the edge of design superiority.

here is a rundown of pros and cons of each:

aw s10 pros:

• sunk cost = $0
• the core features i want are all here
• sleep apnea [so i dont die from sleep]
• sleep detection/scoring
• push emergency notifications about heart rate irregularities, crash and fall
• hrv detection
• ecg (minor)
• can track specific workouts i care about (running & hiking)
• can quickly check time, date, set alarms, take calls, and acts like a mini smartphone (minor)
• pretty sleek design overall

cons:

• battery life is <2 days
• some data are just not there on the watch. blood o2, hrv can only be viewed on the phone because of the lawsuit situation.
• the health app UX is a mess, unable to aggregate data to be summarized. the app UI is still ok. very simple.

whoop 5 pros:

• much longer battery life
• most of what apple watch offers but often with greater accuracy and data density. i am very curious about the HRV and sleep data uptake difference between apple watch and whoop.
• free hardware upgrade
• looks a tiny bit cooler and minimalistic (minor)
• biological age evaluation (cool but gimmicky?)

cons:

• no sleep apnea detection
• expensive subscription
• no ecg in peak tier
• all data can only be viewed on phone (not sure about the ui design), if something urgent happens it cannot take calls for me

as you can see, none of them are perfect, so i will see how this goes!

one month of hardware testing will give me the answer of whether to keep my s10 or switch to w5 :)

so that's about it. more updates coming.