i think what some predictors have wrong about generative AI is the expectation that it will create content in its current one-to-many form - one movie, one piece of music, a book, consumed by many... it seems more likely that AI won't replace the creator in the traditional model but rather will become the personalized content creator for each individual...

what defines a "journalist" today? does a person repeating what someone else said, without taking steps to verify their statement, qualify? ... the reporting of "fired twitter employees" (which turned out to be pranksters) was neither verified nor supported by other investigated facts. "journalists" simply repeated untruths across live media. why? because it was sensational...

the greatest source of value & wealth creation in the 22nd century could be driven by terrestrial nucleosynthesis... demonstration plasma fusion systems are coming online over the next few years. early design tests of these projects are showing remarkable progress.

north america is in the midst of a worsening mega drought, increasing the probability that any of a number of events could lead to a black swan-like economic/social crisis... (1/n) first, h/t to @pkedrosky for being such a consummate source of data on this topic. in april 2020, scientists alerted us to an emerging "North American megadrought" detailing 2000-2018 as the driest period in 500+ years science.sciencemag.org/content/368/64… (2/n)

if enough people believe in something, it will become manifest.

regardless of institutional forces against that belief. our ability to communicate and convince others to believe in something is what makes us human, and has allowed us to defy "natural forces" for ~100k years.

there isn't a lot of transparency or accountability in the vax rollout at federal/state/local level, but the US does have enough vax doses to effectively end the pandemic in 45-60 days. here's why, key metrics to observe, and a reminder on how we can do it... today, pfizer CEO reported producing 70M vax doses as of 12/31 & another 33M doses since start of the year = 103M doses in total! they're making ~100M per month & have reportedly already handed over ~70M doses to Fed govt. finance.yahoo.com/news/pfizer-pr…

The Pandemic War is an actual war we must fight and win. the virus is surging, and evolving. we must surge back, and evolve faster than the virus. here's my proposed plan and analysis... first, in general, perfect has been the enemy of good, that's why we're losing this War. perfect test results. perfect vaccine. perfect treatment (no one can die; follow regular standards of care). we must end perfect and get good enough. it's the only way to win.

BTC has likely created more millionaires (10,000+?) and deca millionaires (1,000+?) than any company or startup in history. these are individual “retail” investors realizing these returns, not institutions, which may drive the Rise of Retail... 1) retail interest in speculative (high-risk/high-reward) investments will continue to sky rocket. BTC’s meteoric USD-denominated value climb sets a new benchmark. everyone is already chasing the next dragon.

the US is administering ~250k covid vaccine doses per day. we prob need to get ~100-150M ppl vaccinated = 200-300M doses = 800-1200 days at current rate. last mile clearly a problem as many doses sitting with states, who have discretion over admin process. states + hospitals don’t know what to do. this likely becoming key issue in coming days. lots of finger pointing. problem may be paperwork burden, process and procedural overhead. or just complete lack of uniform planning. distribute and hope clearly not a great strategy

amazing paper. 40-60% of population that have NOT had SARS-CoV-2 already have activated T Cells to the virus! likely due to cross-reactivity w/ other "common cold coronaviruses". may explain large % having easy time clearing virus and/or mild/no symptoms. cell.com/action/showPdf… "... CD4+ T cell responses were detected in 40-60% of unexposed individuals. This may be reflective of some degree of crossreactive, preexisting immunity to SARSCoV-2 in some, but not all, individuals..."

"re-opening" is not a binary thing. People, Places, and Processes (PPP) define our current and future state. which people can go out, where they can go, what processes are in place when they do go out. we aren't in a binary lockdown today. in sf, for example, anyone can go to the grocery store or "exercise outside". there are some processes in place (i.e. must wear masks to enter the grocery store; but not req'd to do temp checking on entry). our "lockdown" is already nuanced.

love @UCSF and @UCSFMedicine grand rounds video,

here are my tl;dw notes if you don't have the time to watch: epidemiology:
- asian response to pandemic was informed by their experience with SARS. already common practice to wear masks in public, knew how to rapidly test & trace, isolation protocol planned & ready; thus, better curve flattening over US

antibody testing update:
- now seeing some asymptomatic infections in general pop
- don't have enough high quality day to conclude anything
- <60% of tests distributed are reporting
- there is clear bias in the data
- there are sensitivity issues with test kits
(details follow) data:
- 437 reported results
- 11 positive, 401 neg, 24 inconclusive (user error)
- of positive:
-- 4 out of state (ppl shipped test to sick friends)
-- 3 known cases (broke protocol)
-- 2/392 in SF, 2/40 on peninsula
-- 100% of positives reported SOME symptoms vs 71% of negs

how do we test everyone? we need to test active infection (RNA) AND past infection (antibodies) cheap/quick to really trace/track. antibody tests now getting approved & coming to market (yay!).. these are cheap (<$10) easy-to-use (finger prick at-home) & fast (10mins)... best solution to test for active infection may NOT be PCR. PCR is best CURRENT solution - costs $6, but requires central lab, 6 step procedure, 45 minutes, and requires people, swabs, and reagents. can we get cheap/fast at-home active infection test? yes, we should, soon...

SF C19 antibody testing update: lots of mixed data due to non-conformity to testing parameters. improving checks on how we are getting tests out and results back. here is where data stands as of tonight: 338 tests, 7 positive, 20 inconclusive, 311 negative. HOWEVER, 3 of 7 positives were already PCR+ coronavirus patients (should've been excluded) and remaining 4 live outside SF. here are SF-only statistics:

IgM and IgG antibody tests are not a replacement for RT-PCR tests in diagnosing an active SARS-COV-2 infection. the IgG/IgM tests can be used to identify a past infection - useful for finding asymptomatic ppl; typically ~10% false negative and <1% false positive a big benefit of these tests is time and cost - they take 10 minutes and cost pennies to dollars. as such they can be deployed widely and used for figuring out the true infected population count and tracing/tracking infections over time and space

we should also be optimistic on treatment for covid-19. korean doctors created a protocol for treatment using antivirals and chloroquine, drugs which are widely available and affordable. elsevier.com/__data/assets/… as a result of widespread testing, containment, and the adoption of their treatment protocol, south korea is seeing the lowest fatality rate of any country with major outbreak worldometers.info/coronavirus/

another important data point no journalists seem to be covering - the data about hospital conditions in affected areas. we read a lot of anecdotes about despair and death. these are awful stories. but how affected are we and how are the models playing out? *with data* as of today, the wait times for most ERs in seattle is still under an hour. as of this tweet, UW at 40 minutes.. hospitalstats.org/hospital-ratin…

as of today, 5 NBA teams have had a coronavirus RT-PCR test done on their players and 7 players are positive (a 9.3% positive result) NBA players are the ONLY broad population sample being tested for the coronavirus, their positive rate is a strong data point for the general population