akash kulgod
@akadogluk
Followers
456
Following
3K
Media
131
Statuses
1K
chief canine comrade @dognosis. letting dogs teach ai how to smell so we can detect disease at scale
Bangalore, India
Joined March 2020
walter freeman in how brains make up their minds puts it best - the mechanisms of intentionality first emerged in the olfactory system with other senses moving in and co-opting its OS. biology first started with smelling while ai development is leaving it till the end. it's the
Terence Tao says today's AIs pass the eye test -- but fail miserably on the smell test. They generate proofs that look flawless. But the mistakes are subtle, and strangely inhuman. “There's a metaphorical mathematical smell. it's not clear how to get AI to duplicate that.”
0
0
2
ik we have bigger problems (the air being literal poison in many of our cities) but an insanely high baseline of noise as the default is one of my biggest pet peeves of urban india. radical difference in mental peace when i head to the farm and the loudest sound is bird song.
My sleep scores during recent travel were in the 90s. Now back in SF I am consistently back down to 70s, 80s. I am increasingly convinced that this is due to traffic noise from a nearby road/intersection where I live - every ~10min, a car, truck, bus, or motorcycle with a very.
6
1
14
"Because pugs probably don't sit around thinking "wow, I wish I could breathe better" - they just live within whatever constraints their breeding created. And I wonder if there are whole dimensions of. I don't know, curiosity? Creativity? Ways of connecting ideas? That got
0
1
4
v excited to soon achieve the "only wears company merch" phase in my arc
3
0
9
found my read for the metro tomorrow.
I've finally published a long-gestating piece about why India has largely failed at research & development, and inventing novel products downstream of R&D, over the past few decades, as well as what we can do about it. It's called Overcoming India's Technological Cowardice (1/2).
0
0
5
New @dognosis canine BCI kit dropped - version prayer. Records <5kΩ EEG data from 8ch <1min setup time. Custom fit w MRI neuronavigated montage that stays conformable during movement. ORIC SleepyEEG by friends at @myupsidedownlab w 2000Hz sampling rate. Made in 🇮🇳 for the dogs
0
0
11
since cancer-sniffing dogs is having a lil moment, thought i should also share that there's even a guinness world record for the most medical conditions detected by dogs - 28 - held by the amazing @MedDetectDogs
0
0
1
at this point, i'm probably personally responsible for breaking this news to ~1000 people (face-to-face, hopefully more via writing!) and the three groups of people who tend to know this are - clinicians (~20%), working dog people (~90%) and people who watched that one netflix.
0
0
1
what i find crazy is how under-the-radar this knowledge still is. like it seems to have the ingredients to be massively viral + well-known but somehow only 2-5% of even v well read people know that dogs can detect multiple diseases despite 2 decades of research ? why.
1
0
13
seems like you guys should talk if you haven't already @sohamsankaran @shivassangwan 🇮🇳🚀.
Announcing the 1947 Rise Fund. The world’s first fund focused on backing reverse brain drain talent and highly ambitious Indian tech entrepreneurs. 1947 Rise Fund is backed by one of the world's best GPs, founders, and operators.
0
0
2
while at @dognosis we believe that dog noses powered by engineering can already solve for early detection, no machine noses needed, we also believe we have the keys to teach AI how to smell. But unleashing the dogs to pass on the olfactory baton is a trail is for another day.
0
0
3
and when your sensor is a friendly lab called Jessie who doesn't charge you an exorbitant salary, the unit economics looks very good indeed
1
0
4
olfactory tech is the one that excels at the trifecta of sensitivity, UI, and unit-economics. VOCs are the earliest signatures of a tumour's aberrant metabolism, can be captured from a few minutes of breathing, and can require just a few additional seconds of sensor time.
1
0
2
Why focus on the above 3? So you don't miss any cancers (esp early stage when it matters most), you get everyone on-board (no dealing with poop), and you can achieve population-scale without breaking a payor's pocket. And of everything we have at our disposal in the 21st century.
1
0
1
The best way to increase the priors of a cohort having cancer is to first test for cancer using an orthogonal mechanism that optimizes for sensitivity, user-interface, and unit-economics.
1
0
1
But you can still get breast cancer before you turn 40, in fact it's more common now than ever, and you can get lung cancer without ever having smoked before. Cancer may be correlated with all kinds of risk-factors but ultimately it's a random cascade of mutations gone awry.
1
0
1
The solution is to not test everyone. You want to push your prior of 1/100 to as low as possible, what the FDA calls "clinical enrichment". We already do this w current single organ screenings (much worse priors that multi-cancer) - age-thresholds, smoking history and so on.
1
0
1