|
|
Post by sellhighdrinklow on Jul 30, 2021 8:28:42 GMT -5
DXCM adding another $2 billion to their market cap this morning.
|
|
|
|
Post by cretin11 on Jul 30, 2021 9:37:53 GMT -5
DXCM adding another $2 billion to their market cap this morning. Indeed. That alone is more than enough to buy MNKD, but it’s probably not in the cards. |
|
|
|
Post by sellhighdrinklow on Jul 30, 2021 14:57:10 GMT -5
DXCM adding another $2 billion to their market cap this morning. Indeed. That alone is more than enough to buy MNKD, but it’s probably not in the cards. Update - DXCM now up $5.7 billion in market cap on the day. |
|
|
|
Post by joeypotsandpans on Jul 31, 2021 21:06:49 GMT -5
Indeed. That alone is more than enough to buy MNKD, but it’s probably not in the cards. Update - DXCM now up $5.7 billion in market cap on the day. DXCM market cap 49.85B SENS market cap 1.31B Could be the best paired trade of the century 😉 |
|
|
|
Post by agedhippie on Jul 31, 2021 22:21:53 GMT -5
Al might have seen Afrezza as a near cure for Type 2 yet here we are years later and Mannkind management has still not been done to prove that. Until there is large scale trial data to support that claim it just remains a theory and nobody will act on it (other than VDEX). You cannot have Afrezza as part of an AP because an AP is meant to be zero touch and that rules out Afrezza as you both have to manually inhale, and also enter the fact into the AP system. That's not an AP, that's not even closed loop. It's hybrid which is where we are today. I'm curious whether the state of the art in AP development includes use of machine learning in the patient device. Making a distinction here that they could use machine learning to develop the model for (e.g. "meal detection") but once in the patient device it would be a traditional algorithm rather than one that further adapts using patient specific machine learning. If the AP were to use machine learning to adapt to each patient, that could accommodate use of Afrezza without notifying the AP with better results if patient is somewhat consistent with its use. Though that would likely be true for a patient adaptive system that they be somewhat consistent in eating habits in general. I do see that patient specific machine learning likely sends off safety/regulatory flares, so perhaps that isn't yet being seriously considered. To the best of my knowledge nobody is using machine learning. In principle it's possible that the pump manufacturers could do it, but currently they simply model the carb absorption curve for the food, and the insulin activity curve for the insulin. Getting a good data set for training would be hard as you point out. |
|
|
|
Post by sayhey24 on Aug 1, 2021 8:12:32 GMT -5
Mannkind was founded in 1991. There has been a great deal of money spent over the last 30 years on hardware and algorithm development. Machine learning, AI, call it what you like. Near everyone doing AP research has gone down that road with the first being Al Mann. Back in his day Al could write one hell of an algorithm. The thing is they all learned what Al knew 30 years ago.
The problem from 30 years ago will be the same problem 300 years from now, subq insulin absorption. As Al said, its too damn slow. It also varies from time to time. The only way to solve this problem at meal time is to not rely on subq insulin administration.
DXCM had a great quarter. IMO, what better time to announce a deal with Mike as Labor Day approaches.
|
|
