Based on the slowing #codid19 death rate, The University of Washington just reduced their projection of US deaths by the end of the epidemic (Aug) to 82k covid19.healthdata.org
(last week's projection and this week's shown)
Good explanation from the UW team for why their projections are coming down. Includes both the effects of social distancing and the estimates that hospital capacity will not be exceeded in most places healthdata.org/covid/updates
The University of Washington just lowered their projection for US deaths by the end of this phase of the epidemic *again*, to 60k, based on latest data. This is now *half* of their original estimate
While AI is having a summer, robotics is stuck in winter. While the perception problems have been largely solved (thanks to AI), the other robotics problems -- understanding, planning, grasping, energy, flexibility, judgement, etc -- need breakthroughs.
Maybe it's because "robotics" is not a thing. I'd argue it's a category error, too broad to be useful (like "electricity").
Cars are not drones are not industrial arms are not vacuum cleaners.
Efforts to create universal robotics solutions, like ROS, struggle w/complexity, serving too many masters & rarely being the best solution for any one specific real-world prob. Even firms that start w/ROS usually replace it w/something bespoke once they achieve product-market fit
It's now been almost a year since the New York Times put together this technically impressive yet totally wrong interactive article about how long it would take to get a Covid vaccine.
To their credit, the NYT allowed you to try, by clicking, all sorts of (implausible, they suggested) accelerants like fastrack regulatory approval and building factories ahead of approval. But no combination of options got you to how fast we actually got the vaccine.
I'm trying to understand why so many people underestimated our ability to get a vaccine (to say nothing of such an effective one) in less than 12 months. Part of it is that mRNA vaccines were still new; they didn't have confidence that they would work.
Was it always obvious which side of "pandemic wishful thinking" liberals & conservatives would be on (more pessimistic than the facts justified vs more optimistic)?
It's now clear that liberals tended to catastrophize & conservatives minimized, but why? Just a Trump-era thing?
Here's one good answer, involving feelings about "agency" and individualism
Here's another one from Nature, this one leaning on the theory that people were following political leaders, who during the Trump administration were tending to be critics or apologists depending on their party: nature.com/articles/s4156…
Still life of a simplified OpenBot mod thinking that 2-wheel drive makes a lot more sense than 4-wheel drive. More nimble, cheaper, easier. openbot.org
The phone does all the work -- computer vision, deep learning, motion control, etc
I'm incredibly impressed by the new Creality CR-6 3D printer. It's the first 3D printer I've had that *just works*. Silent (just fan noise), perfect quality, no tweaking or setup, just like a good paper printer
Game-changers turned out to be:
-Self-leveling w/tap-touch
-10 yrs of software improvement (Marlin)
-Good touch-screens
-Finally getting nozzles/filament drives right
-Dual z-steppers. The fewer belts the better
-Enclosures unnecessary (waste of space/money)
-Heated glass beds
Also under $400. Really can't see any reason to pay more these days