by Lydia DePillis, CNN Business
There’s a pretty big lesson here in capex vs. opex (or a lesson in AWS margins [or a both]):
Still, [Craig Lowery, with Gartner] says, only about 10% of the federal government has transitioned its data to the cloud. That means a huge market opportunity remains.
The federal government currently spends about $5.8 billion on data centers annually, according to the research firm IDC. That’s not forecast to grow much. But cloud spending is at $8.9 billion, and growing rapidly.
The contrast is astounding. I get it - when you have on-prem IT you’re not re-buying the same data-center over and over again, and it’s unlikely the “data center” cost listed above is an actual all-in cost of their IT (whereas cloud-spend puts more cost-sources into a single invoice and is easier to fully account for), but the fact the numbers aren’t even on the same scale was shocking to me.
by Carly Cassella, ScienceAlert Found via HN
Original study: eLife 2019;8:e45183 DOI: 10.7554/eLife.45183
When a medical treatment, device or procedure is no better than previous or lesser alternatives, it’s deemed a ‘medical reversal’. These discredited practices are a major barrier to better and cheaper healthcare, but actually identifying them is surprisingly difficult and rarely done.
A recent study, designed to create a more comprehensive list, has unearthed nearly 400 established treatments, devices and procedures that fit this bill.
And further:
Like any other study on medical reversals, the most recent one comes with its limitations. It only looked at three journals, for instance, and the authors admit that other researchers may have categorised the results slightly differently.
I wonder what the authors could have found if they’d looked at more journals.
Guatemala’s coffee farmers are at the mercy of one of the world’s most volatile commodity markets. Over the past two years, the price has been pushed down by the increase in cheap, mechanized coffee production in Brazil — the Saudi Arabia of coffee — the strength of the U.S. dollar and increased production in Vietnam, Honduras and Colombia. It’s a perfect storm that has eaten away at the value of the beans even as the price of lattes and Americanos in U.S. shops has risen.
Immigration and migration problems are more complex than simple sound-bytes - and the collapse of coffee prices in Guatemala make a difference (although the collapse of a lot more than coffee in Venezuela are making an even bigger impact on the region).
by John Lynn, Healthcare IT Today
A US House appropriations bill will include language removing the ban on spending federal dollars to implement a national patient identifier.
This has led to folks speculating on if the time will finally come to develop a coordinated approach to patient-identity (a key barrier to medical-record interoperability and data-sharing [or, a key-barrier to reaping the benefits of medical-record interoperability and data-sharing]):
While this is good news for healthcare, I think we should temper our expectations a bit. Just like we shouldn’t assume that a national patient identifier will be a problem for privacy, we also shouldn’t assume that having a national patient identifier will solve all of our patient identification and patient matching problems. That’s 100% not the case. It will help, but it won’t solve a lot of the patient matching problems.
I’m not super-familiar with the history here (other than being familiar with the cultural disdain for federal identifiers in the US), but I liked this article from a couple of years ago on the topic
by Peter Waldman, Bloomberg Businessweek
[Xifeng Wu] hasn’t been charged with stealing anyone’s ideas, but in effect she stood accused of secretly aiding and abetting cancer research in China, an un-American activity in today’s political climate. She’d spent 27 of her 56 years at MD Anderson. A month after resigning, she left her husband and two kids in the U.S. and took a job as dean of a school of public health in Shanghai.
The article makes it sounds like the Center encouraged international collaboration (especially in areas that Wu specialized in - prevention and epidemiology). But over time the organizational culture shifted, and the focus of the Center’s research moved towards commercialization of treatment:
A few months later, the center gutted its international research program and put what was left of its collaborative-project arm under a business department. Bogler and former colleagues within the center say the focus then shifted away from research collaborations and toward business opportunities. MD Anderson spokeswoman Brette Peyton said in an email that the center’s global programs haven’t changed.
And this focus then collided with the FBI’s new justification for racial profiling:
In [FBI Director Christopher Wray]’s telling, China’s challenge to the U.S. today is unlike any this nation has faced. Whereas the Cold War was fought by armies and governments, the contest is being waged, on China’s side, by the “whole of society,” the FBI director said, and the U.S. needs its own whole-of-society response.
The article is pretty good and has a lot more details than anything I can quickly summarize.
by Steve Burke, GamersNexus
A range of 11W-15W, or at least 11-14W, explains the big heatsinks and fans on most of the motherboards, but we still need to look at specs. At present, the chipset doesn’t downclock during low load – it’ll burn at 11W nearly constantly, but our understanding is that AMD is working to fix that. X570 has been AMD’s biggest challenge with launching Ryzen 3000, resulting in at least one delayed release already. The difficulties stem from the added complexity and power requirements of PCIe Gen4. Comparatively, X370 and X470 were about 5.8W under load.
Spicy! I’d be interested in reading more about AMD’s struggles to get X570 out. I’m also curious if the fix to allow downclocking the X570 while not under load would be delivered as a BIOS updated to chips already delivered, or if it would only be in new product (assuming the fix doesn’t make it to launch? I think public availability is a few weeks out, now).
by Susannah Luthi, Modern Healthcare
The way billing and finance works for healthcare delivered in hospitals is really weird. The surgeons, anesthesiologists, and other specialists taking care of a patient are frequently operating as independent financial entities, who issue a bill to the patient (or the patient’s insurance) separate from the hospital.
This means that just because your insurance has an arrangement with the hospital, that doesn’t mean your insurance has a similar arrangement with all of the physicians you might see during your visit - and in some states the patient ends up stuck with the bill.
For some especially egregious examples the Vox article on Zuckerberg San Francisco General is good.
Additional reading: