From the article linked by Techdirt: “executives have sought to almost entirely subordinate clinical case managers’ judgment to the computer’s calculations”
Sounds like the issue is the executives. What does this have to do with “AI”? Also, the company that built this tech (naviHealth) was started in 2012. Their product existed long before large language models were created.
”“A little-known piece of trivia,” Altman announced. “This smell, after it rains for the first time. You know what’s that called?”
“Is this going to freak me out?” Williams said.
“Petrichor,” Altman said. “It’s my favorite smell. You only get to smell this once or twice a year, because it has to not rain for a while, and then rain. It’s the smell of summers in St. Louis.”
It's not that these aren't interesting subjects or potentially useful in adult life. They just shouldn't be the priority. Is it potentially useful in adult life to know how your car's IC engine works? Sure, but nowhere near as useful as knowing how to get a good rate on your car loan.
It’s just another example of the “participation culture” that modern parenting and social media have made commonplace. Who needs to do anything real when you can just upvote or retweet? You get the same sort of participation trophy that you’ve been taught to aim for since childhood.
* Line management: are you willing to be the man and promulgate the party line? You can be mostly apolitical in this role.
* Middle management: are you willing to be the pawn of a specific member or two of senior management and do their bidding (which probably isn’t fully aligned with the party line)?
* Senior management: do you know how to intelligently break the rules in order to stand out from the crowd of middle managers? This could be by developing a broad following within the lower levels of the company through self-promotion, through cultivating specific relationships with the CEO and/or board, or by (in rare cases) delivering on highly visible and truly remarkable results for the company.
With respect to point #2, the real reason why revamping any regulatory regime "top to bottom" is difficult is because so much organizational, legislative, and administrative cruft has built up over the years in the federal government that means no one individual has the authority to drive reform. See FDA's discussion paper from Jan 2017 on revamping the LDT regulatory regime (https://www.fda.gov/media/102367/download) for more on the inane complexity created by the federal bureaucracy: "For example, a test made by a conventional IVD manufacturer would be regulated by FDA initially. If a laboratory made a significant modification to that test, it would then be regulated by CMS. If the original manufacturer then made another significant modification, the modification would be regulated by FDA."
Also: "In 2015, FDA established an Interagency Task Force on LDT Quality Requirements with CMS, the Center for Disease Control and Prevention, and the National Institutes of Health"
Any manager worth his/her salt can tell you that a task force involving four different agencies is unlikely to ever result in any meaningful change. A cornerstone of effective management is to designate one person who is responsible for execution and hold them accountable. This is exactly how the executive branch, from the president on down, is supposed to function (with oversight from the judicial and legislative branches, of course), but few parts of it are like that after centuries of bolting on overlapping agencies and departments.
As someone who has used a variety of VTC products (Zoom, Webex, BlueJeans, Teams, Skype, etc.) for several years on a daily basis (lots of external VTCs with different companies who use different VTC systems), Zoom is by far the best. The audio and video quality is head and shoulders above the rest (both on PC and mobile) and the interface is dead simple for even the least tech-savvy users.
My company uses Zoom, and there have been many instances where, during a VTC call set up by someone at another company (that doesn’t use Zoom), we have switched mid-meeting to Zoom because there’s something wrong with the other VTC system (someone can’t join, can’t hear, can’t speak, can’t share their screen, etc.). And the other options haven’t gotten noticeably better over the years either.
The death curve is also deceptive - there were likely many patients who were recorded as dying of the flu, or of influenza-like illness, that actually died of COVID-19, due to a lack of testing. The bottom line is that it’s very difficult to make a conclusive statement based on the data collected prior to widespread testing.
As for that Medium article, it’s written by a tech exec with zero education or professional experience in anything related to medicine or the life sciences, so I would treat it with a heavy dose of skepticism. Many of the deluge of preprint articles from academic institutions written by epidemiologists or public health experts would likely be a far better source of info.
The most important factor for financial time series forecasting is undoubtedly access to clean data. This is what sets Renaissance apart. There’s no need for particularly sophisticated math - they’ve been doing it for 3 decades.
My phone's contacts list. Make friends with as many people at your current company as possible, and when they leave, make sure to keep in regular touch with them. I enforce self-discipline on this by putting regular touchpoints in my GTD app. I try to get coffee with or talk on the phone to former coworkers once every 4-6 months, depending on how well I know them.
Great—let me know when you've changed the zoning laws, fixed urban planning more generally, and how the whole society is organized around massive subsidies for driving and inefficient sprawl housing with long-term unsupportable infrastructure. In the meantime, let's provide affordable transit that's accessible regardless of which neighborhood you live in: high-frequency, reliable buses.
Wrong—the claim has nothing to do with what rich or poor people like to use.
The claim is that the only way to provide affordable public transit that's sustainably accessible to people who aren't rich (that is, even when gentrification occurs in the city core) is to provide surface transit (buses), which can serve a much larger area than light rail.
> Maybe if we had more human-scale, walkable, transit-accessible neighborhoods everywhere, there would be enough of them to meet the high demand and more such neighborhoods could support mixed-income residents.
Sure—let me know when you find the funding required to build all those "human-scale, walkable, transit-accessible neighborhoods everywhere." In the meantime, let's provide affordable transit that's accessible regardless of which neighborhood you live in: high-frequency, reliable buses.
> He claims only the rich use light rail. But what is it about a light rail that causes a poor person to look at it and walk away deciding it's not for them? It simply doesn't make sense. Build LRT to a poor area, people will use it.
Are you sure you read the article…?
> The area around the LRT lines definitely attract investment, but if you look at who actually uses the line a few years in, it’s mostly rich people. Why? Because they’re the only people who can afford to take it – not because the fares are too high, but because real estate in the immediate walking area around stations becomes too expensive…Unless you have bus routes or other last-mile ways of getting to the LRT, then it’s going to be a public transit option that’s only available to people who can afford to live nearby. And the nicer you make the line, the higher an income threshold that’s going to require. (Unless you do the hard work of actually integrating the LRT line via last-mile bus routes into all of the other neighbourhoods that aren’t gentrifying.) LRT investment on its own doesn’t expand public transit; it gentrifies it.
Sounds like the issue is the executives. What does this have to do with “AI”? Also, the company that built this tech (naviHealth) was started in 2012. Their product existed long before large language models were created.