"In the experiment, we vary the socio-economic status (SES) of the patient and whether a second opinion signal is sent. Furthermore, measures of market, practice and dentist characteristics are collected. We observe an overtreatment recommendation rate of 28 % and a striking heterogeneity in treatment recommendations. Furthermore, we find significantly less overtreatment recommendations for patients with higher SES compared to lower SES for standard visits, suggesting a complex role of patients' SES. Competition intensity, measured by dentist density, does not have a significant influence on overtreatment. Dentists with shorter waiting times are more likely to propose unnecessary treatment."

Oh I didn’t know that.

The claim “stronger than bzip2” is strange. What does it even mean?

Also, comparing parallel decompression benchmarks with bzip2 instead of pbzip2 seems unfair.

Not sure about "on a dime". AI companies are currently spending billions of dollars every year, with none making any profit.

Sure, AI will get better. I even say that at the end. But, the replacement scare is unfounded. My point is mostly, AI will make humans more human rather than more unemployed.

Comments URL: https://news.ycombinator.com/item?id=49277342

Points: 3

# Comments: 3

Comments URL: https://news.ycombinator.com/item?id=49216406

Points: 2

# Comments: 0

I wrote about the harms of elitist social models and labels like that before: https://ssg.dev/trolls-dont-exist-9d879b960fd1/

"Symbiote or Parasite?

The Popular Reaction to Computers

Most people would concede that computers are strange new entities that have a significant impact on human life. However, I would suspect that most of them would classify the species of computer as a harmful partner rather than as a symbiote. Why has popular reaction to computers been so negative?"

----

"Next there is the common human fear that massive computers will be "Big Brother watching us." Indeed, there are a number of deep legal and moral questions concerning invasion of privacy, but I will not attempt to discuss these in this book.1 It is important to remember, however, that computers do only what they are told to do. If governments and big businesses are allowed to invade our privacy, it is our own fault for allowing them"

----

"A great deal of work has gone on, under the general title of "artificial intelligence," to try to teach a higher form of intelligence to computers. The decade of the 1960s showed that this task is vastly more difficult than some of the proponents of artificial intelligence had guessed. Although there have been a number of notable successes, in many cases the human effort to teach the computer has turned out to be' horrendous, and the amount of labor required for a computer to simulate human intelligence is discouragingly great. Often it reaches a point at which the effort is simply not worthwhile."

----

"On the other hand, it makes no sense for human beings to spend several days teaching a computer a task that a human being could do himself in an hour. In the present state of the art most of the attempts to teach higher intelligence to computers come under this heading. However, this rule has a clear-cut exception: if there is a task that human beings have to carry out over and over again, it may well be worthwhile instead to spend several days teaching a computer how to do it. It is also worthwhile to invest considerable effort in developing "artificial intelligence" in computers in the hope of possible future payoffs."

[1] https://bsky.app/profile/did:plc:4acfam7ybm4rtqrjpuss5hyb/po...

why not Zigbee or Thread?