Curating the Internet: Science and technology digest for January 17, 2020
How to stimulate knowledge sharing in an organization; An article suggests that AI may be on the cusp of a new AI winter; Google to end Chrome support for tracking cookies by 2022; Questioning the relationship between blue light and sleep; and a Steem essay discussing the use of coal as an energy source
Straight from my RSS feed | Whatever gets my attention |
Links and micro-summaries from my 1000+ daily headlines. I filter them so you don't have to.
First posted on my Steem blog: SteemIt, SteemPeak*, StemGeeks.
- Motivate Your High Performers to Share Their Knowledge - In a recent working paper, Harvard's Christopher Stanton discusses two common barriers to information sharing in a company. Underperforming employees can hesitate to ask for help, and high performers can be reluctant to offer it. Stanton and colleagues tracked two interventions in a call center environment. In the first, employees were paired, filled out worksheets describing strengths and weaknesses, and then compared notes on how to improve. In the second, employees were paired and offered a combined bonus on the team performance. The first protocol aimed to encourage underperformers to seek help, the second coaxed top-performers into offering it. Both approaches improved results during the study period, but workers who followed the second approach reverted to baseline performance five months later, whereas people who followed the first approach sustained their performance. Stanton speculates that the first approach helped people formulate better methods of productivity, whereas the second approach just caused people to work harder during the study period. The researchers suggest that businesses can harness this knowledge by setting up curated meetings where people have a reason to talk to each other and honestly assess strengths and weaknesses.
- Researchers: Are we on the cusp of an ‘AI winter’? - In the decades since the term, artificial intelligence (AI) was coined, the hype surrounding it has gone through peaks and troughs. The peaks in hype are often termed as "AI summer, and the troughs have been known as "AI winter". According to this article, the hype peaked in the 2010s, and there are signs that another AI winter may be on the way. As an example, the article suggests that many 2010 era thinkers were anticipating a sort of artificial general intelligence (AGI) that could function outside of a narrow domain, but in 2019, less people were predicting this in the foreseeable future, and the AIs that exist are still tightly focused on specific problems. NYU's Gary Marcus, is quoted to say that there is wide and growing awareness that new techniques are needed, and Edinburgh's Verena Rieser says that there is a feeling that the field has reached a plateau. Even among researchers who see a slow-down, however, not all agree that it is pushing towards a full AI winter, and some, such as IBM's Deep Mind group still have an optimistic perspective, saying that we have only scratched the surface of what may be achievable. -h/t Communications of the ACM: Artificial Intelligence
- Google plans to drop Chrome support for tracking cookies by 2022 - After a previous policy was criticized as "privacy gaslighting" (see Curating the Internet: Science and technology micro-summaries for August 26, 2019), Google now says that they'll discontinue support for tracking cookies in the Chrome browser by 2022. In August, Google said they would be deploying a so-called privacy sandbox in the chrome browser that can infer user preferences privately through the use of machine learning. Google's previous stance was that third-party tracking cookies were preferable to the alternative, a mechanism known as browser fingerprinting. Now, however, the company says that the privacy sandbox will make third-party tracking cookies obsolete, and they will discontinue support by 2022. As-of yet, there has not been enough public progress on the privacy sandbox for security researchers to assess the functionality. Google says that feedback from industry insiders has been positive. To me, without knowing the detailed implementations, the privacy sandbox sounds superficially similar to the way that the Brave browser delivers ads while simultaneously protecting user privacy.
- Forget What You Think You Know About Blue Light and Sleep - There is a widespread belief that light near the blue range of the visible light spectrum can cause difficulty with sleep, but a December paper in Current Biology challenges the claim. The paper's authors exposed mice to warm-toned yellow light and cool-toned blue light, and they found that the yellow light disturbed the sleep more than the blue. To explain this, they hypothesize that warm-toned light resembles sunlight and stimulates wakefulness whereas cool-toned light resembles twilight.
- STEEM Energy - Coal - In this post, @scholaris tells us about the history and use of coal as an energy source, beginning as far back as Europe of 73,000 years ago and with wide-spread use as early as 3,500 years ago in China. The post notes that coal is easy to light and creates productive energy, light, and warmth when burned. The article goes on with illustrations and descriptions covering the process of burning coal to spin a steam-turbine which drives a generator and creates electricity. It then discusses the importance of safe disposal of the waste from a coal-powered generator, listing the cases of Kingston, TN and Eden, NC as examples of what happens when the disposal is not handled safely. In Kingston, the post says says that as many as 200 clean-up workers have been sickened or killed after the company forbade them to wear protective gear during a spill cleanup. In contrast, Eden also experienced a spill, but the cleanup was so quick and safe that the response received praise from environmentalists. Surprisingly, the article quotes Scientific American in saying that, "Coal ash is more radioactive than nuclear waste." This is because trace elements of radioactive materials are present in coal, but they become more concentrated as the coal, itself, burns away. The article also asserts that clean coal technology has not yet been deployed in the US, but has in other places like Korea and China. Finally, it notes that coal is abundant, cheap, and after centuries of use, there is still enough supply to last for centuries more. Continued use is nearly certain, but clean technologies can make coal safer for people and the environment. (A beneficiary setting of 10% has been applied to this post for @scholaris.)
In order to help bring Steem's content to a new audience, if you think this post was informative, please consider sharing it through your other social media accounts.
This post will also appear on my pre-Steem blog, chescosteve.blogspot.com.
And to help make Steem the go-to place for timely information on diverse topics, I invite you to discuss any of these links in the comments and/or your own response post.
Beneficiaries
- Burn STEEM - null - 10%
- Cited author(s) - scholaris - 10%
- Posting and/or scheduling service (steempeak.com) - steempeak - 5%
- Steem/API services (anyx.io) - anyx - 5%
- Steem/RSS services (steemrss.com) - torrey.blog - 5%
- SteemWorld (steemworld.org) support - steemchiller - 5%
- STEM curation - steemstem - 5%
About this series
Sharing a link does not imply endorsement or agreement, and I receive no incentives for sharing from any of the content creators.
Follow on steem: remlaps-lite, remlaps
If you are not on Steem yet, you can follow through RSS: remlaps-lite, remlaps.
Thanks to SteemRSS from philipkoon, doriitamar, and torrey.blog for the Steem RSS feeds!
0
0
0.000
I expect AI 'summers' are almost completely the result of irrational exuberance. I point out the video you posted of Pigeonbot, which used actual feathers. Viewing the way in which the bot flies is comparable to a graph of the competence of general AI, as the manner of flight is directly dependent on the quality of the AI controlling the bot. Notably, the bot wasn't even autonomous at all, but directed by a human via a controller.
I believe general AI is at least decades away, and despite hubris regarding the state of tech today, comparing biological systems to information tech is a good way to judge the competence of our AI devices. We're orders of magnitude short of relevance. The vast majority of biological complexity remains utterly beyond our comprehension today, despite our immense progress. We have only just grasped that epigenetics and expression exist, and have a lot to learn before we can successfully build mechanisms remotely competent to improve on our general ability to understand and control things. Literally orders of magnitude more to learn than we have already.
Thanks!
Thanks for the reply!
I agree with your points. Additionally, I think a big part of the recent AI hype has had to do with the asymmetric access to free data. Data seemed worthless to the individual, but immensely valuable to the aggregator, so the aggregators have had free reign to harvest it for decades. Now that the value of aggregated data has been recognized, I think there will be ever-increasing friction as the original data owners seek to reclaim their share of that value.
In the long run, I think that AI development has to slow down dramatically when the source data stops being free to the aggregators.
You make a good point about the pigeon-bot. So far, human engineers haven't been able to fully replace something as simple as feather, so how far must we be from replacing a complex structure like the brain?
There's a story in tomorrow's scheduled post that will offer additional support to your idea that we are still very far from a complete understanding of biological intelligence.
My thoughts are piqued by your phrasing. Aggregators of data are also aggregators of financial assets, and this link you well reveal is not only a one way provisioning, but detracts from the enjoyment of their lives the sources of that data experience. I reckon this is also effected financially.
Looking at it this way is subtly different, and perhaps potential of better consideration. It seems there should be a means of improving the provisioners of data and financial assets lots. Imma think on it.
Thanks again!
Thank you, @remlaps-lite and @remlaps, for considering my writing as part of your blog. I am honored at the selection. There are many articles I have remaining on the Energy series before blogging about other critical topics of our society.
I look forward to reading your work in the future.