Taking offense inappropriately in academia; Palmer Luckey on Apple Vision Pro; Allegory of Good Government; & Gurri, Zadeh, Yglesias, & Chen on "equity", charisma, "overcrowding", & upward mobility...
re: Noah Smith: "‘Misinformation is a job; correcting misinformation is a hobby…"
What about the paid "Trust and Safety" staff on social media? They can choke off disinformation reducing the need of the "hobbyist" to debunk mis/disinformation. While misinformation propagandists do have paid shills, there is an awful lot of unpaid misinformation by people with agendas, and just plain ignorance. [I accept that they may be parroting the output of paid misinformation providers.]
While AI/ML can never keep up with similarly AI produced misinformation, it certainly can reduce the cognitive load on the unpaid debunker. It would be really nice for a browser extension to flag misinformation whenever it is presented, providing a more truthful link with a text snippet to counter the flagged misinformation. It will require non-poisoned data, but that is the task of either governments or reliable corporations. While we can argue over the interpretation of facts, those facts should be unviolated and uncorrupted.
AT - I am guessing that a browser extension to flag misinformation would quickly be turned into the American version of "Social Credit Score" with the campaign against if quitely funded by Musk and his ilk who would all stand to lose if we reduced the ease with which misinformation can currently be passed around..
It will always be an arms race, with the defender trying to catch up. AI/ML just speeds things up. We let the genii out of the bottle with the section 230 that allows social media to be not liable for propaganda and disinformation. I don't think this was wrong, but the consequences are difficult to deal with. But is was ever thus, ever since the printed word could mass produce written thoughts and ideas. I really don't want to see authorities shutting down the modern electronic equivalent of pamphleteers, as we see in more authoritarian nations. What I would prefer is better means for individuals to cope with bad information. Personally, I believe we have to fight this individually, but with help from media. Just as I can avoid propaganda from Fox "News" and Prager University, and get less biased information from good newspapers, what I want is a way to use those selected media to help me as a guard from disinformation grabbing my attention. This would act as a filter as I browse the web. In essence, it would provide an extra layer of links and information that would highlight mis- and disinformation. Of course, which filter[s] I select will differ from those others select. That is inevitable. But again, this has always been the case, it is just that information now flows more like a from a hosepipe than a cup, vying for my eyeballs. Recall when Snopes was around to debunk lies? I see AI as an extremely fast and effective way to do the more sophisticated version of a Snopes debunking. Facts should be paramount, and then interpretations and opinions should be secondary. I doubt my interpretations and opinions are correct all of the time, but I do want to have them biased towards the "truth", and I particularly want to avoid being swayed by lies. If the public sphere could operate more like the scientific method, self-correcting, that would be ideal. Sadly that can only happen over the long term with experience, and even then that experience can be lost after a couple of generations.
AT - I agree, trying to shut it all down will never be the solution, it is the road to authoritarianism.
I think the answer an be found by looking within.
I agree sources like Prager U and Fox push misinformation for the right. What about starting with misinformation from the left? Our side is just as guilty as the right.
Great thinkers like Noam Chomsky exposed how the media on the left and right frequently give media viewers false choices. For example the media debated about whether or not the US should go to war with Iraq, when they should have been debating whether or not the justification, WMD, was real in the fist place. Or during the financial crises the media had media viewers debating whether or not to bail out banks during the financial crises. When they should have paid equal attention to whether or not to bail out home owners.
In both cases the media on the right and left manufactured a false debate that turned out to ignore other more relevant issues.
Now we are facing misinformation hyper weaponized by social media. Maybe the better choice is to start to self report Misinformation on our own side. Start holding media accountable for their own misinformation. What role did lliberal main stream media play in getting the US into Iraq or into bailing out banks and not home owners. What about Russia Gate, was that reporting in line with what ultimately came out in the Mueller report. I say we start holding our own side accountable.
We may never get the right to reflect on their own consumption of misinformation but we should be able to get our own side to re think what we are doing with misinformation.
"Maybe the better choice is to start to self report Misinformation on our own side. Start holding media accountable for their own misinformation."
Without some help, that is the problem highlighted by Noah Smith. Maybe AI helps, but what you have is the equivalent of moderating at scale - in this case reporting at scale. What does the reporting do - if anything? I still think it is better to have a tool that just flags suspect content, perhaps at the granularity of the sentence level. The data to do this is a curated set based on specific reliable content. Since so much content in books has been digitized, a curated set of books in addition to reliable online content would be the data for the AI to draw upon. [No "hallucinating" allowed]. STEM material is relatively easy to curate. It is the "squishier" subjects that are harder. Can an AI handle the opposing theories in economics or politics? Would accounts of historical facts be different between historical accounts and how to decide which is closer to the truth?
The left used to be kookier, but I think that has been eclipsed by the right. There is a difference between advocating for, say Marxist Socialism, and bad faith propaganda for neo-fascism. This is no longer a "both sides are equally bad" position. Pick areas of disagreement between left and right. Abortion, climate, how to manage the economy. Which side is promoting more outright lies? Yes, we should all be called out on misinformation, but I think the right wing is the greater purveyor of misinformation.
AT - Do you think that people on the right would opt in to real time fact checking? I don't think so. And if the fact checking forced people on the left to rethink old tropes that many of us see as Gospel ("winability" should determine who we support in a primary, we can't afford M4A, Trump was controlled by Putin during Russia Gate. etc...) then i think you would get back lash from the left.
Have you ever debated a real partisan. They don't care about TRUTH. They want to win or "own" who ever they are talking to... They don't want to find common ground or room for compromise.
With AI we can start to score media sites for both misinformation and for bias. For example Fox online and TV reports disproportionally on illegal immigrants or crime in communities of color relative to issues Americans care far more about like Wages and health care.
I think we can use AI to evaluate far more content and SCORE publishers for Misinformation and Headline Bias... And that would certainly influence what people on the left and right consume.
Nobody likes real time fact checking. Even ChatGPT gets huffy if you correct its BS. I most people who can apply logic use it to win an argument. Facts can be selected to do that, even if biased. However some issues are not determined by facts and logic, but emotion, for example politics. I have watched debates where the arguments used have not swayed the audience more than an iota, even when one side of the argument seems compelling.
I used to have a good friend who was very partisan, and he would simply deny he was partisan. Facts would just be denied with different facts. On some issues compromize could be gained, but very often not. The other problem was knowledge. It is hard to persuade some if they have little knowledge of the facts and can therefore ignore them.
I have a book for scientists to persuade non-scientists. It suggests not just using facts to appeal to the intellect, but emotion and other means to appeal to the heart and the "gut". People like me with a science background prefer factual arguments. Others find those arguments unappealing.
AI can be applied to check facts, but not bias. After all, we select information that appeals to our interests. Bias would be very obvious when rating academic subject magazines. Even news media will select different topics for articles and news.
I don't know what the answer is other than educating children, but as we have seen, that is subject to political pressure. Liberal and conservative views will always differ. Isn't that clear from history?
I agree with Gurri on identity. It is a a "perpetual conflict machine" Class, however is not a perpetual conflict machine. The media demonstrates that daily, and recently exemplified by the massive amount of attention paid to a few billionaires drowning on a submarine while daily not reporting on the deaths of despair witnessed as rural America or deaths of misery from our homeless population.
This media was able to unite Americans to fight wars in Vietnam and Iraq and fund a proxy war in Ukraine. But it can't seem to unite Americans over providing FREE health care to an aging and increasingly obese and unhealthy population.
I am reading Daron Acemoglu and Simon Johnsons book Power and Progress where they look at why societies care about the plight of the lower classes or sometimes simply do not. Unlike "identity" class wont go away. Century to century, nation to nation, economic system to economic system it wont go away. Identity will die because as Gurri points out, it is pointless. Class is not pointless.
Wong: Translation: some lenders may not get all their money back especially, if they do not let borrowers reduce rental rates of the property they have let for and especially if local land use regulation do not allow owners to shift the uses of the property that is rented. The latter is a policy issue. The former is not.
"It looks like something major changed with the appearance of Homo erectus, either on the supply side – improvements in food supply making brains more affordable – or on the demand side – a greater fitness payoff to a high energy brain – or both."
More likely it is the change in diet, and the use of fire to cook foods by <i>H. erectus</i> allowing for a change in jaw muscles and attachments, with energy from a smaller digestive gut being redirected to the brain.
This does not discount other factors, especially sex selection for "smarter" individuals. However, the larger brain size may also have been limited by the size of the birth canal through the pelvis. Childbirth was a major cause of women's mortality until the 20th century.
re: Noah Smith: "‘Misinformation is a job; correcting misinformation is a hobby…"
What about the paid "Trust and Safety" staff on social media? They can choke off disinformation reducing the need of the "hobbyist" to debunk mis/disinformation. While misinformation propagandists do have paid shills, there is an awful lot of unpaid misinformation by people with agendas, and just plain ignorance. [I accept that they may be parroting the output of paid misinformation providers.]
While AI/ML can never keep up with similarly AI produced misinformation, it certainly can reduce the cognitive load on the unpaid debunker. It would be really nice for a browser extension to flag misinformation whenever it is presented, providing a more truthful link with a text snippet to counter the flagged misinformation. It will require non-poisoned data, but that is the task of either governments or reliable corporations. While we can argue over the interpretation of facts, those facts should be unviolated and uncorrupted.
AT - I am guessing that a browser extension to flag misinformation would quickly be turned into the American version of "Social Credit Score" with the campaign against if quitely funded by Musk and his ilk who would all stand to lose if we reduced the ease with which misinformation can currently be passed around..
It will always be an arms race, with the defender trying to catch up. AI/ML just speeds things up. We let the genii out of the bottle with the section 230 that allows social media to be not liable for propaganda and disinformation. I don't think this was wrong, but the consequences are difficult to deal with. But is was ever thus, ever since the printed word could mass produce written thoughts and ideas. I really don't want to see authorities shutting down the modern electronic equivalent of pamphleteers, as we see in more authoritarian nations. What I would prefer is better means for individuals to cope with bad information. Personally, I believe we have to fight this individually, but with help from media. Just as I can avoid propaganda from Fox "News" and Prager University, and get less biased information from good newspapers, what I want is a way to use those selected media to help me as a guard from disinformation grabbing my attention. This would act as a filter as I browse the web. In essence, it would provide an extra layer of links and information that would highlight mis- and disinformation. Of course, which filter[s] I select will differ from those others select. That is inevitable. But again, this has always been the case, it is just that information now flows more like a from a hosepipe than a cup, vying for my eyeballs. Recall when Snopes was around to debunk lies? I see AI as an extremely fast and effective way to do the more sophisticated version of a Snopes debunking. Facts should be paramount, and then interpretations and opinions should be secondary. I doubt my interpretations and opinions are correct all of the time, but I do want to have them biased towards the "truth", and I particularly want to avoid being swayed by lies. If the public sphere could operate more like the scientific method, self-correcting, that would be ideal. Sadly that can only happen over the long term with experience, and even then that experience can be lost after a couple of generations.
AT - I agree, trying to shut it all down will never be the solution, it is the road to authoritarianism.
I think the answer an be found by looking within.
I agree sources like Prager U and Fox push misinformation for the right. What about starting with misinformation from the left? Our side is just as guilty as the right.
Great thinkers like Noam Chomsky exposed how the media on the left and right frequently give media viewers false choices. For example the media debated about whether or not the US should go to war with Iraq, when they should have been debating whether or not the justification, WMD, was real in the fist place. Or during the financial crises the media had media viewers debating whether or not to bail out banks during the financial crises. When they should have paid equal attention to whether or not to bail out home owners.
In both cases the media on the right and left manufactured a false debate that turned out to ignore other more relevant issues.
Now we are facing misinformation hyper weaponized by social media. Maybe the better choice is to start to self report Misinformation on our own side. Start holding media accountable for their own misinformation. What role did lliberal main stream media play in getting the US into Iraq or into bailing out banks and not home owners. What about Russia Gate, was that reporting in line with what ultimately came out in the Mueller report. I say we start holding our own side accountable.
We may never get the right to reflect on their own consumption of misinformation but we should be able to get our own side to re think what we are doing with misinformation.
"Maybe the better choice is to start to self report Misinformation on our own side. Start holding media accountable for their own misinformation."
Without some help, that is the problem highlighted by Noah Smith. Maybe AI helps, but what you have is the equivalent of moderating at scale - in this case reporting at scale. What does the reporting do - if anything? I still think it is better to have a tool that just flags suspect content, perhaps at the granularity of the sentence level. The data to do this is a curated set based on specific reliable content. Since so much content in books has been digitized, a curated set of books in addition to reliable online content would be the data for the AI to draw upon. [No "hallucinating" allowed]. STEM material is relatively easy to curate. It is the "squishier" subjects that are harder. Can an AI handle the opposing theories in economics or politics? Would accounts of historical facts be different between historical accounts and how to decide which is closer to the truth?
The left used to be kookier, but I think that has been eclipsed by the right. There is a difference between advocating for, say Marxist Socialism, and bad faith propaganda for neo-fascism. This is no longer a "both sides are equally bad" position. Pick areas of disagreement between left and right. Abortion, climate, how to manage the economy. Which side is promoting more outright lies? Yes, we should all be called out on misinformation, but I think the right wing is the greater purveyor of misinformation.
AT - Do you think that people on the right would opt in to real time fact checking? I don't think so. And if the fact checking forced people on the left to rethink old tropes that many of us see as Gospel ("winability" should determine who we support in a primary, we can't afford M4A, Trump was controlled by Putin during Russia Gate. etc...) then i think you would get back lash from the left.
Have you ever debated a real partisan. They don't care about TRUTH. They want to win or "own" who ever they are talking to... They don't want to find common ground or room for compromise.
With AI we can start to score media sites for both misinformation and for bias. For example Fox online and TV reports disproportionally on illegal immigrants or crime in communities of color relative to issues Americans care far more about like Wages and health care.
I think we can use AI to evaluate far more content and SCORE publishers for Misinformation and Headline Bias... And that would certainly influence what people on the left and right consume.
Nobody likes real time fact checking. Even ChatGPT gets huffy if you correct its BS. I most people who can apply logic use it to win an argument. Facts can be selected to do that, even if biased. However some issues are not determined by facts and logic, but emotion, for example politics. I have watched debates where the arguments used have not swayed the audience more than an iota, even when one side of the argument seems compelling.
I used to have a good friend who was very partisan, and he would simply deny he was partisan. Facts would just be denied with different facts. On some issues compromize could be gained, but very often not. The other problem was knowledge. It is hard to persuade some if they have little knowledge of the facts and can therefore ignore them.
I have a book for scientists to persuade non-scientists. It suggests not just using facts to appeal to the intellect, but emotion and other means to appeal to the heart and the "gut". People like me with a science background prefer factual arguments. Others find those arguments unappealing.
AI can be applied to check facts, but not bias. After all, we select information that appeals to our interests. Bias would be very obvious when rating academic subject magazines. Even news media will select different topics for articles and news.
I don't know what the answer is other than educating children, but as we have seen, that is subject to political pressure. Liberal and conservative views will always differ. Isn't that clear from history?
I agree with Gurri on identity. It is a a "perpetual conflict machine" Class, however is not a perpetual conflict machine. The media demonstrates that daily, and recently exemplified by the massive amount of attention paid to a few billionaires drowning on a submarine while daily not reporting on the deaths of despair witnessed as rural America or deaths of misery from our homeless population.
This media was able to unite Americans to fight wars in Vietnam and Iraq and fund a proxy war in Ukraine. But it can't seem to unite Americans over providing FREE health care to an aging and increasingly obese and unhealthy population.
I am reading Daron Acemoglu and Simon Johnsons book Power and Progress where they look at why societies care about the plight of the lower classes or sometimes simply do not. Unlike "identity" class wont go away. Century to century, nation to nation, economic system to economic system it wont go away. Identity will die because as Gurri points out, it is pointless. Class is not pointless.
Brad,
This is the best info available on the Bay Area office market. https://officetimes.com/corporate-office-perspective/
Wong: Translation: some lenders may not get all their money back especially, if they do not let borrowers reduce rental rates of the property they have let for and especially if local land use regulation do not allow owners to shift the uses of the property that is rented. The latter is a policy issue. The former is not.
re: Blood and Brains.
"It looks like something major changed with the appearance of Homo erectus, either on the supply side – improvements in food supply making brains more affordable – or on the demand side – a greater fitness payoff to a high energy brain – or both."
More likely it is the change in diet, and the use of fire to cook foods by <i>H. erectus</i> allowing for a change in jaw muscles and attachments, with energy from a smaller digestive gut being redirected to the brain.
This is supported by changes in dentition, mandible shape, and mandible movement. <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5943523/">Unexpectedly rapid evolution of mandibular shape in hominins</a>.
This does not discount other factors, especially sex selection for "smarter" individuals. However, the larger brain size may also have been limited by the size of the birth canal through the pelvis. Childbirth was a major cause of women's mortality until the 20th century.