Just a Thought

A thread for thoughts or questions that pop into your head that don’t necessarily warrant a thread. Here’s one that came to mind just now.

In some of the news I see today, I react with dread. For example, I saw this tweet:

Sometimes I feel like we’re doomed because we’re stupid or dysfunctional. But then this thought came to mind: 40 years ago, somebody likely felt the exact same way I did. I mean, if you go back to the Cold War, a nuclear apocalypse wasn’t something people worried about, over-population, the ozone layer disappearing. My guess is that you could go back forty years from that point, and someone would be feeling–with justification–a similar level of dread and doom.

Here’s what I’m wondering. What if the world has always had really dreadful problems, but the difference is that an individual, at a certain point in their life, gains enough knowledge and insight to actually become aware of that? In other words, the world isn’t really worse off per se, but it feels that for the individuals who have reached a point of awareness.

Actually, I don’t know if I’m comfortable with this. The problems now could indeed be worse than in the past; and even if they’re not, they could be bad enough to be seriously catastrophic. That is, there’s no guarantee that human life will continue.

Still both situations could be true: Individuals who reach a certain point may realize how bad things are, and things may also be uniquely bad. I don’t know, it’s just a thought.

38 thoughts on “Just a Thought

  1. if you go back to the Cold War, a nuclear apocalypse wasn’t something people worried about, over-population, the ozone layer disappearing.

    I disagree. I remember us hearing about all three of these things repeatedly. Forty years ago we were nine, so maybe you don’t remember (we were so young!), but that stuff was out there. I remember looking at my dad’s can of shaving gel, and reading “This product contains no chlorofluorocarbons known to harm the ozone layer.” When I first read that, I thought the ozone layer must be something dermal, so I asked my dad which layer of our skin was the ozone.

    The Day After was 35 years ago, which I agree isn’t 40 years ago, but we didn’t go from no concern to major concern in just those 5 years.

    And I hazily remember a PSA similar to the Indian at the side of the highway crying because people were littering, only it was about overpopulation. Oh wait! I just did a quick search and this came up. Is this what I remember?

    Anyway, your main point is definitely something worth considering even though I disagree with this one supporting argument.

  2. You’re right–many people were aware of those problems. I remember being aware and anxious about them myself as a child. What I’m talking about is developing a greater depth and breadth, not only of the problems, but the means to solving them. For example, my level of understanding of human nature, government and politics, is broader and deeper now than it was in the past, and this understanding can make one even more pessimistic, perhaps intensifying a sense of doom. Does that make sense?

  3. How does the press eschew both-sides approach in an asymmetrical situation while still maintaining trust with the public?

    I think I’ve brought this topic up several times before, so I’ll put this here, instead of in a new thread. If the press covers an issue or politicians, where the facts, logic, ethics heavily support one side over the other, how does the press report the story this way–versus presenting both sides as relatively equal–without losing the trust of the public? This unbalance reporting will seem like the press is taking sides, which will undermine the public’s trust. I also think journalists have a psychological warning system that will make unbalanced reporting really difficult, even if it is totally justified and appropriate. How do we get around these problems?

  4. At some point, “offline” will be “in” and “online” will be “out.”

    Here’s a kind of trajectory for online activity that I’m noticing:

    1. Information will become easier and easier manipulate and manufacture;
    2. Information security and privacy will either decline or will always be significant vulnerable;

    The consequence?

    1. People will have more difficulty knowing what is real and what is fake–and if that trend continues it will be impossible to distinguish the two;

    2. Individuals, businesses, governments (including foreign governments), etc. will be able to manipulate and track individuals at an unprecedented level.

    If these things are true, a reasonable response would be for people to migrate away from the internet. This creates an opportunity for offline services, institutions–to address what these people have given up. Individuals and groups providing information/news should think of ways to capitalize on this. Businesses can create offline services as well. I’ll try to write more about this later.

  5. Okay, Reid. This is just for you.

    Imagine two coins. Coin A and Coin B.

    Coin A is tossed 1,000,000 times. About 500,000 times it lands on heads, and about 500,000 times it lands on tails, in no discernible pattern.

    Coin B is tossed 1,000,000 times. About 500,000 times it lands on heads, and about 500,000 times it lands on tails, in no discernible pattern.

    Coin A is then tossed 1,000,000 times more. This time, it lands on heads 1,000,000 times.

    Coin B is then tossed 1,000,000 times more. This time, it again lands on heads about 500,000 times and again lands on tails about 500,000 times, in no discernible pattern.

    Now someone forces you to wager your life on the next coin toss. The coin must land on heads, or you forfeit your life, but you get to choose which of the two coins is tossed. Do you choose Coin A or Coin B?

  6. I have no idea how what you’re saying relates to why my previous post is “just me”…unless you mean: My crazy questioning of coin flips is just like what I’m suggesting above. 🙂

  7. It’s just for you because you’re the only one I have this point to make with. Are you going to answer the question?

  8. If you choose Coin A, you don’t actually believe that 1,000,000 heads in a row on a truly random coin would have to pull slightly in favor of tails. I understand why you have stuck to that position for 20 years, but with your life on the line, you don’t actually believe it, do you? Because by your theory, you’re more likely to flip tails with Coin A. Unless you have a death wish.

  9. I’m a little confused, here–and I honestly I didn’t really think a lot about my answer. I chose A because heads kept appearing. I assumed it’s not really a balanced coin. Or, was the first 1,000,000 flips supposed to prove that the coin was balanced? Can you explain what your thinking was in setting up the question.

  10. Yes. The first 1,000,000 flips demonstrates that it’s a fair coin. And not thinking about your answer is the right approach anyway, because I asked the question to see what you believe about the coin.

  11. I didn’t conclude that the first 1,000,000 flips established that, though. The situation you pose is bizarre and highly unrealistic. If you recall, in the examples I give, I offer a much smaller number of flips landing all on one side–and I do that because I’m trying to be realistic. 1,000,000 heads in a row would cause me to doubt the fairness of the coin. How could something like that happen? If the we could flip the coin an infinite amount of times, then that isn’t so unusual or unrealistic, but we can’t do that with a real coin.

  12. So if we’re certain it’s a fair and balanced coin, and let’s say we see 40 consecutive heads in row. Let’s also assume that we have data. Specifically, we’ve flipped a coin a trillion times (generational project) and we’ve found that coin 40 in a row has only appeared two or three times. I’m going with coin B.

    The key part to this has been the thinking/explanation. My thinking has been the following:

    1. If we find the higher number of consecutive flips landing on one side is less likely as the number of flips increases–i.e., 2-10 flips in a row may be so unusual, but as the number increases 15, 20, 25 30–in a row–the likelihood becomes a lot less*, then

    2….if we’re flipping a coin and we get to 40 in a row, the chances that the streak will continue start to diminish, if the coin is truly fair.

    The point isn’t that I really believe this is the case. What I’m looking for is a good explanation why this isn’t case.

    (*The key here is that infinity is excluded as a range. Theoretically, with an infinite number of flips than 1,000,000 heads in a row would not be unusual or rare. But I’m saying infinity isn’t real–an actual coin doesn’t have an infinite number of flips.)

    (I can’t believe you’re bringing this up. I thought you hated this topic.)

  13. Making Good Governance Exciting to More People

    There are things that occur in government–things that good politicians do–that just aren’t sexy or exciting for the average person, but are really important to everyone. For example, ensuring that an international system based on the rule of law and human rights is strong and functioning is really important for every American. Why? Because I believe helps prevent major wars (as in world wars)–wars that would likely draw in the U.S., which would obviously affect many average Americans. And there are other reasons this type of system is important.

    My sense is that many Americans don’t really care about such a system; they don’t care if a politician does something to strengthen or stabilize it. And this is just one example of the type of policy that many people are largely indifferent to.

    What comes to mind is a way we can make this more meaningful and interesting to the average American–to get people to care about these things when voting for politicians. I don’t really have any answers, but I hope we can figure out a way to improve this….

  14. I don’t know. I think if we focus instead on helping our fellow citizens understand what good political candidates and governmental appointees are, they don’t really have to know the details of what makes it all work.

    You walk into a high-rise in downtown Honolulu without knowing what the building inspectors, construction supervisors, engineers, architects, plumbers, or electricians did to make it a safe building. I’d say this stuff is equally (and more immediately) important.

    Whatever we’ve got going on now with government is flawed, but it can be made better without educating everone on specifics. Don’t you think?

  15. I think if we focus instead on helping our fellow citizens understand what good political candidates and governmental appointees are, they don’t really have to know the details of what makes it all work.

    I suspect we’re thinking of this in very similar ways (but I’m not sure). I expecting people to know a lot of details about how government works or about policy is realistic–that’s not what I’m asking for. My example about international system involves broad, conceptual understanding, more than understanding specific details.

    I also agree that we should give people a better understanding of what makes a good politician and public administrator. Actually, when I thought of this post, I thought of George H.W. Bush. I honestly thought he wasn’t a very good candidate/president primarily because he was dull and un-charismatic. It’s clear to me that this is a bad way to judge him, and he might have did a more than I realize that helped our country and the world. But those things either didn’t get enough attention, or they didn’t interest me enough. What I’m suggesting is that we find a way to make that type of information more interesting and meaningful to voters–mainly by helping them understand why these things matter. Another possible way could be to create certain benchmarks–benchmarks that may not see so exciting, but are nevertheless important–track them, celebrating them when we move closer to achieving them.

  16. A friend of mine recently went on a trip to New York City. She goes twice a year to catch theater, but this was the first time she took a day to visit Ellis Island. “The Statue of Liberty is a lot smaller than I always pictured it,” she wrote.

    I feel the same way.

  17. Enlist Hollywood in the Fight Against Authoritarianism

    The fight against authoritarian regimes, ideas and information will be a key element. Remember WWII propaganda films? I’m a little uncomfortable about any sort of political propaganda, but I think liberal democracies will probably have to engage in this, and I think Hollywood should join in. And really, in terms of entertaining stories with compelling heroes and villains, there is a treasure trove. With action/spy films alone, there’s a bonanza. As an example, look below (I believe it’s real):

    This sounds like something out of a Quentin Tarantino film or a maybe a Marvel Comics movie. Some like Vladmir Putin, who is a combination of a KGB agent and crime-family boss is a great concept for a villain.

    Action/spy films aren’t the only genres with potential. Dramas with journalists as the heroes–with themes reaffirming Enlightenment values; or historical films that celebrate the best moments of our democracy; or films that tear down walls between whites and non-whites, reducing the fear of the Other.

  18. Originally I wanted to write this post as a separate thread, but I fear I’m flooding the site with too many posts on politics or other topics that others aren’t interested in. The post has to do with observations and insights I’ve gained from consuming a lot more news, starting at the 2016 election. I wanted to talk about one specific observation–namely, the difference between consuming a lot of news, from a variety of sources on a regular basis, versus less consumption, from less sources. Here are several takeaways:

    1. I got a better sense of who is trustworthy and reliable. Over time, seeing the journalists and publications that are reasonable and reliable become much more apparent, or at least I feel far more confident in determining these things. For example, when I first started consuming a lot more news, I had no idea about the credibility of The Federalist, conservative website, and some other editors/writers like Sean Davis and Mollie Hemingway. Now, I’m pretty confident that they’re not really reliable or reasonable, but I’d never feel confident saying that if I didn’t consume as much news, on a regular basis, as I have. (I should say that prior to 2016, I got a lot of national news from magazines like The Atlantic, as well as reading some articles from national newspapers. Now, I read far more of the latter, as well read more articles from more magazines [like National Review] and follow a variety of journalists, pundits and academics on twitter.)

      The situation reminds me of my experience with getting NFL gamepass. Once I subscribed, I watched far more games (close to every game, every week), and when that I happened I grew far more confident in evaluating how good the players and teams were. I could compare players and teams to all the other players and teams in the league; I developed a baseline for different positions, offenses, defenses, etc. Now, this past season, the number of games I watched started to decline a little, and I felt less confident about my opinions. I expect the same would be true if my news consumption decreased.

    2. I didn’t have to rely on or trust journalists as much. I still hadd to rely on them to gather information, and I also had to rely on their analysis–but to a lesser degree. And I could determine not only who was more trustworthy for myself, but I could also determine which claims and arguments were credible or even compelling from the ones that weren’t. Now, it’s important to note that I could be wrong in these assessments–my biases could be clouding my judgments. But the point is that I felt more confident about these judgments, because I had been reading many sources on a regular basis for a relatively longer period of time. In my opinion, what’s happening here is that I’m forming a mental track record of the various sources, keeping score on who’s accurate and reasonable. I couldn’t do this if I weren’t consuming as much news, from many different sources on a regular basis.
    3. The people who aren’t consuming a lot of news are more vulnerable and have a greater chance of being confused and uncertain. This applies to very smart, well-educated individuals as well. If these people aren’t consuming a lot of news, I don’t see how they can feel really confident about the information they’re receiving–unless they’re really comfortable trusting certain sources, almost in a blind way. (By the way, blind trust–in reputable sources–isn’t necessarily as bad as I’m making it sound.) This is a problem, and I’m not sure there is any good answers for this.
  19. I think your numbers 1 and 2 are related. And they’re legit positions, but I’ll add one more that I hope you’re adding to your arsenal of discernment. The more you read credible sources, the more able you are to pick up hints in non-credible places that they are not credible just because of the language they use and how they use it. There’s a reason the NYT and WaPo phrases their stories the way they do, and most non-credible sources don’t work as diligently to stick to this kind of phrasing.

    A long time ago, we had a very long conversation on the phone about editorializing on the front page. You didn’t think it was a big deal, but it has never stopped bugging me. Editorial language in what is supposed to be a news piece is one betrayal of credibility, and you’ll see it a lot in most non-credible places. The more you read good journalism, the more obvious bad journalism is.

    Of course, the NYT and WaPo have more resources with whom to edit that stuff out. The Star-Advertiser is less careful about it, partly because of fewer resources, but then I’d say that also makes the Star-Ad less credible than, say, the Washington Post. This is not to say the Star-Ad is not credible; it just means that by not sticking to the conventions and ideals of good journalism, it compromises its credibility.

    The NYT, Wall Street Journal, and Washington Post in most (if not all) news stories try to demonstrate to the reader that they have exercised due diligence in fact-checking, getting verification from two sources, and asking for response comments where they should.

    And I don’t mean to keep getting back to this, but the more you read technically sound writing, the more you’ll agree with me that non-credible news sources are far (far!) more rife with errors in grammar, spelling, and punctuation. I very often don’t have to read more than a couple of paragraphs before I can decide that a source is not credible, based on the writing alone. This is not a perfect way to filter out the bad, but it’s a pretty reliable first defense.

    Have you noticed the same, with the increase in your consumption?

  20. I read an interesting content review study in 1994. Someone analyzed the languaged used in TV promos for shows like Entertainment Tonight and A Current Affair, and compared it to TV promos for convential news programs. The tabloid programs promoed themselves with words like “exclusive,” “sensational,” and “shocking,” while the news programs stuck to much more neutral language. That seems obvious, but someone actually quantified it. I notice this kind of thing all the time in websites I give no credence to.

  21. The more you read credible sources, the more able you are to pick up hints in non-credible places that they are not credible just because of the language they use and how they use it.

    I suspect that’s happening on some level–I’m honestly not fully cognizant of all the factors that help me determine if something is reliable or not.

    Editorial language in what is supposed to be a news piece is one betrayal of credibility, and you’ll see it a lot in most non-credible places. The more you read good journalism, the more obvious bad journalism is.

    If you can let me know some specific examples of publications or articles that demonstrate this connection. I’m open to checking this out.

    On a related note, for what it’s worth, in my mind, I don’t usually make a clear separation between journalism and opinion-pieces/analyses. That is, I’m not thinking, “OK, now I’m reading objective reporting,” or “I’m reading an opinion piece.” I mean, I’m aware of this sometimes, but it’s not that big of deal. Also, when journalists appear on TV (especially in group discussions), aren’t they sort of blending journalism and editorializing. The language they use seems to suggest this at least. In any event, I don’t have a problem with the editorializing. To me, the substance and accuracy of their points, the logic and reasonableness of their arguments–to me these things matter for more than subjective language.

    And I don’t mean to keep getting back to this, but the more you read technically sound writing, the more you’ll agree with me that non-credible news sources are far (far!) more rife with errors in grammar, spelling, and punctuation.

    Maybe I don’t even make to these sources, because I don’t really notice this. What sources do you have in mind? By the way, typos in major publications are not really that uncommon in my view, especially their online content.

  22. Idea: Software That Will Build Scorecards

    We’ve talked about tracking the comments and prediction of NFL draft experts, and I’m wondering if there is a way to build software that will collect these comments, and then put them in a table or chart. Let me sketch out what I mean.

    Step 1

    Input names of draft experts (any amount of them). Program will search internet for all the comments made by these experts about specific players.

    Step 2

    The program will then sort and display comments by specific experts about specific players.

    Step 3

    User can customize the tables. For example, let’s say you want to see all the comments all the experts gave about all the defensive ends, you could enter these specifications and the program would show you a table with all of this information.

    Step 4

    A separate category would be made for how each of the players are doing in the NFL. The user could either evaluate and render a judgment for how good each player is. The program could collect comments about the player online and then collect them in one spot. Over time, this assessment may change. For example a player may get significantly better in year 3.

    It should be obvious that this program could be used for other subjects besides football. For example, we could input comments made by journalists about the Trump presidency and track those comments over time. The program would make all of this more automatic and efficient.

  23. We’re Clueless About Effects of New Technology

    Or at least I’m clueless about the effects of new-ish technology…It’s hard to describe how I feel, but I would say that I have this sensation that there are many problems, without really knowing clearly what they are. It’s as if the fog is clearly, and these problems are emerging, becoming a little clearer. However, for me, I feel like I’m at the beginning of this process, that there are more problems that I’m totally unaware of and totally don’t understand. What’s troubling is that I get the sense that many of our political leaders really don’t understand these things, either. And it’s not just our political leaders.

    I’ll give one small example.

    AI Won’t Be Able to Effectively Filter Hate-Speech or Fake News

    I believe Mark Zuckerberg pointed to AI as a solution to filtering out hate-speech from Facebook. I’m very skeptical that this will be the case. His saying this, made me think of sports officiating–specifically the more judgment calls that need to be made. I’m also very skeptical that AI will someday be able to make these calls.

    Why do I feel this way? For one thing, I don’t think we could ever adequately articulate these rules, including defining key terms. If that’s true, what this means is that an referee (human being) will have to make a judgment call. AI could replace the human in making the judgment, but the latter won’t be objective–it won’t eliminate error. To believe it would seems to fundamentally misunderstand the nature of officiating these calls.

    I suspect the same is true for filtering out hate speech or fake news. There isn’t an objective way to do this. There isn’t an error free way to do this. The definition and nature of hate speech and fake news is subjective, not objective or factual. So while an AI could replace humans–and that may be preferable–we shouldn’t think that the problems we have relating to subjective judgments by humans will go away.

  24. Based on the little I know of AI, it can do a better job than humans could in this situation. If you feed the system information of human judgements filtering hate-speech it will be able to mimic this in a much better way than humans could. It’s weird because we think the computers will just look for key words or key phrases, but again based on the little I know of AI, it doesn’t just do that. If that’s all it did, it wouldn’t be considered AI. AI can actually make decisions based on past data.

    I watched a NHK documentary on AI. Two examples that stood out are 1) the computer was given a millions of games of “Go”. It learned moves never ever done in tournament games before and used it to beat the best “Go” player in the world. It learned it, versus how Deep Blue played chess. Deep Blue was able to go through all the possible combinations of moves and made the best one. That’s not AI. And 2) it learned where passengers will probably be located that needed to catch a cab based on past history. I think it used mass transit and traffic patterns and where people are located at different times to formulate where passengers will need a cab. And I’m pretty sure that changed day to day.

  25. Here’s another way to look at this: Do you think it will be possible to get almost 100% agreement on a decision regarding hate speech or something like pass interference? To be clear, there will be some situations where you would get close to 100% agreement on the right call. However, I think there are going to be some calls where that’s just not the case. And in these situations, the expertise of the human being making the judgment is not the decisive factor. The issue here is that some judgments are inherently ambivalent. In these situations, I expect that some people will always disagree with whatever judgment is made.

    By the this is one of the things that makes hate speech different from winning a game of go or catching a cab. With the latter, we have clear objectives–i.e., we know when someone wins at go or catches a cab. I don’t think the same is true for what constitutes hate speech.

  26. No I’m not saying AI will get it right all the time, but it can probably do better than 50 (or any number) of individuals reading and interpreting the same thing. Now if you are saying in cases where there is no agreement the 50 (or whatever number) will vote, I’m guessing that could be a better resolution.

    I was going to write this on my original post as well, but I wouldn’t doubt if AI will be able to understand much if not all of what is written. That’s the beauty of AI, is that it can learn like you and I. My guess is AI can learn and interpret poetry or movies I have no business watching. I think the notion of how we think of computers warps the understanding of what AI can do.

    AI invented moves that the Go champion started using in future matches. The Go announcer (as well as the champion and the audience) was baffled and aghast when the computer made the move.

  27. I think the notion of how we think of computers warps the understanding of what AI can do.

    I’m not coming from the angle of what the computer can do–but the nature of what the computer will have to analyze–namely, things that aren’t clearly defined, and can never be; things that don’t have clear-cut right or wrong answers. Winning at go or catching a cab–those things have clear-cut. Hate speech, what makes something something good art versus bad, or even what makes art at all–my sense is that these things resist clear and precise definitions, by their nature. Do you see what I’m saying?

    No I’m not saying AI will get it right all the time, but it can probably do better than 50 (or any number) of individuals reading and interpreting the same thing.

    Here’s one way this would make sense. If you had a 1,000 messages, and a 100 of them contained obvious hate speech (i.e., close to 100% of people would clearly identify as hate speech), and the AI had significant more success at finding and flagging these 100 messages, then I would say the AI is meaningfully better (assuming it also didn’t inappropriately flag significantly more messages).

    Additionally, if the AI could do this more cheaply (and I would assume at some point the AI would be cheaper than hiring 50 people), then this would be a case where AI would potentially be valuable.

    But it’s unclear that an AI will even be able to identify the obvious examples of hate speech–because again the nature of hate speech is something abstact and really hard to define–in my view, harder to define than winning at chess or go.

  28. The same doubts you would have for a person to interpret something correctly (ie: Is this hate speech or not) would be the same doubts you would have for AI. So the notion that computers are flawless sort of goes out the window with AI. They can formulate opinions or interpretations might be a better word, on things it learns. It won’t be black or white. So it will have to interpret a facebook posting as in or out like a person would. So like when the AI “guesses” where a cab fare will be, it’s far from a guarantee that someone will be there. That would be scary. It just gives you its best interpretation of the facts, sort of what you would do if you were reading a facebook post that you needed to determine whether it gets in or out. Or something a cabbie would do without AI, it would formulate where a fare would be based on past experiences (ie: where I picked someone up yesterday or last week at this same time).

    I agree speech is much harder to define, but I’m guessing AI “guys” would say if you and I can learn to interpret something, AI can as well. I could be way off base, but that’s my understanding of AI. So if I was to “teach” AI how to interpret poetry, I would give it a million different poems with maybe two or three interpretations of each poem. AI will try and learn to interpret a poem it never saw before based on what it learned from the original million poems. That’s something you and I would sort of do as well, just on a much smaller scale.

  29. The same doubts you would have for a person to interpret something correctly (ie: Is this hate speech or not) would be the same doubts you would have for AI. So the notion that computers are flawless sort of goes out the window with AI.

    I’m not as confident as you that many people would react this way. I mean, if you ask them, they’d admit that computers wouldn’t be perfect, etc. But people also say that standardized test are only one form of assessment and not the most important one, and behave in a contrary fashion. Same with sports statistics. Or look at sports officiating. We acknowledge that we won’t be able to completely eliminate errors and that judgment is part of officiating, but then why are people so mad about the few errors in judgment?

    My concern is that people will assume that because we have a powerful computer, acting objectively, it can solve the hate speech problem–with “solve” meaning almost completely eliminate it. In other words, I suspect many people aren’t thinking that AI will basically perform as well as humans. (Is that how you’re thinking about this?) If so, why hasn’t already solved this?

    Do you agree that the problem of identifying something like hate speech is subjective–that a computer operating in precise, objective fashion–won’t necessarily be better at solving this problem than a human?

    I agree speech is much harder to define, but I’m guessing AI “guys” would say if you and I can learn to interpret something, AI can as well. I could be way off base, but that’s my understanding of AI.

    Meaning, you think the AI experts believe that all problems are essentially the same–that if AI can become good at playing go, it can also identify hate speech? Somehow this doesn’t sound right–or am I skeptical about the logic here. They seem like fundamentally different sorts of problems–the latter is close to understanding language, involving the understanding the meaning and intention of the words and sentences. That’s not only highly sophisticated, but doesn’t seem like a different sort of sophistication than learning to play go? It kinda does to me.

    If I’m wrong about, then would AI also be able to moderate online discussions–identifying uncivil discourse, unreasonable claims, hostile gestures, etc. That would be really cool, if so.

    1. I’m not as confident as you that many people would react this way.

      I agree. I think most people view computers as producing whatever it can produce in a more black and white fashion. It’s either correct or wrong. But my understanding of what sets AI apart from normal “computing” is that it can learn on it’s own. How it does it, doesn’t seem possible, but that’s what I gather AI can do.

      I suspect many people aren’t thinking that AI will basically perform as well as humans. (Is that how you’re thinking about this?) If so, why hasn’t already solved this?

      Yes AI will basically perform as well as humans in terms of interpretation. But I’m guessing AI will eliminate biases that humans may have and process things much faster. Why it hasn’t already solve this? My hypothesis is that AI programing is very specific. The AI that learns to play Go and the AI that learns where a cab fare will be is completely different. I’m sure there are similarities on how it learns, but that’s where it ends.

      The examples of Go and cab fare isn’t specific to learning to interpret writings and yes it’s different. Those examples are more to show AI’s capabilities versus computing as we know it. Deep Blue could only replicate what was done before. It could only make the same chess moves or more important series of moves as humans have done before it. If you only programmed Deep Blue with elementary chess players’ abilities, that’s how it will play. AI though will learn to play better. So if it started with only elementary chess players moves, it will get better the more it played.

  30. But my understanding of what sets AI apart from normal “computing” is that it can learn on it’s own. How it does it, doesn’t seem possible, but that’s what I gather AI can do.

    What it sounds like you’re saying is that, to you, the virtue of AI isn’t that it is objective, but that it can learn; and because of this, you believe it has a good chance of filtering out hate speech. Is that right?

    Yes AI will basically perform as well as humans in terms of interpretation. But I’m guessing AI will eliminate biases that humans may have and process things much faster.

    So it sounds like you’re saying, essentially, that the AI will be better than humans–because it will be less biased (i.e., more objective) and faster. (Note: It wasn’t clear, but I was asking that if you thought AI would be just as good as humans, why hasn’t FB just solved the hate speech problem via humans. That is, why build AI to solve the problem if they’re just going to be just as good as humans. But you seem to be saying AI will be better.)

    If I’m reading you correctly, this goes back what I was saying before about identifying hate speech is inherently subjective, so thinking that objective evaluation will be superior misunderstands the nature of the problem. Judging hate speech–like judging a catch in football, determining what is art–will never be objective or quantifiable–at least, my sense is that they’re not like that. Therefore, an objective, thinking entity won’t be superior. You see what I’m saying?

    The examples of Go and cab fare isn’t specific to learning to interpret writings and yes it’s different. Those examples are more to show AI’s capabilities versus computing as we know it.

    I think I understand where you’re coming from. My sense is that a computer learning, on its own, how to improve in playing a complex game like chess, or making predictions about taxi stops routes, doesn’t mean it can identify hate speech. To be more specific, my sense is that the former involves finding patterns, and then extrapolating novel moves from those patterns. In contrast, identifying hate speech involves more than identifying and using patterns. What it also involves is understanding the meaning of words and sentences. Words don’t have simple meanings. They’re much more flexible and dynamic. Ten people can have ten slightly different ways of understanding and using the same word. This gets even more complex when the words are put together in sentences. Additionally, when people write online, they often write like how they speak in a face-to-face conversation. However, the online version doesn’t have the information that comes from non-verbal cues like tone of voice, facial expressions, and body language. I believe this is why communicating online can be even more difficult for people. I might be wrong, but what do you think? Do you think I’m wrong in thinking the difference learning game tactics and learning to identifying hate speech is so different that just because AI can do the former that doesn’t mean it can do the latter?

  31. Identifying Areas of Consensus Vs. Political Debates

    There seems to be a general belief that debates between individuals with opposing is a good way to inform the public about political issues. While I believe the approach has value, my sense is that there is a greater need to identify the areas of consensus for opposing sides. For example, what are the common set of facts that both sides agree upon? What facts are the ones they believe are most relevant and important? Are there opinions or arguments that one side can concede to the other? This sort of thing. Instead of debates–or at least in addition to them–I think news agencies and public service groups should spend greater effort on ways to identify areas of consensus.

  32. The “Chinese” Circle

    China taking it’s cue from Dave Eggers’s novel, The Circle

    Can anyone imagine a non-dystopian outcome from something like this? Are we just too biased by dystopian stories like 1984? I don’t doubt some good things could occur, especially if you look at society overall, but the government doing something like this is creepy–and I don’t think this is primarily from dystopian novels. I think it has to do with the nature of government and people in power. Being an American likely plays a big role in this. For example, people in Asian countries or former Communist countries might be more amendable to this. I wonder. Whatever the case may be, I think that even in the healthiest liberal democratic system, the abuses would far outweigh whatever benefits would be gained by the society overall. If we’re talking about a more authoritarian government like the Chinese, I’d say it’s more disturbing situation.

  33. Documentary or book on Play Calling in Football

    I’m not expert, but my sense is that fans really don’t understand what goes into successful play calling (for both offensive and defensive coordinators). Because I think a book, documentary, or maybe a multi-part netflix docuseries on a classic game, focusing on the chess match between the OC and DC (and probably head coaches as well). The series could go into the preparation (maybe have flash backs to the week of practice or what was seen in the film room). The idea is to reveal the thought process of both the OC and DC, with the idea of showing the nature of play calling.

    My guess is that many fans will be surprised at what they learn, especially the fans of analytics. My sense is that play calling involves getting a feel for the play calling for one’s opponent. There’s a kind of groove or rhythm, and the if a coordinator can tap into that, they can then anticipate and make the right calls. To get a sense of this requires accounting for many variables–the score, field position, time remaining, field conditions, tendencies of the coordinator and players, injuries, and even how the players seem to performing that day. I think coordinators are trying to get into the head of their counterparts, and I suspect the best coordinators (defensive or offensive) can do this really well. (I got this sense from an article about Mike Zimmer. After watching a few offensive series, Zimmer could predict the plays that would be called, even if he had never seen the game before.) I think analytics could play a role in this, but play calling seems to be about playing against another human being(s)–understanding and anticipating what they’re going to do.

    Whether this is right or wrong, I’d like to see a docuseries that explores the nature of play calling.

    By the way, the kernel of the idea came to me after reading Buzz Bissinger’s Three Nights in August about Tony LaRusa and the St. Louis Cardinals. After reading the book, I thought film about a famous game (or even a few innings of that game), showing the thought process of the managers and players, would be really cool.

  34. If adversaries of the U.S. are willing to steal or fabricate information and strategically use and release them to inflict maximum damage on politicians, institutions, or democratic processes, then should institutions like the press play a more active gatekeeping function for the public square? (Note: This may also apply to legitimate domestic actors as well.) Or should they disregard the potential for weaponizing information, taking the stance that this is not their jobs? (A part of their jobs is vetting information, so they would filter out fabricated information.)

    I can guess Mitchell’s response to this, but I think the internet has made the second approach far less tenable. I’m not sure if the press understands this or if there is consensus about this. Additionally, I’m worried that commercial pressure might conflict and override the this important custodial role.

Leave a Reply

Your email address will not be published. Required fields are marked *