Gregory's Server
|
21 posts total
12
It's a terrifying development. LLMs are literally designed to generate *plausible-sounding* *bullshit*. They have no accountability and even less allegiance to truth than crooked cops—but they will be much, much better at writing the kinds of falsehoods that will bring a conviction.
Show previous comments
 @ct_bergstrom when i saw this i went "yooo look at these fuckers" and showed it to my bf  So this is crazy fcuked. Despite their denial, Gizmodo caught the College Board (who administer the SAT, etc.) sharing scores and GPAs with Facebook, TikTok, etc. https://gizmodo.com/sat-college-board-tells-facebook-tiktok-your-scores-gpa-1850768077
Show previous comments
 @ct_bergstrom Some pretty crazily specific wording in their denial ""we do not share with X, Y and Z or anyone else using pixel (sp?) or cookies." This immediately implies that they do share via some other mechanism...and then they get caught sharing using the mechanism they denied to begin with.... @ct_bergstrom that's absolutely wild. Isn't it a crime to do this, especially when it would involve minors, especially with data privacy since you would have to acknowledge you consent to having data shared with third parties?  Yo that's fucked up. "We don't do the thing." "Here's video evidence of you doing the thing." "Oh we're caught? lol yeah we do the thing." We need the GDPR so bad in the U.S. it's just silly. Welcome to everyone arriving from...elsewhere. Here, lots of stuff about social media and science, but aloso ravens, crows, birdwatching, and birds, watching. My colleague Kevin Gross and I have a new preprint up on the arXiv. Just for fun, rather than pure text simple text explainer, let's go some slides for a talk I'm giving at https://www.icssi.org/ tomorrow. Here's the paper itself: Rationalizing risk aversion in science. https://arxiv.org/abs/2306.13816 The basic issue at hand is high-risk, high-return science. There is widespread sentiment, and even some scattered empirical evidence, that scientific research within academia is too cautious and that higher-risk, higher-return research would yield more progress more quickly. Listening to very smart people talk about #GPT4 I'm reminded of the joke about a checkers-playing dog. A guy has a dog that plays checkers. "My goodness," everyone says, "that's amazing. What a brilliant dog!" "Not really," he replies, "I beat him four games out of five." That's GPT4. It's capacities are amazing and completely unexpected. But it's also so limited. You shouldn't back the dog in a checkers tournament, and you shouldn't use an LLM as a medical assistant or in many other ways. I feel like this joke captures a great deal of the dialectic that I see here and in the broader discussion around LLMs and AI right now. I spend most of my time writing about how baffled I am to watch Microsoft and Google betting their futures—and to a degree, ours—on this dog's performance in the World Checkers Championship. Other colleagues are legitimately amazed that the dog can play chess at all, and want to understand how well it plays, and how it manages to do it in the first place.
Show previous comments
 @ct_bergstrom that's quite a read. Thanks for sharing. Will be adding her books to my and my kids reading list.  I'm fascinated by the reciprocal nature of #birdwatching. Birds are every bit as much watching us as we are watching them. In my photography I try to capture bidirectional interaction in my photography. I photograph birds watching me, but only when they can freely approach, and freely depart. I'm interested in their curiosity, not their fear. This is a wild bird that approached me.
Show previous comments
@ct_bergstrom Our chickadees here are pretty much exhibitionists, flipping and flapping wildly as they bump and grind their bird parts together mid-flight, just about bouncing off my head in the process. @ct_bergstrom Actually reading The wake of Crows (https://www.thomvandooren.org/publications/books/wake/) and it is very good! @ct_bergstrom A really cool thing I read about is the parrots of Central and South America. They aren't "domesticated" per se, as in they weren't intentionally bred like dogs or livestock, but in urban areas, especially places straddling a city and the forest, these wild animals seem to just "know" how to interact with humans, they'll perch near you to ask for food and can recognize who feeds them most, and will preferentially go to the people who feed them as well as tourists who do so as well.
Show previous comments
 @ct_bergstrom I don't miss Twitter. But I occasionally miss some of the people who have made the bad decision to stay on Twitter.  @ct_bergstrom So instead of you wading through 10s of thousands of tweets for a good one, just let the internet do it's thing. Here is your must-read article for the day, a profile of @emilymbender, and her efforts to deflate the ridiculous hype around large language models such as ChatGPT. It's also about the people who are behind that hype, and about what their way of thinking has the potential to do to us. It's worth reading all the way to the end. https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html
Show previous comments
  @ct_bergstrom @emilymbender I've been speaking and writing lately about how tech companies try to shift the discussion on misinformation, polarization, harassment, etc., away from the systems and structures that are inherently toxic and toward questions of individual behavior. This way they can blame their own users for any pathology and steer clear of calls for systemic change. Today Elon Musk has come through with a perfect illustration for my future talks. This should be obvious, but having an algorithm that behaves that way is a DELIBERATE CHOICE. It would be easy enough, for example, to implement basic sentiment analysis so that the algorithm doesn't boost content that you have reacted negatively to in the future. Musk is playing it both ways. He keeps the algorithm that boosts inflammatory content and drives the online conflicts that draw views and clicks, while pushing the blame for this off onto the individuals involved. That sucks. For the holiday, a thread on how to befriend crows. -- Befriending crows is a wonderful thing. I have many crow friends at home and at work. They bring joy at unexpected moments and can rescue a miserable day even without shaking down the dust of snow that Robert Frost described. This thread is an updated version of one I posted at the bird site in July 2019. #birding #birdwatching #birds #urbanbirding #crows #corvids #crow #corvid #crowfriends
Show previous comments
@ct_bergstrom Fascinating. Thanks for the great thread. I never knew how responsive and intelligent they were. Animals and birds are more aware than us humans give them credit for. Great pics too. Inspired by today's travel problems, I wrote something about why corporate customer service systems fail during crisis events. I like the formatting over at post.news (much as I dislike the ownership), so you can read the piece there: https://post.news/article/2JLN7sXUcrlRe2HmLSdcjBzZ5LA I'll also "serialize" it here, below.
Show previous comments
 @ct_bergstrom Shame it's a walled garden so we can't actually read the post over there (from here at least).   At age ten, I started #birding after moving to Australia and falling in love with the colorful and charismatic parrots of Canberra. Upon moving back to Michigan a year later, I kept it up and loved it—but after Australia, I just couldn't really get into sparrows. Too little-brown-bird for me. Funny how time changes you. These days, I adore the precise and exquisite tweeds of North American sparrows. Below, a lovely Lincoln's sparrow taken in Seattle last week. Today was the straw that broke the camel's back for me. This morning Elon Musk tweeted that his pronouns were prosecute/Fauci. You can’t study evolutionary biology under the roof of the Discovery Institute and you can’t have meaningful and productive scientific collaboration on a platform run by right-wing troll who denies science when its results are inconvenient to him and just simply to hear his audience cheer. Twitter has been an amazingly valuable platform for scientists over the years, and it was integral to coordinate global scientific activity during the early part of the COVID pandemic. Sadly, we live in a world where one malignantly narcissistic billionaire can destroy all of that for lulz and the adulation of Marjorie Taylor Greene. I've locked my account and will be deleting the content as I figure out the right way to do it. 12
|
@ct_bergstrom and that is just the beginning. Have you seen this? https://sakana.ai/ai-scientist/ (link to arXiv is also in the post)
@ct_bergstrom funny how the people who have the most money in AI are also the ones who have funded or made money from disinformation 🤔
@ct_bergstrom ok so not only do LLMs suck more energy than we can decently produce, and destroy the internet by flooding it with more BS than we can debunk, but its next target is now science.
Of course, LLMs are not a scient thing trying to fuck with humanity. We are doing this to ourselves in enlightened consent.