Let's Know Things cover art

Let's Know Things

Let's Know Things

By: Colin Wright
Listen for free

About this listen

 A calm, non-shouty, non-polemical, weekly news analysis podcast for folks of all stripes and leanings who want to know more about what's happening in the world around them. Hosted by analytic journalist Colin Wright since 2016.

letsknowthings.substack.comColin Wright Politics & Government
Episodes
  • AI-Associated Delusions

    AI-Associated Delusions
    Jul 15 2025
    This week we talk about AI therapy chatbots, delusions of grandeur, and sycophancy.We also discuss tech-triggered psychosis, AI partners, and confident nonsense.Recommended Book: Mr. Penumbra's 24-Hour Bookstore by Robin SloanTranscriptIn the context of artificial intelligence systems, a hallucination or delusion, sometimes more brusquely referred to as AI BS, is an output usually from an AI chatbot, but it can also be from another type of AI system, that’s basically just made up.Sometimes this kind of output is just garbled nonsense, as the AI systems, those based on large language models, anyway, are essentially just predicting what words will come next in the sentences they’re writing based on statistical patterns. That means they can string words together, and then sentences together, and then paragraphs together in what seems like a logical and reasonable way, and in some cases can even cobble together convincing stories or code or whatever else, because systems with enough raw materials to work from have a good sense of what tends to go where, and thus what’s good grammar and what’s not, what code will work and what code will break your website, and so on.In other cases, though, AI systems will seem to just make stuff up, but make it up convincingly enough that it can be tricky to detect the made up component of its answers.Some writers have reported asking AI to provide feedback on their stories, for instance, only to later discover that the AI didn’t have access to the stories, and they were providing feedback based on the title, or based on the writer’s prompt—the text the writer used to ask the AI for feedback. And their answers were perhaps initially convincing enough that the writer didn’t realize the AI hadn’t read the pieces they asked them to criticize, and the AI systems, because most of them are biased to sycophancy, toward brown-nosing the user and not saying anything that might upset them, or saying what it believes they want to hear, they’ll provide general critique that sounds good, that lines up with what their systems tell them should be said in such contexts, but which is completely disconnected from those writings, and thus, not useful to the writer as a critique.That combination of confabulation and sycophancy can be brutal, especially as these AI systems become more powerful and more convincing. They seldom make the basic grammatical and reality-based errors they made even a few years ago, and thus it’s easy to believe you’re speaking to something that’s thinking or at the bare-minimum, that understands what you’re trying to get it to help you with, or what you’re talking about. It’s easy to forget when interacting with such systems that you’re engaged not with another human or thinking entity, but with software that mimics the output of such an entity, but which doesn’t experience the same cognition experienced by the real-deal thinking creatures it’s attempting to emulate.What I’d like to talk about today is another sort of AI-related delusion—one experienced by humans interacting with such systems, not the other way around—and the seeming, and theoretical, pros and cons of these sorts of delusional responses.—Research that’s looked into the effects of psychotherapy, including specific approaches like cognitive behavioral therapy and group therapy, show that such treatments are almost aways positive, with rare exceptions, grant benefits that tend to last well past the therapy itself—so people who go to therapy tend to benefit from it even after the session, and even after they stop going to therapy, if they eventually stop going for whatever reason, and that the success rate, the variability of positive impacts, vary based on the clinical location, the therapist, and so on, but only by about 5% or less for each of those variables; so even a not perfectly aligned therapist or a less than ideal therapy location will, on average, benefit the patient.That general positive impact is part of the theory underpinning the use of AI systems for therapy purposes.Instead of going into a therapist’s office and speaking with a human being for an hour or so at a time, the patient instead speaks or types to an AI chatbot that’s been optimized for this purpose. So it’s been primed to speak like a therapist, to have a bunch of therapy-related resources in its training data, and to provide therapy-related resources to the patient with whom it engages.There are a lot of downsides to this approach, including the fact that AI bots are flawed in so many ways, are not actual humans, and thus can’t really connect with patients the way a human therapist might be able to connect with them, they have difficulty shifting from a trained script, as again, these systems are pulling from a corpus of training data and additional documents to which they have access, and that means they’ll tend to handle common issues and patient types pretty well, ...
    Show More Show Less
    18 mins
  • Pay Per Crawl

    Pay Per Crawl
    Jul 8 2025
    This week we talk about crawling, scraping, and DDoS attacks.We also discuss Cloudflare, the AI gold rush, and automated robots.Recommended Book: Annie Bot by Sierra GreerTranscriptAlongside the many, and at times quite significant political happenings, the many, and at times quite significant military conflicts, and the many, at times quite significant technological breakthroughs—medical and otherwise—flooding the news these days, there’s also a whole lot happening in the world of AI, in part because this facet of the tech sector is booming, and in part because while still unproven in many spaces, and still outright flubbing in others, this category of technology is already having a massive impact on pretty much everything, in some cases for the better, in some for the worse, and in some for better and worse, depending on your perspective.Dis- and misinformation, for instance, is a bajillion times easier to create, distribute, and amplify, and the fake images and videos and audio being shared, alongside all the text that seems to be from legit people, but which may in fact be the product of AI run by malicious actors somewhere, is increasingly convincing and difficult to distinguish from real-deal versions of the same.There’s also a lot more of it, and the ability to very rapidly create pretty convincing stuff, and to very rapidly flood all available communication channels with that stuff, is fundamental to AI’s impact in many spaces, not just the world of propaganda and misinformation. At times quantity has a quality all of its own, and that very much seems to be the case for AI-generated content as a whole.Other AI- and AI-adjacent tools are being used by corporations to improve efficiency, in some cases helping automated systems like warehouse robots assist humans in sorting and packaging and otherwise getting stuff ready to be shipped, as is the case with Amazon, which is almost to the point that they’ll have more robots in their various facilities than human beings. Amazon robots are currently assisting with about 75% of all the company’s global deliveries, and a lot of the menial, repetitive tasks human workers would have previously done are now being accomplished by robotics systems they’ve introduced to their shipping chain.Of course, not everyone is thrilled about this turn of events: while it’s arguably wonderful that robots are being subbed-in for human workers who would previously have had to engage in the sorts of repetitive, physical tasks that can lead to chronic physical issues, in many cases this seems to be a positive side-benefit of a larger effort to phase-out workers whenever possible, saving the company money over time by employing fewer people.If you can employ 100 people using robots instead of 1000 people sans-robots, depending on the cost of operation for those robots, that might save you money because each person, augmented by the efforts of the robots, will be able to do a lot more work and thus provide more value for the company. Sometimes this means those remaining employees will be paid more, because they’ll be doing more highly skilled labor, working with those bots, but not always.This is a component of this shift that for a long while CEOs were dancing around, not wanting to spook their existing workforce or lose their employees before their new robot foundation was in place, but it’s increasingly something they’re saying out loud, on investor calls and in the press, because making these sorts of moves are considered to be good for a company’s outlook: they’re being brave and looking toward a future where fewer human employees will be necessary, which implies their stock might be currently undervalued, because the potential savings are substantial, at least in theory.And it is a lot of theory at this point: there’s good reason to believe that theory is true, at least to some degree, but we’re at the very beginning phases of this seeming transition, and many companies that jumped too quickly and fired too many people found themselves having to hire them back, in some cases at great expense, because their production faltered under the weight of inferior automated, often AI-driven alternatives.Many of these tools simply aren’t as reliable as human employees yet. And while they will almost certainly continue to become more powerful and capable—a recent estimate suggested that the current wave of large-language-model-based AI systems, for instance, are doubling in power every 7 months or so, which is wild—speculations about what that will mean, and whether that trend can continue, vary substantially, depending on who you talk to.Something we can say with relative certainty right now, though, is that most of these models, the LLM ones, at least, not the robot-driving ones, were built using content that was gathered and used in a manner that currently exists in a legal gray area: it was scraped and amalgamated by these systems so that ...
    Show More Show Less
    18 mins
  • Hurricane Tracking

    Hurricane Tracking
    Jul 1 2025
    This week we talk about the NOAA, FEMA, and the SSMIS.We also discuss Arctic ice, satellite resolution, and automated weather observation stations.Recommended Book: Superbloom by Nicholas CarrTranscriptThe National Oceanic and Atmospheric Administration, or NOAA, is a US scientific and regulatory agency that tackles an array of environmental, climatic, and weather-related issues, alongside its responsibilities managing oceanic ecosystems.So it’s in charge of managing fishing protections and making sure endangered species within US waters are taken care of, but it also does scientific exploration—mapping the ocean, for instance—it monitors atmospheric conditions and keeps tabs on the various cycles that influence global and US water, air, and temperature happenings, and it tracks macro- and micro-scale weather events.That latter responsibility means NOAA (which is the modern iteration of several other agencies, including the US Environmental Science Services Administration and the US Weather Bureau) also manages the US National Weather Service, which is the sub-agency that sends out hazardous weather statements when there are severe storms or tornadoes or other weather-related events of note in a given area, and which also provides weather forecast information that local experts on the ground use to make their own predictions.Most of what the National Weather Service puts out is in the public domain, which means anyone can access and use it, free of charge. That’s a pretty big deal, because the data they collect and informational products they distribute, including all those hazardous weather statements, are at times life and death, but they’re also a big part of what makes standard local weather services possible in the US—they help the FAA and other agencies do their jobs, and they help everyday people understand how hot or cold it’s going to be, whether to pack and umbrella for the day, and so on.To accomplish all this, the NOAA and its sub-agencies make use of a bunch of facilities and other tracking resources to collect, aggregate, and interpret all those data points, crunching them and spitting them back out as something intelligible and useful to their many end-users.They’ve got weather observation stations across the US, many of them automated surface observing stations, which are exactly what they sound like: automated stations that collect data about sky conditions, wind direction and speed, visibility, present weather conditions, temperature, dew point, and so on—most of these are close to airports, as this information is also vital for figuring out if it’s safe to fly, and if so, what accommodations pilots should be making for the weather and visibility and such—but they also collect data from smaller weather stations scattered across the country, around 11,000 of them, many operated by volunteers under the auspices of an effort called the Cooperative Observer Program that was established in 1890, and that’s paired with another volunteer data-collection effort called the Citizen Weather Observer Program.There are also weather buoys and weather ships lingering across the surface of the ocean and other bodies of water, tracking additional data like sea surface temperature and wave height at various points. And there are weather balloons which collect additional information about happenings further up in the atmosphere, alongside the many satellites in orbit that capture various sorts of data and beam that data down to those who can make use of it.Again, all of this data is collected and crunched and then turned into intelligible outputs for your local weather forecasters, but also the people who run airlines and fly planes, the folks out on boats and ships, people who are managing government agencies, scientists who are doing long-term research on all sorts of things, and everyday people who just want to know if it’ll be sunny, how hot it will be, and so on.There’s one more major client of the NOAA that’s worth noting here, too: the Department of Defense. And that relationship is a big part of what I want to talk about today, because it seems to be at the root of a major curtailing of weather-related data-sharing that was recently announced by the US government, much to the chagrin of the scientific community.—US President Trump has long voiced his skepticism about the NOAA.There have been claims that this disdain is the result of the agency having called him out on some bald-faced lies he told about hurricane projections during his first administration, when he reportedly altered an NOAA hurricane impact projection map with a Sharpie to support a misstatement he had previously made about a hurricane impacting Alabama; the hurricane in question was not anticipated to hit Alabama, Trump said it would, and he later altered a map in order to make it look like he was right, when all the data, and all the experts, say otherwise.Whether that’s true or not, the NOAA ...
    Show More Show Less
    16 mins

What listeners say about Let's Know Things

Average Customer Ratings

Reviews - Please select the tabs below to change the source of reviews.

Audible.com.au reviews

Amazon Reviews
No Reviews are Available
In the spirit of reconciliation, Audible acknowledges the Traditional Custodians of country throughout Australia and their connections to land, sea and community. We pay our respect to their elders past and present and extend that respect to all Aboriginal and Torres Strait Islander peoples today.