• 0 Posts
  • 151 Comments
Joined 2 years ago
cake
Cake day: July 31st, 2023

help-circle
  • I am very, very concerned at how widely it is used by my superiors.

    We have an AI committee. When ChatGPT went down, I overheard people freaking out about it. When our paid subscription had a glitch, IT sent out emails very quickly to let them know they were working to resolve it ASAP.

    It’s a bit upsetting because may of them are using it to basically automate their job (write reports & emails). I do a lot of work to ensure that our data is accurate from manual data entry by a lot of people… and they just toss it into an LLM to convert it into an email… and they make like 30k more than me.


  • The few times I’ve used LLMs for coding help, usually because I’m curious if they’ve gotten better, they let me down. Last time it was insistent that its solution would work as expected. When I gave it an example that wouldn’t work, it even broke down each step of the function giving me the value of its variables at each step to demonstrate that it worked… but at the step where it had fucked up, it swapped the value in the variable to one that would make the final answer correct. It made me wonder how much water and energy it cost me to be gaslit into a bad solution.

    How do people vibe code with this shit?



  • As I understand it, it’s atomic Fedora with virtually everything you might need to game on Linux baked in (no need for layering) and more or less preconfigured. Off the top of my head, proprietary Nvidia drivers, Steam, Lutris, Hero launcher, support for Xbox One wireless controller dongle, plus a number of useful tools like Tailscale. An app with a catered list of gaming-oriented flatpacks, one click updating. Also a lot of effort into replicating the Steam Deck experience for handheld devices or devices connected to a TV.

    I believe they also do Aurora, which is similarly geared toward workstations with a ton of container-related tools like distro box readily available to easily use containers instead of layering where possible. The same tools may be available in Bazzite but I never checked. I have Aurora on my laptop and use a dedicated gaming device with Bazzite.

    I’m not a Linux veteran by any means but I was hopping distros looking for something I could install on my family’s computers I tried atomic Fedora. When using it for myself, I became frustrated with the number of tools I use that needed to be layered or run in a container and eventually found myself on Bazzite and Aurora. So far so good.












  • Another comment links to an article that explains it generally. NETs and “microclots” can clump together.

    One possibility first raised by physiologist Resia Pretorius of Stellenbosch University in South Africa in 2021 is microclots. These are tiny, abnormally persistent blood clots that are smaller than those seen in conditions such as stroke or thrombosis, yet large enough to hinder blood flow through capillaries.

    Meanwhile, in 2022, Thierry and his colleagues showed that patients with long COVID have elevated levels of neutrophil extracellular traps, or NETs. These are sticky webs of DNA and enzymes released by white blood cells to capture and contain pathogens invading the body.

    Normally, NETs do their job and then quickly break down, but when they are released in large numbers or persist longer than needed, they can contribute to blood flow problems such as thrombosis and atherosclerosis.

    The new research – a collaboration between Pretorius and Thierry – suggests that these two separate markers, NETs and microclots, may interact in the blood of long COVID patients.



  • It is like encouragement for the thing you were already likely to do, which is the goal of targeted advertising.

    It’s the claim of targeted advertising. The person I saw talking about this actually ran the numbers, comparing two very similar geographic markets. In market A they paid for advertising, but in B they did not.

    When comparing market A to market B, market A had a marginal increase in sales for the advertised product vs. market B. However, they were charged for orders of magnitude more conversions than the actual increase in sales.

    The idea is that when compared to something like actual click-through purchases, where a user literally clicks on an ad and then buys a product, it’s extremely deceptive.



  • There is actually an argument that advertisers like Google are abusing micro targeting to extract advertising revenue from clients while, at least in some cases, delivering few actual new customers.

    Here’s the process.

    1. Google sees that your profile (browsing habits, demographics, search patterns, etc) suggest you are interested in product A.
    2. Google blasts you with advertisements for product A, essentially marking your browser session and claiming you as a recipient of their advertising. Ever look at a particular product and find you are being advertised for that product incessantly for a while?
    3. If you happen to buy product A around the time that your session was shown an advertisement for that product, Google claims you as a conversion and gets paid for convincing you to buy the product. Advertising works!

    So if Google’s algorithm thinks you are already going to buy product A, they show you an ad for product A constantly because it means they’ll claim you as an advertising success and get paid extra.


  • Are we now protesting that they reversed their decision?

    …no? I’m not really protesting so much as offering what I think the other person is trying to say. I think they are saying that Google crossed a line, and walking it back doesn’t change that fact.

    In my opinion, Google has crossed countless lines over the last 5-10 years. I’m looking for alternatives that meet my own needs. That search has accelerated over the last few years, when the things Google has done have been most egregious. This isn’t a protest. This is disillusionment. I’m abandoning ship.