Lettuce eat lettuce

Always eat your greens!

  • 9 Posts
  • 633 Comments
Joined 3 years ago
cake
Cake day: July 12th, 2023

help-circle





  • I can’t speak for other fields, but I’ve worked in IT as a sysadmin for about a decade at a bunch of different companies, big and small.

    I’ve never worked at a place that was close to “overstaffed” nearly every place I’ve worked we’ve needed at least 2-4 additional people.

    Everybody was overworked, overwhelmed with tickets and projects, working 50+ hours a week constantly.

    But upper management and executives love claiming that staffing is maxed out and needs to get more lean. Like, dude, our IT team is handling dozens of tickets a day, running 5-10 different infrastructure projects simultaneously, and keeping near-decade old equipment alive because we were denied our third budget request in a row.


  • I personally think that general consumers will never use LLMs in any significant number. I think that LLMs will exist in two distinct spaces, FOSS for devs and other technical people who want to run there own infra locally - and B2B for everything else.

    The few big AI companies that manage to last will be selling access to their models for much higher prices. Probably similar to current proprietary commercial software like VMWare, SolidWorks, VEEAM, Splunk, etc. Companies will pay hundreds, possibly thousands of dollars per seat depending on the niche offering and amount of usage.

    Suppose that a company developed an LLM that is trained & tuned specifically to do legal work, and suppose it produced work that was around 95% the quality of a typical paralegal. If that company charged $6,000 a year per license to work on their platform, that’s expensive, but if you’re a small firm with say, a dozen full time lawyers, then for the yearly price of a single average paralegal, you could have each lawyer using that software to do most of the work that the paralegal would have done. I can see those kinds of applications happening more and more.

    This assumes though that LLMs will continue to improve at a significant rate for a long time into the future, (5-10 more years) which isn’t at all obvious, and there is some evidence that it’s already starting to hit a ceiling.

    There are other ways it might work, like if there is a method of compression that is discovered that reduces the necessary RAM and Compute needs by 2-3 orders of magnitude. So models that are considered very large today (100-300 billion params at full quality) might be able to run effectively on a single 32GB GPU that costs a few thousand dollars.

    So the cost to run these models is reduced immensely, and a single small data center could run enormous models with 1,000,000+ context windows for tens of thousands of users at once.

    But that cuts both ways, which is something that any AI company is going to have to deal with. Once small free models get good enough to do the vast majority of a task, a user is going to start weighing the cost/benefits, and the prospect of just buying a box and throwing one of these models in for a few grand will be very appealing.

    I think there may be a good market out there for “AI boxes”, compact computers designed to run a tuned LLM, set up with a little special sauce so the interface is user-friendly, etc. Companies could sell these with support contracts to legal firms, indie Dev studios, startups, small government agencies, etc.

    Idk, it’s so up in the air right now, and everything is constantly changing so fast. It’s impossible to predict where things will be in 6 months, let alone 6 years from now.




  • As opposed to now, where companies copyright and patent their medications and sell them ultimately to…taxpayers who pay them billions of dollars a year in just out of pocket costs, let alone the scheme that is the American private insurance/healthcare system.

    If taxpayers had to fund drug companies or research institutions for R&D without the insane middleman that is the private healthcare/insurance system, it would cost a fraction of what it does now.

    On top of that, this assumes that people won’t do research for the good of society vs becoming filthy rich, which is a false assumption driven by Capitalist propaganda. Remember that the ultra wealthy CEOs and executives of these drug companies aren’t the ones doing any of the actual work or research. That is all done by scientists and engineers, who make a decent living, but none of them are incredibly rich from it, classic Capitalist exploitation at work.

    Often times these drug companies (and the private equity firms that own them) don’t even primarily do R&D, they just purchase the patents and IP rights to drugs that are already on the market, and once they do that, they jack up the price often by hundreds of percent to increase their cashflow.

    That cost gets sent to the insurance companies, which of course, they pass on to consumers, raising our healthcare prices year after year.

    I want medical researchers, scientists, and engineers to make a good living, a very good living, their work literally saves and improves hundreds of millions of people’s lives worldwide. But you don’t need a CEO or executives, or private equity firms owning that space and making insane amounts of money, you just literally don’t.

    There are millions of very smart and passionate people around the world who want to do this kind of work because they enjoy it, and they want to make a difference. Providing an open and rigorous academic and scientific structure to study and practice this is all you need. That already exists today, many medical breakthroughs came from publicially funded research institutes, which is the way it should be.







  • Gaming PC - Nobara (Fedora base with lots of gaming-specifc kernel optimizations baked in.)

    Personal laptop - Linux Mint

    Business laptop - Linux Mint Debian Edition

    Junk/Test laptops - Void

    Home lab main hypervisor - XCP-ng (Highly customized Fedora under the hood.)

    NAS - TrueNAS (Debian under the hood.)

    Virtual servers - Mostly Debian, but a few Alma Linux VMs to get that RHEL experience. Ubuntu Server for my self-hosted gaming servers.

    Steam Deck - SteamOS (Valve’s immutable spin of Arch.)


    1. Typically, but not always. Some trans women are biologically intersex. (This also depends on how you define “biologically male” which is not totally straightforward.)
    2. It matters in some contexts, not in others. Their physician should know, because various hormone treatments cause different effects in people’s bodies, and certain health conditions effect biologically male or female people differently too. That’s nobody else’s business but the patient and their trusted medical providers. As far as their dignity, opportunities, and general acceptance, it doesn’t matter. Trans folks deserve the exact same rights, opportunities, and acceptance as anybody else.
    3. Usually people who bring this up aren’t acting in good faith, so I don’t engage with them. On the rare occasion where somebody is genuinely curious and wants to learn, I answer them in the same way as I am doing right now.
    4. Because the word “woman” denotes multiple concepts, like the word “parent”. If a child is adopted at birth and is raised by a couple, the child and their community will refer to those people as the child’s parents. This is not a false statement, because the word “parent” doesn’t only mean the direct biological progenitors of a person. Parent also is a social role, hence the verb form “to parent somebody.” This is also why we have the terms, “biological parent” and “adoptive parent” to add additional information when it’s necessary.

    Trans women are women in the sense that they are filling their society’s sociological role that surrounds the expected concept of a woman. That will be different depending on many factors, and will have many different aspects including their pronouns, fashion and clothing, voice, makeup, hair, activities, and so forth.

    Just like any other woman, they will chose which social roles they desire to fit into, and which ones they don’t, and all of that is completely acceptable.


  • Wish they handled it better, but I knew about this a while ago, and the price is more than reasonable.

    A decade without a price hike is extremely generous, especially at how cheap their plan was.

    They are a FOSS company that makes a fantastic product I’ve been happy with for years, I’ll gladly pay less than $2 a month to support them. Their server code is licensed with the AGPL, the strongest copyleft license there is, which gives me a lot of confidence.

    Worse case scenario, they enshitify down the road, we are protected via the open source implementations. We’ve seen this many times in the past, Red Hat > Alma & Rocky Linux, Citrix Xen Server > XCP-ng, Terraform > Open Tofu.

    Pay for your open source software, folks 💖