diff options
Diffstat (limited to 'content/blog/2024/llms-scare-me.md')
| -rw-r--r-- | content/blog/2024/llms-scare-me.md | 110 |
1 files changed, 110 insertions, 0 deletions
diff --git a/content/blog/2024/llms-scare-me.md b/content/blog/2024/llms-scare-me.md new file mode 100644 index 0000000..32637c6 --- /dev/null +++ b/content/blog/2024/llms-scare-me.md @@ -0,0 +1,110 @@ ++++ +title = "LLMs are everything that it wrong in the world of computing" +author = ["Michał Sapka"] +date = 2024-12-23T21:53:00+01:00 +categories = ["blog"] +draft = false +weight = 2001 +image_dir = "blog/images" +image_max_width = 600 +Abstract = "On why I am loosing hope in software" +Listening = "For help me - Calm Nights" +Listening_Url = "https://www.youtube.com/watch?v=7A2YmgCOImc" ++++ + +For decades corporations have been doing anything in their power to make computers _worse_. +Software used to much faster, much leaner than it is now. +Hardware performance is progresing by leaps and bounds, but somehow the software is slower. +Yes, programs we run (cough, web applications, cough) are more complex, but the complexity increase is smaller than the compute improvements. + +Software used to also be substantially better written. +When sending patches to clients was a risk by itself, releasing barely working software was akin to shooting yourself in the foot. +Now bug are expected. +Again - I am comparing simpler software to what we have now, but the complexity increase is smaller than increase in budgets and team sizes. + +Software also used to be: + +- much more open +- purchasable +- not requiring internet +- not as streamlined +- not designed by psychologists to be addictive +- not designed to be a constant stream of revenue +- swiftly improving +- possible to be created by teens in their bedrooms +- simple (as in having a limited use case) +- small +- running on different hardware architectures (yes, there used to be more than 3 systems) + +I am often accused of romanticizing past, but I am far from that. +Software of the golden age had lots of problems, it was far from perfect. +I am critical of the path we took - as it was aimed not at _better software_, but at _better return of investment_. +We could have had much better working computers, but here we are. +Everything is barely working but dollar pours. +Some call it enshittification, others distribution. +All I know is this is not going to stop. +It is never enough. + +This brings me to ChatGPT, LLMs and all that crap. +I have to admit, that the way it works is extremely cool. +Some vector math is able to fool us into thinking we are interacting with a human being? +How cool! +I'd love bo be amazed by it, but if you we look at it critically, it a is terrible peace of software. +It is slow, expensive, and non-deterministic. +If I released a code to production with has _some chance_ to work or not, I would need to fix it. +If the same software yielded different returns _by design_, I'd need to rewrite it. + +I am fully convinced that LLMs are not a path forward at best, and a huge step backward at worst. +A lot of people are looking at this tech with scared eyes, but there's no chance LLM will be better than they are. + +However, as in the above examples, better doesn't always win. +I am afraid that too much money was invested in this dead-end technology for it to fall. +The entire computing world is the hands of a few gigantic companies, and each of them has fully risked their future on it. +Google, Microsoft, Amazon, NVIDIA, AMD, Apple - by the end of 2024 there is no big computer company left that is not "AI first". +Will they let their investments go to waste? + +Google search is useless; Windows is a clown show; Facebook/Twitter are not be touched by humans. +Yet they own _billions_ of dollars. +They are able to throw piles of dollars in the data center furnace and wait. +The fact that they also benefit from those dollars doesn't make it any harder. + +I know people who already can't work without asking ChatGPT and they lived through better times. +But I also know teens who never had a fast computer; for whom slow a web app is the only status-quo they have experienced. +For them the current crap software is simply _software_. +I am really afraid that the next generations may never know deterministic software. +Companies can simply continue to add "AI" features (since AI is now a synonym for GenAI), and raise the kids with it. +Will the kids care that it sucks? +Will they even know? + +Will they give a rat's ass that the GenAI has only two goals: first, to become abstraction over all interactions with computer; second - to make people obsolete? +I think the first one is the goal, while the other is what GenAI dealers sell to investors. +Microsoft and Google want to be between everything we do with out computers; that was their strategy for a long time. +It's a self-propelling nightmare: the worse the vision, the bigger is the money river, which fuels making computers unusable. +We have already lost so many alternatives - Google is trying to fix itself with GenAI after it destroyed itself with ads, after it destroyed all competition. +The web was already great. +Microsoft Windows ate all other x86 OSes except of Unix - and we had dozens of great ones. +Apple iOS or Google Android are mandatory - but this was not the case a few short years ago. + +Progress is not linear and _better_ doesn't always win. +In most cases whatever gives more return to the investors does. +And no other technology has more investor risk attached to it than LLMs. +AI has a lot of usecases, but it is thrown away to make way for GenAI. +This steamroller may break computers, the economy, and the environment. +But there is nothing we can about it, as none of us is a billionaire. +All we can do is to dance how it pleases them. + +Unless something **big** happens. +With infinite pockets, the giants can just [burn the money](https://indiadispatch.com/2024/12/22/big-techs-292-billion-ai-spending-spree-meets-the-revenue-desert/) - literally. + +--- + +I will keep this site althole-free, and I will personally run software that has zero LLM. +I don't use copilot, I had only a dozen chats with GPT. +I will try to continue keeping LLM subjects here to a minimum. +This is my life and this is where I have any control. +This is the only rebellion I am capable of. + +--- + +In unrelated news - SystemD now tries to convince people that having file-based logs is deprecated. +And in few short years we will have Linux "administrators" who don't know how to use grep(1). |