Its just gambling, for the stupid to loose there money, for the rich to have a little bit of fun, for the smart ones to take it from others while having some way of moral excuse to justify it.
The stock market at least gives you ownership/partial of a real thing.
The depressing thing is, when you see all these cliches in real: Go to some casino and you will see the guy having a coke nail talking gambling garbage like 'were is the bank teller? My luck is coming back' :(
I feel like discussiong ai with people who hated the seat belt, enjoyed smoking inside, giving babies alcohol and doing operations on babies because they couldn't feel pain.
So many more people have clear stances against AI but the energy consumption of bitcoin was seldomly a hotly debated topic.
Lets be clear, technology advances. You can embrace it and learn it and use it, or you lose.
For me most shocking thing is: I wouldn't have expected so many people getting AI wrong or not getting it at all. 20 Years of software engineering and i have never experienced a technology so weird and crazy and with such a fast progress than AI/ML/Neural Networks.
Is it perfect? No. But if you would have asked ANYONE just a few years back "Hey i give you a 100 Million dollars, you will write some software however you like and I expect it to be able to solve any problem (an easy one for the start) but i will not tell you what it will be" we would have build something like IBM Watson and it would have just failed.
but we have a breakthrough in neural networks progress. I can ask it in german, in english, in shitty english and shitty german. I can generate any picture i want, i can generate videos.
This technology bridges computer and human. Its the FIRST TIME EVER I could build something like a Star Trek Computer.
Just think this 5 or 20 years further and its clear that these models will be better, relevant better than today. You will (if still needed) be able to stand in a VR world and talk to a computer and it will understand you well enough and it will be able to generate what ever you want.
Of course we still need to solve a lot more issues but honestly, before ChatGPT it felt like we solved software but it was not clear how it will continue. There was something missing.
And the way we build out compute, can be a milestone for material science, pharma, biochemistry etc.
Also were not building out compute in any meaningful way, theres only actually been a couple of fundamental advancements to how we understand NN on a fundamental level, whats really changed is theres now a few companies with enough resources to throw an ungodly amount of compute power at a clearly already plateauing line of tech. To the detriment to every other opportunity.
Its not plateuing at all. And the next things are already in the working like world models.
All this compute can also be used for other neural networks like finding new materials or drugs etc.
But what i meant was, that the outcry against AI is so much louder than bitcoin. The argument was brought up for sure but it was ignored plenty. No one forbid a company to restart a gas power plant in NYC to JUST burn it for bitcoin.
Honestly why I would really apprechiate something like this, hn is not an explain platform.
For sure, some words or feedback on what you understood (did you get it right) etc. yeah.
But otherwise, if you do not understand a research paper, you have to do the same hard work as everyone else. Sitting down, going through it paragraph by paragraph and learning it. This takes massive time.
and for a high level overview, chatgpt and co are really really good getting papers.
If you need help getting more out of ai, you can use chatgpt and co to go through papers and let yourself eli5 paragarphs. 1blue3brown also has a few great videos about transformer and how they work
I found a few papers in this direction with perplexity like this one https://ceur-ws.org/Vol-4005/paper1.pdf and it doesn't seem to be that relevant for now.
The progress of a handful models seem to be so much better (because limited compute, we have only a handful of big ones, i presume) that these finetunings are just not yet relevant.
I'm also curious if a english java + html + css + javascript only model would look like in size and speed for example.
Unfortunate whenever i ask myself the question of finetunging tokens (just a few days ago this question came up again), deep diving takes too much time.
Claude only got lsp support in november i think. And its not even clear to me to what extend. So despite the feeling we are moving fast, tons of basic ideas haven't even made it in yet
if you have a corpus of code snippets to train the manifold (Laplacian) on (and a good embedding model), it is definitely possible to try something like this.
The stock market at least gives you ownership/partial of a real thing.
The depressing thing is, when you see all these cliches in real: Go to some casino and you will see the guy having a coke nail talking gambling garbage like 'were is the bank teller? My luck is coming back' :(