As a side point, some people here seem to think this post specifically came from 24 contributors. The text at the bottom seems to indicate this and I initially got the same feeling.
However, that's actually a description of the site itself, not the post. There are 24 essays, one per contributor.
This is definitely something we're considering. The initial use cases we saw were people using Skald to build AI features into their own apps based on content that was generated within their application, but we're now seeing more interest from companies wanting to import data that exists somewhere else into Skald, so this certainly would make sense to build!
It's easy to have strong feelings about this post because, well, we all have to pay bills and often times aren't the biggest fans of the means (work) we have to do that. My immediate reaction, like many here, is to just go "well tough luck kid".
But I'd like to offer some sympathy. I certainly have grappled with thoughts like these and have also been guilty of posting a rant on HN at a moment when I've been down!
I do wonder if part of this is influenced by the AI craze that has companies substituting junior engineers for LLMs and how hard it is to get hired fresh out of university these days. I do feel for those who genuinely want to grow and become better engineers since it does seem like companies are betting less and less on developing young talent.
Then there's the whole philosophical discussion about work and meaning and everything. Thoughts around this are certainly very present in our minds during our 20s (P.S. I'm still in this decade of my life too). There are many alternative paths, but they often aren't for everyone. I know people who live with very little, and don't consider steady work a high priority at all. Many of them are happy, but most of us couldn't cope with the lifestyle. You then have the path of starting your own thing, but that path is usually more painful and terrible for your finances too.
It's all tradeoffs. It sucks, it hurts. And I'm sorry that the market is terrible right now for those starting out. Good luck.
I think Futureme "letters" are actually emails, no? Or at least they do both. I've sent my future self emails via Futureme (and they called it a "letter" on the email). Nevertheless, this is still cool, congrats OP!
I highly recommend people try this exercise. I got an email this year which I sent myself five years ago. It was tiny, and still mind-blowing.
Ah also note that the one thing that would put me off from using this service is actually trusting that you'll (you as in Resurf) still be around 10 years from now to deliver the email.
Yeah great point...
It's kind of a catch-22, in that resurf has to be around for a long time, for people to think it will be around for a longer time still, haha.
Right now though, I set the maximum date in the future as one year from now, so should be able to guarantee that at least.
Similar thing happened to me, albeit like 6 or so years ago.
I was very young and had been "in the space" for a couple years. The concept of "decentralization" in general appealed to me as someone frustrated with the abuses of centralized power. I liked the idea of a transparent ledger and wanted to apply it to non-profit initiatives. I even spoke about this at the European Central Bank as a 20 year old (got there via anonymous paper submission -- imagine the surprise).
But very quickly I grew frustrated with how much of the space was scams, people looking to get rich quick, etc. I had been more interested in the applications of blockchain that went beyond crypto but even stopped paying attention to cryptos altogether as a speculative investment. Funny enough, a few people who were "evangelized" by me made a fair bit of money.
I'm thankful though because I got into programming at 18 to write smart contracts, and that actually changed the course of my life quite drastically.
The only useful application of crypto/blockchain tech I've ever seen simply takes advantage of the transparent ledger as a way of tracking receipts.
One of the perfect applications is in shipping and logistics where receipt validation, transparency, and certification, is an essential part of the business.
Is there somewhere that explains this concept? When I "receive" something, there still must be trust that I correctly mark it so. By that point, doesn't it require trust, and therefore you may as well just use a trusted third party data store?
In this case do you even need a RAG? Most models will have been trained on Wikipedia anyway.
Give Jan (https://www.jan.ai/) a try for instance. You'll need to do a bit of research as to what model will give you the best perf on your system but one of the quantized Llama or Qwen models will probably suit you well.
reply