friendly_chap

I have bad news for you: for the majority of the population mental effort was always a foreign concept. That's exactly what got us in this mess in the first place.

Me: Paste a go compile error

Gemini: Wow, yeah, haha! That's the final boss of Go compilation errors!

The problem is they nerfed 5.5 about a month ago. The change was immediately visible to me: context compacting started to happen about 3x as frequently.

I think 5.6 is still not even close to 5.5 xhigh pre-nerf.

They mostly use public repositories though.

Absolutely, think "lick" vs "leak". I think the author means Hungarian maybe uses very similar looking letters to denote this (ie "lik" and "lík").

In Hungarian also every vowel comes in pairs of short-long: a-á (what vs high), e-é (ever vs eight), "o-ó" (moss vs most), "u-ú" (put vs you), "ö-ő" (fur vs ... well long version has no English equivalent I think but German does: schön).

I can imagine that happening when the motivation is external (praise). When I write open-source it's because I have to write it out of myself.

Do I expect praise, kudos, fame, whatever? I do and that happens, I have been hired countless of times because of open-source. Even my friends (!) have been hired because of open-source stuff I wrote and they contributed to.

But the main motivation is internal - I just have to see it take shape. Like a writer who can't resist painting or a writer who can't stop writing.

Do I have regrets still? Yeah because I could have used the time for better things. But that can be said about any hobby.

Repo: https://github.com/1backend/1backend

After some 12+ years of collecting microservices platform ideas in my head and implementing them at various companies and open-source projects, I decided to create a system that contains all of them.

1Backend is the result. Mostly built it for myself so I have a foundation to build projects on but I'd love if others would also use it!

I think that makes sense, AI could be used to evaluate and score the importance of contributions imho. In fact this is probably a business idea in itself...

I think the idea is fairly cool, although it's hard to come up with a way to judge contributions. How would you solve the legal part? In the end a business entity would still need to own things.

Would you perhaps sync ownership from time to time from the DAO to the real world?

I'm working on a few closed source projects that are all built on https://github.com/1backend/1backend, my dream microservices platform I've had in mind for 12 years before starting to implement it.

On the frontend I'm rather happy with Angular but backend-wise I was never really happy. So it was time to build it. Been developing it for more than a year now.

> If you want to be fair, you could add that information. Otherwise, readers will assume you mean that we nefariously censor negative posts about YC, when the truth is the opposite.

I'll do that, sure! Dang, I have deep respect for both you and HN for many-many reasons (HN literally changed my life and made it 1000x). So don't take this as me having a grudge. I just like accountability.

> but it does sound like you noticed something that should not have been there.

I don't follow this, sorry.

If anyone is reading from YC, I'm happy to review their submissions, I was perhaps the first who raised concerns and was on top of the thread with my comment.

(See this post I made on LinkedIn about this: https://www.linkedin.com/posts/crufter_today-y-combinator-de...)

That post was pulled partly because of my comment. I commented this:

"While I see the economical usefullness, this sounds like the worst possible application of AI.

Using AI to surveil is building hell on earth. AI should be used to help people work less/easier, not whip them into working more."

Which ended up on the top of the thread. Was surprised to wake up this morning and see it gone.

LinkedIn post I made about this:

https://www.linkedin.com/posts/crufter_today-y-combinator-de...

Hey! Email is not public on HN, if you want to publish it you gotta include it in your about section :). Also how are you with German? I used to speak some but forgot most. Can you get by with English?

Why is it miserable? I left in 2020. Life was pretty good there at the time, I just got bored of London. What changed?

I'm thinking about moving from Hungary (Eastern Europe). I used to live in London in my twenties but came home (what is home? I miss the UK) to settle down. But the leading right wing party is poisoning this country for Chinese money so I'm forced to move.

I think with all its problems US is still the best country in the world, but moving there is not an option for me. Even with a H-1B visa or whatever that's called, I don't like the idea of being tied that tightly to my employer. The US is still the source of innovation (imho for cultural reasons). I feel like my spirit is closest to the American one. Love for freedom, self-sufficiency etc.

Since the USA is not an option I'll settle for one of the European countries. Candidates are Spain, Switzerland or Austria.

Reasons are: Spain has cheap properties and amazing nature, Switzerland has an extremely high quality of life and wealth, and Austria is close to my home country.

I feel personally attacked ; ) I'm also building a tool to run/build on top of LLMs (https://github.com/singulatron/superplatform) and I opted for containers too. TBF I'm mostly targeting backend developers (who am I kidding, I'm mostly building this for myself).

The desktop version has its own configuration management software to install docker or WSL and all the dependencies you talk about, so I feel your pain.

/self plug

Hah! Nice idea! I built something with a similar mindset but instead of calling cloud AI providers my aim is to provide a self-hostable complete AI platform: https://github.com/singulatron/singulatron

I know that might sound like putting the server back to serverless. But I would say it's being your own serverless provider - once you have the platform installed on your servers you can build frontend-only AI apps on top.

Hope you don't mind the self-plug. Your approach definitely a ton of advantages when starting out (no infra to manage etc).

> Sorry if this sounds harsh

I braced myself for actual harsh critique there but that was rather mild! Phew.

> Or if the entire value is based on something that's easily replicated

Not THAT easily replicated but given how well funded our competitors are and we are bootstrapped I would prefer to wait until most startups participating in this wave burn through their funds instead of publishing relatively novel but not overly expensive to replicate twists.

We'll come up with better explanations of how the thing works without hinting at technical difficulties we faced and their solutions.

The rightful point was made, thank you. You were not harsh at all.

Hey! At the moment we don't release much about the inner workings but that stance might change soon as we have a habit of going from closed -> open pretty quickly ;).

Shameless self-plug but your questions (especially combined with the product posted) are the raison d'etre for our whole company: https://singulatron.com/

We are running smaller models with software we wrote (self plug alert: https://github.com/singulatron/singulatron) with great success. There are obvious mistakes these models make (such as the one in our repo image - haha) sometimes but they can also be surprisingly versatile in areas you don't expect them to be, like coding.

Our demo site uses two NVIDIA GeForce RTX 3090 and our whole team is hammering it all day. The only problem is occasionally high GPU temperature.

I don't think the picture is as bleak as you paint. I actually expect Moore's Law and better AI architectures to bring on a self-hosted AI revolution in the next few years.

The main focus is the daemon which we plan to make distributed. We will also support multiple database backends and other distributed primitives.

Thanks gpaslari,

We are running known models like LLama, StableDiffusion etc. So pretty much the same :)

We are not planning on building our own models yet.

So any web app that runs on multiple nodes reinvent k8s? I'm not sure about that.

Hey!

At the moment the distributed features are entirely missing from the repo, sorry to disappoint on that front despite being part of the readme. It's hectic!

But to answer your question: I plan to sidestep most difficulties of the distributed story by offloading into the usual distributed tools - horizontally scalable DBs and distributed locks, queues etc. This will have the added upside of being able to use your favourite tools for which you already have dashboards, alerting etc.

There will be no fancy resource management (an other comment touched in "reinventing kubernetes" which is definitely not a goal at all), but there will be some simple optimizations in terms of where to run which prompt depending on if the model is already running on a node or not etc.

So there will be some knowledge of the nodes' capacity (at least available VRAM etc), but the picking of the prompts from the queue will be a simple unit testable algorithm.

At least this is the plan but it's only half implemented in my mind.

I'm not planning to :)). Singulatron is not about scheduling, etc.; it operates at a much higher level of abstraction. To put it in less fancy terms: it's an app, albeit one designed to run on many servers simultaneously.

Hey folks,

Op here. Trying to build a distributed AI platform - lots of work, fairly early phase but I think it's already useful. Hope you'll like it.

Thanks!

I have actually open sourced one recently which supports queues - its both a desktop app and a daemon: https://github.com/singulatron/singulatron

We are working on improving getting the idea across. It's just a search engine with a bunch of extra filtering capabilities - at the moment those filters are all related to the website owner companies' metadata (industry, size etc).