Hacker Newsnew | past | comments | ask | show | jobs | submit | more ericb's commentslogin

runsc / gVisor is interesting also as the runsc engine can be run from within Docker/Docker Desktop.

gVisor has performance problems, though. Their data shows 1/3rd the throughput vs. docker runtime for concurrent network calls--if that's an issue for your use-case.


In ollama, how do you set up the larger context, and figure out what settings to use? I've yet to find a good guide. I'm also not quite sure how I should figure out what those settings should be for each model.

There's context length, but then, how does that relate to input length and output length? Should I just make the numbers match? 32k is 32k? Any pointers?


For aider and ollama, see: https://aider.chat/docs/llms/ollama.html

Just for ollama, see: https://github.com/ollama/ollama/blob/main/docs/faq.md#how-c...

I’m using llama.cpp though, so I can’t confirm these methods.


Are you using it with aider? If so, how has your experience been?


Ollama breaks for me. If I manually set the context higher. The next api call from clone resets it back.

And ollama keeps taking it out of memory every 4 minutes.

LM studio with MLX on Mac is performing perfectly and I can keep it in my ram indefinitely.

Ollama keep alive is broken as a new rest api call resets it after. I’m surprised it’s this glitched with longer running calls and custom context length.


I'm a fan of TrilliumNext, which is open source, for this:

https://github.com/TriliumNext/Notes


Thanks for sharing. Their encryption service is a nice source of inspiration on note encryption https://github.com/TriliumNext/Notes/blob/56d4d7c20f775eed73...


Same, this gets my recommendation. Trilium is one of the most under-rated tool I know of.


I would add an LLM like QwQ-32B to the mix--that has a ton of compressed knowledge embedded in it.

I would also store it in a steel Oscar the Grouch style trash can for a cheap faraday cage, which gets you protection from solar flares, and EMP blasts.


LLMs are a bad deal when you look at how much power you need to run that inference. A device that could barely run one instance of QwQ-32B at glacial speeds will be able to serve multiple concurrent users of Kiwix.


Quick question: which car companies are working on self driving cars? All of them, and two other companies ( Apple and Google ).

Which militaries are working on battle field AI. All of them.

Could a 64Gb dual xeon run say 50 to 100 users of kiwix?


To serve multiple users, probably not.

But--if you don't think of asking Hacker News every single thing you need to know beforehand, I think you still want the LLM to answer questions and help you bootstrap it.


Learning things from scratch is really hard too, just a copy of wikipedia gets one absolutely nowhere if you don't know what to search for.

Having something that you can plainly ask how to start that will point you in the right direction and explain the base concepts is worth a lot more, it turns raw data into genuine information. Yes it can be wrong sometimes, but so can human teachers and you can always verify, which is a good skill to practice in general.


See the Wired article on the rewright of German history. And The George Galloway article. The enshitification has not only begun, it's in rising force.


> educating people on what the best choice is and letting them make it themselves.

If we go that way, we should have some sort of Department, to make sure Of the Education being consistent.


Very cool! Just digging in. Does it works with the new JSON format clickhouse introduced recently?

Also, what service did you use to make the video, if you don't mind my asking?


Thanks!

I haven't tested the new JSON format in ClickHouse yet, but even if something doesn't work at the moment, fixing it should be trivial.

As for the video service, it wasn’t actually a service but rather a set of local tools:

- Video capture/screenshots - macOS default tools

- Screenshot editing - GIMP

- Voice generation - https://elevenlabs.io/app/speech-synthesis/text-to-speech

- Montage – DaVinci Resolve 19


Very cool concept! There's a lot of potential in reducing the try-debug--fix cycle for LLMs.

On a related note, here's a Ruby gem I wrote that captures variable state from the moment an Exception is raised. It gets you non-interactive text-based debugging for exceptions.

https://rubygems.org/gems/enhanced_errors


I like it too. Memorable is good! Why not just put Hydraulic in front of the name of each other product? Hydraulic Deploy. Hydraulic Build. Etc. Seems scalable.


That's a really fascinating, but horrible, point.

Every AI decision becomes a way to shirk responsibility, even more than just automated ones (because then its "your" rule).

Welcome to the brave new world of AI-decision laundering.


The same can be said of any bureaucracy's function. It isn't your fault that you made an abhorrently stupid decision, you were just following the directive. Not to say it isn't a problem, but that it isn't new.


That makes it a great point, not a horrible one.


I think they mean its reality and implications are horrible.


Yes, thank-you.


You mistook my meaning.

Great thought from op.

Horrible implications.


> Over the last 20 years, 100-200 such cables globally have been damaged annually.

[Citation Needed]



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: