urvader

Maybe we should create an RFC for it?

> Mirabilis / AOL really needs to fix this problem.

Berget Code is running normal https://berget.ai/code

What about cache? When you change the context the prefill stage will be much slower?

About the caching things. This is a flaw in the backend, not in the Frontend. If the backend is only relying on prefix cache you need to look for more fine grained cache solutions like HiCache or Lmcache. If your backend discards the whole cache because of one date is changed, you really can’t blame it on the Frontend. It is an implementation detail that you push to the wrong side.

Great post! Today we just launched an European alternative to Claude Code - Berget Code- https://berget.ai/code

GTA VI is postponed again, release next year

I'm not sure this is at all a good idea but thanks to the great discussion here at Hacker News - it is now up on npm:

npm i aspipes

I wrote this with an LLM but manually changed the README. Thanks for pointing this out, it is now updated.

Thanks for pointing this out, I updated the examples now to this syntax.

The bitcoin block chain should be on this list

This model will be one of the few open models where the training data is also open which makes it ideal for fine tuning.

Thanks!

Well we have already made the source code open and free and also encouraged the city to release an app with our source code as base. They weren’t interested in that. They would rather license the app, support and maintenance to us. We have quoted a fixed sum per month for that service and we plan to use that money to reimburse everyone sending PR:s we merge.

The law is not specific at all in regards to the format of the document. So to talk about an “API legally” has no meaning. In a private scenario it makes sense but what we are talking about here is public documents which are sent through an API. The city has responsibility to only send information I have (as a parent) legally right to see. How I parse it and present it is up to me as citizen (through an app or save it as json and upload to an excel file or such)

One implication of this project could be that government agencies in Sweden can not have private API:s.

To use more proprietary methods (private api:s) will have no effect on the constitutional law. You still have received a public document as a citizen.

If you encrypt it, you have to, at some point, also send the keys to the user. The key has the same legal protection as the rest of the document so encrypting the data has no implication on the legal discussion.

No, the app has no communication to us, we don’t even have a server. This means that from a legal standpoint we aren’t publishing any information. We only help our users to present their own data in a better format (than json).

(I am Christian Landgren, cofounder of the project)

You are right, the city believes they have ownership of the data, mainly because they fail to understand that they aren’t showing data in an app, but rather publishing data in an API. In Swedish law, once you have released data from a government, the receiver have the right to do whatever they want with the that data (as long as it isn’t violating any other laws).

The city in this case is responsible to check that the data is safe to share publicly and once they have- the data is not theirs. This is regulated in the constitutional law regulating free speech which goes back to year 1766.

This means that they can’t really apply the same logic as a private company can when publishing data in their api. A private company can still keep license over what can be done with the data they publish. A city can not do that because of these constitutional laws.

I am one of the programming parents, please help us get the word out to Wired, TechCrunch etc this ridiculous request can not be accepted. We have to put pressure on politicians to stop this now.

Fork our repos: https://GitHub.org/kolplattformen/

Connect me to journalists: https://twitter.com/landgren