tailscale offers many good things.

for example "tailscale" drop for free, share files between devices on your network

also https://tailscale.com/docs/features/tailscale-funnel which is cloudflare tunnel alternative, no more localtunnel or other things

so i am not surprized people like it as product, most good features are free.

Had very high hopes for tailscale drop but it’s worked literally once of the 7 times I tried to use it before I stopped trying.

Mostly machines on the same network.

yesterday i setup tailscale on a gcp VM.

i enabled route advertising and managed to ping my google cloud instances using tailscale

problem is the packet loss rate is high. SSH tunnel worked!

soo i think there is something wrong with tailscale

I thought they acquire the pirate bay.

Same can be said for java, yet google own android.

This time it's different.

Linux was already stable enough 10 years ago as daily driver, i used Arch.

everything worked just fine, i remember only having issue with graphic drivers and glitches

I never really wanted anything more from it but when i moved to Mac, i saw how it prevents me from opening apps i downloaded from trusted site and every now and then i need to set xattr to open the files, and go through bunch of lockdowns.

Now freecad has improved so much, with all AI coding and all opensource will improve DRASTICALLY and very fast.

using AI which stole everyone's code to develop OpenSource is morally right thing to do vs using it at private companies. It will attract more devs.

I have tried Arch btw

I work in india, most village panchayats are running Ubuntu without any issues and we don't pay anyone any money.

Third world country. We, in first world country, are not dependendent of Ubuntu. We are supported by Microsoft. (see Limux). /s

I consulted at some Indian startups. Such bad queries were written which were costing them so much compute

I just rewrote all queries with claude code and next day and compute decreased to 1/5th.

Kind of a weird flex - you're basically saying that you aren't any better than the Indian teams at writing efficient queries or optimizing existing ones, the only talent you've demonstrated is ability to ask Claude Code. It's like you are announcing your own obsolescence.

It's true, claude is already outperforming me

It's funny such a big corporations can't let such a small tool live.

Google is such an evil company, it is not even provided anything great anymore.

Anti-gravity paid plans suck, GCP is billing heavy. Today google sucks at most things

Their Android playstore hardly updates statistics once a day, so much for such a big data company with unlimited sources lol

It's a tool used to build other tools, some of which have non-trivial amounts of users.

It's also a tool used by e.g. journalists and government agencies that dabble in stuff like research and evidence, and it would probably be more cumbersome for everyone involved if Google instead had to process requests and provide copies of material for these purposes.

Otherwise they'd probably have made life much harder for the yt-dlp-developers already. Not that I think they're nice in any way, but I don't think they're seriously trying to fully eradicate yt-dlp or related software.

That people on this forum convinced themselves that it's a reasonable take to vibe code a useful geometric kernel is profoundly depressing.

Indeed. I would love for it to be true, but aside from opencascade^1 all the professional kernels are proprietary and not in the training set, so LLMs can't just regurgitate them.

^1: Which I really appreciate, but let's be real, it is far behind eg. parasolid.

Correct me if I'm wrong, kernel is already there, and it's decent enough. What's lacking is a good user interface to it.

FreeCAD is perfectly good user interface for opencascade. The problem is that as your geometry gets more complicated you start running into the kernel limitations.

Just wait for sometime, there will be one.

There are many geniuses on HN for tacking up challenge

Vibe coding only seems to work, insofar as it does when the training data includes multiple exemplars of solutions to a given problem.

As noted elsethread, there's only one geometric kernel which is decently far along and opensource and it's over 1 million LOC --- I doubt it's being included in any training data, and I doubt that an LLM could regurgitate such a large project which would then compile w/o errors and then work as expected --- the number of tokens required to get such a project to an initial state is a marked hurdle as well.

So you think AI can do this – why exactly is a genius needed?

genius is needed in pilot seat to guide AI :)