tiluha

My current vibe/playing around setup is deepseek v4.1 flash with opus/sol as advisor agent, dsv4.1flash agents for self review and final review with opus/sol. I love how fast and cheap deepseek is. Using omp as harness.

Privacy policy is not great with deepseek api, but you are always just trusting their word with any hosted llm and in theory i could at least self host the models i use from deepseek.

Ladybird is the browser this whole submission is about. Did you not even read the headline?

It really is crazy. Placing my phone on my chest lying down i can clearly see my heartbeat on the graph

Makes no difference in the fundamental dislike i have for the concept

Mine do, at least when it's very dark

Im pretty sure this is not the case on most providers, where "dedicated" VPSs demand a significant premium over the default "shared" VPSs

This is something i have been thinking about lately. How well do these performance optimizations work in the cloud on a shared system?

You are not alone. It evoked a physical reaction in me

This does not match my experience in Germany. If somebody gives you their phone number it is just expected that you can reach them on WhatsApp and i have yet to meet anyone that doesn't use WhatsApp.

You need 25k€ to register a GmbH, but you can nowadays register a UG which only requires 1€ and can later be converted to a GmbH

One time they didn't get the needle in smoothly and it hurt a little bit, but it really wasn't too bad

In Germany you get to eat from a tasty assortment of food after donating. When going with friends it makes for a nice little event. Fair deal

This is true, but scale comes after production. Once you have the video encoded on a server with a stable connection the hard part is over. What netflix failed to do is spread the files to enough servers around the globe to handle the load. I'm surprised they were unable(?) to use their network of edge servers to handle the live stream. Just run the stream with a 10 second delay and in that time push the stream segments to the edge server

Yes, depending on thresholds which vary by state. For some you have to figure out taxes from the first transaction, for others from the 100th and then there are some that consider revenue, like $100k. At least at a conceptual level, taxes in the US seem much more complex than the EU, where you have only one tax rate per country.

On hetzner it gets you at least 4 cores, 8GB RAM and 80GB local SSD. For $49 you can almost get a dedicated server with 8 cores and 64GB RAM. More than enough to handle that load Edit: this is for $8 but general point still stands

For this use case i would use the bunny volume tier which is $5/TB

Testing some Hetzner Cloud instances

16-core ARM CAX41

root@ubuntu-32gb-fsn1-1:~# ./a.out

1 17.8

2 35.6

3 49.9

4 64.9

5 79.0

6 92.1

7 101.7

8 110.2

9 117.6

10 124.0

11 130.4

12 136.2

13 140.9

14 143.8

15 149.2

16 152.9

16 Core AMD shared CPX51

root@ubuntu-32gb-fsn1-1:~# ./a.out

1 18.9

2 37.5

3 54.3

4 65.4

5 77.2

6 92.8

7 100.8

8 92.1

9 95.1

10 105.7

11 93.4

12 100.9

13 89.9

14 97.0

15 99.1

16 107.8

16 core AMD dedicated CCX43

root@ubuntu-64gb-fsn1-1:~# ./a.out

1 36.3

2 71.6

3 56.8

4 50.8

5 63.4

6 57.2

7 53.4

8 50.9

9 55.2

10 61.3

11 64.4

12 65.3

13 66.5

14 68.8

15 69.2

16 64.4

Increasing the data volume for CCX43 to 48GB increases the bandwidth using 16 cores to 75.1. Memory bandwidth seems to scale pretty well with the number of cores on ARM. Interesting that the bandwidth for the shared AMD system is that much higher than the dedicated system

I'd probably not say anything. At most anonymously. I'm not taking the risk as i stand nothing to gain

Finding peers is somewhat slow. When downloading via torrent it takes a while to really ramp up in speed. With a centralized repository you can start downloading basically right away. For many small files the latency is more important than the bandwidth, given that most repositories are not short on bandwitdth.

The trained model would likely be able understand both Alienese and English equally well, but it never learned to translate even one word or context. It might have an internal representation for "eating food" in both languages, but since since no links exists between the languages the embeddings will not be close.

You could try it on earth with if you train a model on two separate languages, being careful that the traning data does not contain any mixed language. But even then, modern Human languages most likely have too much cross-contamination. Would be an interesting experiment nevertheless

Moving away from your established life, friends, family to a country where you will (at least initially) be foreign, where the lifestyle is different are burdens that should not be underestimated.

I'm sure everybody in the EU would like US level salaries, but most wouldn't move for the money alone

Thanks!

What is the library being used to display the map?

Thank you for your answers.

That sounds like a reasonable and balanced approach.

Can you explain the reputation system a bit more? Is all content in freenet public or how would you know that someone posted certain content? How are they punished for it? Do other users actively have to "down vote" them to reduce their reputation or is there an entity which can influence reputation? Is there are single global reputation system or is it more local?

Internet Speeds in Germany are lacking in some areas, but fast internet is generally available. I have a symmetrical 1 gig FTTH connection in the countryside

That's the default in Germany (and Europe?). Writing out the month is somewhat rare, although it does happen.

No. Privacy has to exist at the base layer.

Running a hidden service on Tor should work without problems behind NAT

It's already happening. You cannot get better than 720p on most services on linux, while pirated content is available in 4k