pimeys

Because the prime minister has been running a government that is disliked from all directions. Because the economy has been going down and hard. Because the unemployment rates are the highest in EU.

And now he can shake hands with Google executives in a public photo.

People still live in the old world that thinks that doing business with the American corporations is somehow beneficial, especially with data centers. But the country should focus on building their own data centers, providing GPUs for smaller companies and providing cheap DeepSeek for Europeans instead of giving access to Google.

Here some pics from those empty skyscrapers that are completely missing all the walls and the scaffolding's been gathering rust for years. And all the court cases for the people who bought the "luxury apartments" that never realized.

I'm always surprised that the topography of terror exhibition in Berlin misses this part of the German history. Nazi's and communists are presented, but the robber barons who bought all the real estate don't get a presentation.

Yes. Should also cover what happened in East Germany when the wall went down. And people are now somehow surprised when everybody's kind of pissed outside of the few big cities.

Like... the town where the Google datacenter goes gets like some paved bike road from Google who "takes care of the community". Price tag around the profits of less than an hour of the GPU time in that datacenter.

An article in Finnish: https://www.uusijuttu.fi/juttu/sCcu9YiY-mjF4UD1O-b66ec

A bottle of vodka for your shares, comrade. Etc...

Oh the logging projects also get a LOT of attention. It's just that the current government is center-right / far right, and they just let the logging to continue no matter the sanctions from the EU.

There are a lot of people who are against this, and protest against this. Looking at the polls, seems that it will change in the next elections, like a lot.

Finland has a lot of people who care about forest and Finland needs to pay fines for EU if they don't have enough forest to suck CO2 from the air.

Google cut corners here, and didn't ask for the needed permissions. And the prime minister doesn't want to show his conversation with Google regarding this.

People are really pissed.

Edit: and the electricity prices are going up due to these data centers. And the profits from them don't stay in Finland, but go to American corporations. There should be Finnish data centers owned by Finns instead of having Google to take the cheap electricity and taking profits out.

Yes. Yle is awesome.

Deepseek is what, maybe 10-15 dollars a month for quite a lot of usage if you need an llm and do not want to pay for Anthropic...

I also used to plan with more expensive models like you did, but now I just do planning with DeepSeek too and don't see much difference. I use the omp plan mode and iterate on the text until I'm happy with it. Then clear context, give the plan to DeepSeek again for implementation.

If it's mission-critical code, do a security review with GLM 5.3.

Otherwise go manually through the code and ask for changes if needed.

Works really well for me.

I'm using DeepSeek a lot, and I'm curious how you get a task to go to ten dollars with it? Or are you mixing Opus and Sol tokens to the same task? I have very hard time to go over 5 dollars with DeepSeek and that is already a weeklong project where I share the same context.

Of course I do certain things like use omp and do a compact picture snapshot after 256k tokens automatically. My tasks rarely go over a dollar...

I've ported a few games, wrote a cosmic desktop app, iterated in my own agent software, use the model in our home speaker system, use the model in my phone assistant. My monthly bill to coralbricks is about 14 dollars.

Or just us Europeans who are so sick of American corporations milking every penny with their LLM models and preventing us from using the full model capabilities.

Now we have: a ton of competition in different inference providers, very low costs, and speeds reaching 1k tok/s in certain providers.

We can just get a box of compute even in Germany, and run the exact same model here without anybody stopping us.

It's quite different here in EU. I just gave a presentation in an AI conference in Paris about moving from Anthropic to open-weight models, how it looks like and what kind of savings you'll be looking at. The room was packed, and after discussing with a bunch of VPs, staff engineers, and CTOs in the conference, every single one of them are either looking at or using DeepSeek now.

It's quite hard to have the valuation of Anthropic if they are limited to US.

They do not yet. They are testing it but they go with GLM right now.

Scaleway, Mistral, Lyceum, Verda, Silo AI, Siili, Upcloud, CSC, Gcore, Nebius...

I mean... I ported some of my childhood Finnish shareware games for Linux and Rust with DeepSeek V4.1 Flash. It took about six hours of spelunking and finally I was pretty happy with the results! I did it twice. Tokens for both of the games were around 3 euros.

Because it doesn't make difference for my use. High is good enough! If you do an assistant for yourself, low is more than enough.

It reduces the speed and tends to overthink sometimes which is really annoying. For programming, 4.1/high is the bread and butter in my daily work.

There's also European providers for these models. More than one!

It thinks a bit more, but if the cache reads are free or almost free, and it blasts through 10x faster than GPT Sol/Terra/Luna, it doesn't really matter. Just don't use the max thinking for implementation, go with high or low.

There are multiple others in EU, and in US. Cheaper, more expensive, slower, faster. So much choice.

Seems that V4.1 Flash is quite cheap to serve. Especially how it implements the cache, meaning the cache pricing can be very low depending on how you implement it in your inference engine.

Depends on your use case. But if you do software development with a harness such as oh-my-pi, your cache rates are probably somewhere above 98%. Coralbricks has free cache reads, meaning it's much cheaper than the Deepseek API rates.

Lithosai is more expensive especially in the ultra tier. It's if you want extremely snappy and fast responses, which is important e.g. for a personal assistant where every second counts.

If you generate a lot of output, you're better off with the Deepseek API rates.

Speed and price. You should try the 4.1 Flash on one of the fast providers. It's pretty crazy how fast this model works, and it's great for iterative software development because you can steer it fast to where you want to go.

https://www.lithosai.com/

https://www.coralbricks.ai/

In Berlin they started following the EU law and you can quite often pay with a card. Even in places like biergartens. Still having cash when going to a Kneipe, although last time they also accepted cards which was a bit weird...

I live in Berlin... So naturlich!

I just decided to completely stop paying with NFC. I always carry a few cards with me and Wero is already working in a few spots, which requires just your banking app and a working camera.

I don't even have Google Wallet installed anymore.

Just be careful. I locked myself out from my bank when I installed a second phone that did not have sim card yet and had the app in my old phone.

It requires a sim card and cannot be used from multiple phones. So they put you to this endless face scan loop and then lock you out.

Good to know. I'm using N26 and they work just fine in Graphene OS. What I do is I create a private space for the apps needing play services, which I keep locked most of the time. This acts as a separate profile and when locked, the apps including play services are completely off. My main profile uses only open source apps and no play services.

I actually called N26 (I'm a Metal customer with my own phone support) and asked will they support Graphene OS or no, and they said to me they will and gave me instructions what to keep in mind when installing the app.

His movies are also worth a watch: Hana bi, Sonatine, Brother, Dolls, and Zatōichi are all amazing movies.

Should do a rewatch again, it's been a while...

I like to do some discussions and back and forth with GLM. It doesn't hallucinate as much or gaslight like DeepSeek does. DeepSeek is great if you have a plan and you let it go. Also can plan a bit. But it has this sparse attention and especially the index_topk is too small in common deployments so it doesn't "see" all the context. And it then imagines things which is annoying if you chat with it about stuff you develop in the session.

I recommend trying Coralbricks with GLM due to their cheap input cache prices. GLM can get quite expensive elsewhere.

Yes it brings me joy and sadness at the same time. And anger.