clouedoc

Let's create a Discord/Signal/WhatsApp/mailing list group to help each other figure it out... it's time to end our sleep irregularity once and for all!

glad to see someone giving a source. a lot of people here heard of fiber optic drones and decided that now 100% of drones in Ukraine are using those

Cool! I'll definitely use it for my next japan trip :)

Everyone has the urge to criticize what they don't like, and it looks like you are no exception... :)

Thanks, means a lot to me :-) !

I made a little tool to help me find the best Hetzner auction deals and it turned out pretty well!

I discovered this PostgreSQL CLI by accident the other day, and it's been a real upgrade from `psql`.

Yes this is the solution to this problem. This is what SoundCloud is doing AFAIK.

SigNoz v2?

Maybe it could be useful to be a little more specific? I am using OpenTelemetry + ClickHouse + Grafana.

In the past, this setup was really hard to get right because of the endless configuration options + having to write the Grafana dashboards yourself.

However, you can connect Claude to your server, ClickHouse and even ask it to create the Grafana dashboards.

This is wayyy cheaper than using something off the shelf like Datadog.

At most, PoW makes it a bit annoying to scale: you need to add some form of RPC that delegates solving to a beefy+cheap Hetzner server. If you're really scaling and it's getting expensive, you can rent a GPU to do batch solves.

PoW systems like Anubis are self-soothing.

Same. I got a crazy old Ubuntu desktop when I was 9 or something. It couldn’t run any games and that’s why I learned Python. I wouldn’t be who I am today if I had a machine capable of running Minecraft at the time…

It didn’t look like bragging, if that’s what you were worried about. Thnaks for answering :)

I know it's not really related, but how did you manage to build two startups worth getting acquired in such a short period of time?

Recently, what have been working for me is to take a day off just to explore stuff that interests me related to programming. I've been learning sleep states and working on the Rust track on Exercism.

The productivity boost that you get from Claude Code should be reinvested into learning your subject area deeper, rather than just churning out more superficial code.

It makes me happier this way. Experiencing flow is the best way for me to feel happy, and programming was the most reliable way to achieve that before Claude Code / Codex.

Gym feels lonely. Social sports are better! E.g. badminton

With time I actually came to get accustomed to it and to enjoy my files not reloading automatically with Claude Code changes.

100% agree. Everyone should always backup their production database somewhere where's it's not trivial to delete.

I think this blog post from the author is interesting: https://justin.poehnelt.com/posts/rewrite-your-cli-for-ai-ag...

Exactly. Big GPUs are the #1 reason battery doesn’t last on Linux laptops.

Power management is not done well with the GPU drivers in Linux. If they are not used, they still draw a lot of power, while that’s not really the case on Windows, from what I heard.

I think the best is to get a good Linux laptop, but with an integrated GPU. If you really want to do anything beefy, you can always use an eGPU :)!

Obviously this will never come close in terms of convenience as having an actual M series MacBook…

Wishing you best of luck for the .net migration!

What about MacBook Pro M5? Waiting for that one.

I am sure some people can. I got pretty far in the interview process at xAI, but couldn't get anyone at other AI labs to even reply my emails, though...

(I dropped out at the end because there were issues with HR not wanting to hire in certain geos, and me not wanting to relocate to the USA or London)

On here: https://cc-context-mode.mksg.lu/#/3/0/3

Bun auto-detected for 3–5x faster JS/TS execution

This is quite a claim, and even so, doesn't matter since the bottleneck is the LLM and not the JS interpreter. It's a nit, but little things like this just make the project look bad overall. It feels like nobody took the time to read the copy before publishing it.

More importantly, the claimed 98% context savings are noise without benchmarks of harness performance with and without "context mode".

I'm glad someone is working on this, but I just feel like this is not a serious solution to the problem.

It's hard to deny an offer to become a millionaire in the next 3 years if you just hang tight at xAI, especially if you don't have any offers from competing AI labs. Also, LLMs are converging into an easy-to-replicate commodity. It doesn't matter much who wastes their money on you to build them.

I'm really curious what were the magic words.

Alex had actually tried to brute force the hash earlier, but had downloaded a list of the top 10,000 most popular words to do it, which turned out not to be big enough to find it. Once he had a big enough word list, he got the answer.

They don't reveal the answer.

Maybe selling upvotes? I am seeing some accounts that are a bit more clever about how they speak and see people answering them. That's sad. Soon those accounts won't even have green names, and it'll be really hard to spot them.

What's the meme with SoftBank? Just that they're super bad at investments?

That's funny but it must absolutely drain the battery of her phone, no?

What are the drawbacks of HAPI? Seems too good to be true. Will give it a try.

Same. This price increases is making me think about switching all the workload that I have on Hetzner Cloud into dedicated auctions server.