Aeolun

31 ms·

871.
by Aeolun 1y ago
Umm, Claude Code is a lot better than a lot of enterprise grade code I see. And it actually learns from mistakes with a properly crafted instruction xD
872.
by Aeolun 1y ago
Huh, this must have changed after I concluded my contract with them (several years ago). Huh, archive.org tells me its been unlimited since at least 5 years ago, so I guess I must’ve just seen someone mention 20TB and felt it was a reasonsb
873.
by Aeolun 1y ago
I’m sure that they can write that. But their actual enforcement mechanism is nonexistent. No country is going to work with the UK to arrest someone that does that when the same thing isn’t illegal under their own law.
874.
by Aeolun 1y ago
I dunno, I deliberately talk with Claude when I just need someone (or something) to be enthusiastic about my latest obsession. It’s good for keeping my motivation up.
875.
by Aeolun 1y ago
That doesn’t make any sense. It’s not wikimedia’s responsibility to ensure people from the UK don’t hit their servers by typing wikipedia.org into the browser bar.
876.
by Aeolun 1y ago
Gitlab is not really an option for me. Their pricing is absolutely out of this world.
877.
by Aeolun 1y ago
They don’t really need to surveil their citizens. The indoctrination starts from kindergarden :)
878.
by Aeolun 1y ago
It’d be good for all mental health. Banning the whole thing is probably a lot more sensible, but then people would have to face their own addictions.
879.
by Aeolun 1y ago
The standard for robust and effective age verification is extremely low, given how I’ve seen anyone do age verification. It’s also pointless if we’re talking about the internet, you are essentially outsourcing your porn production to foreig
880.
by Aeolun 1y ago
Depends on your connection I think. Mine do 1Gbit/sec but have a 20TB limit. The 100Mbit ones are unlimited (last I checked)
881.
by Aeolun 1y ago
I don’t think I’ve ever migrated anywhere and not kep the previous thing running for a month unless I wasn’t worried about exactly that thing happening (to be fair, not prod traffic blackhole, but similar effect)
882.
by Aeolun 1y ago
All of Ollama and Stable Diffusion based stuff now works on my AMD cards. Maybe it’s different if you want to actually train things, but I have no issues running anything that fits in memory any more.
883.
by Aeolun 1y ago
56 Gbit/sec costs you about €590/day even on Hetzner.
884.
by Aeolun 1y ago
I think, when you read that, you should be reassured that nobody is going to suddenly tell you to pay, and then still implement caching on your own side to preserve the free offering for everyone else. Seriously, whose first thought on read
885.
by Aeolun 1y ago
If you do 200Mbps on a hetzner server after cloudflare caching, you are going to run out of traffic pretty rapidly. The limit is 20TB / month (which you’d reach in roughly 9 days).
886.
by Aeolun 1y ago
This all seems to be Hetzner Cloud offerings?
887.
by Aeolun 1y ago
> They’re unreliable as soon as you have to deal with their support who have the technical knowledge of a brick. Since I never have to, that’s perfect isn’t it? If you need support from Hetzner you are using the wrong host.
888.
by Aeolun 1y ago
I think it’s reasonable to assume that a free service is not going to deal gracefully with your 100k rps hug of death. The fact that it actually did is an exception, not the rule. If you are hitting anything free with more than 10rps (tempo
889.
by Aeolun 1y ago
I mean, it’s not that bad. It’s bad at all the same things that other models are bad at. I just have no reason to switch away from Claude to GPT-5
890.
by Aeolun 1y ago
I don’t think that makes intuitive sense. Whether I send 50k rows or 10x5k rows should make no difference to the database. But somehow it does. It’s especially annoying with PG, where you just cannot commit a whole lot of small values fas
891.
by Aeolun 1y ago
The model has to be trained on something though. It’s easy to see how this works for art art, because people have been drawing that shit for years, and most people wouldn’t feel to bad about training on it. I don’t think that’s true for the
892.
by Aeolun 1y ago
I feel like the people that would be capable of giving you an answer would be very careful not to actually give you one. You can’t really research when the only thing you can be certain of is the known real cases. It’s much harder to quan
893.
by Aeolun 1y ago
Maybe you have a porn test suite for LLM’s? See which ones are fine with or capable of talking about specific topics? I believe there was something similar for willingness to discuss sciency stuff.
894.
by Aeolun 1y ago
I'm just sitting here hoping that their lowered prices will force Anthropic to follow suit xD
895.
by Aeolun 1y ago
I feel like this is just related to my projects getting bigger. Claude Code is trying to keep up with my project evolving from 2k lines of code to 100k lines. Of course it’s going to feel worse.
896.
by Aeolun 1y ago
Is this on x5? Because ever since they booted all the freeloaders I’ve not once seen the “you are approaching usage limits” message. Anyway, the “you are approaching usage limits” message shows up when you are over 50% of your tokens for th
897.
by Aeolun 1y ago
Claude Code definitely reports cached tokens, and I think CCusage does too, so it wouldn’t make sense for the calculation to be based on full pricing when they have the cached values.
898.
by Aeolun 1y ago
My opinion of Opus is that it takes the correct action 19/20 times, where Sonnet takes the correct action 18/20 times. It’s not strictly necessary to use Opus, but if you have the subscription already it’s just a pure win.
899.
by Aeolun 1y ago
In every price comparison I make. Claude (API) always comes out cheapest if you manage to keep most of your context cached. 90% price reduction for input is crazy.
900.
by Aeolun 1y ago
> this isn't interfering with their open source project Wait until it makes significant amounts of money. Suddenly the priorities will be different. I don’t begrudge them wanting to make some money off it though.

More ›