gilgoomesh

It feels like it's trying to evoke Bioshock vibes or a similar retro-punk dystopia.

It is common for the agent tools, MCP tool, skills, project context and system prompt to take between 5k and 25k tokens. This depends heavily on your agent and tool setup.

But that's before the prompt.

Then after the prompt, every tool call the agent makes adds to the context. Longer turns can easily consume 50k-100k tokens between the agent and various tool calls (reading the filesystem, reading files, reading compiler output, reading memories).

Then each "turn" with the agent stays in context and is fed into the next turn. Two or three turns and you're up near 250k.

Easy. I just need... 50 times more speed.

Better download some more RAM.

Destroying currency is crime in Australia, so no, this has never been a thing, here.

Yeah, I must have misread the press release last week as I thought it would be released at the same time.

It's been 2.5 months since Qwen 3.7-Max. Unless you mean since the preview of this 3.8-Max model.

> lack of European competitors

There's Mistral. Although they've been surprisingly quiet for the last 6 months.

...not going anywhere.

Goldbach's Conjecture is very accessible.

I'm disappointed only that the chain of thought needed to be rewritten. Need to train these LLMs to natively communicate in LaTeX research paper format.

> It’s not hacking, I wasn’t breaking anything

That's a very narrow read of the word "hacking".

We're literally on a website called "Hacker News". We're not all trying to break things.