It is not.

> Learns what the model is trained to do just from the weights!!

expand prompt

> Describes Breakout in great detail

Also... "Each book links back to the comments that mentioned it."

> Please don't put generated text in HN posts.

Claude speak gives me a gag reflex these days.

https://news.ycombinator.com/item?id=48892468

They threw the book at him. He was charged and faced a maximum of 50 years in prison.

OpenAI committed very similar crimes on a much larger scale, yet the most they will see is civil litigation.

Insiders have called it the largest theft of labor ever in history[1].

In the USA, corporations enjoy constitutional protections, the right to free speech etc, yet they are rarely if ever charged with crimes. For examples, see how the courts handle:

* large scale systematic wage theft by corporations vs petty theft by an individual,

* negligence infecting a 10k Americans with HIV by selling known-to-be-contaminated blood products (Bayer)

1. https://arstechnica.com/tech-policy/2026/09/microsoft-exec-c...

Imagining a mode where the agent is very transparent about its direction and process, and I can passively interject at any time to steer the outcome, without having to wait my turn or hard-interrupt.

She stayed in the race for Super Tuesday in 2020 rather than consolidating the progressive vote for Sanders.

Right before, her campaign accepted millions in funding from billionaires after promising not to. Her comment was roughly, as soon as others stop taking billionaire money, I will stop.

So perhaps you are a constituent part of the larger consciousness, and you feel your part of it (e.g. your roles and relationships in society)

Only the US sees this as a competition, new space race, Cold War, etc.

The way these “AI is too powerful now” articles read about Mythos, Fable, GLM, etc is completely incongruent with my experience using them. It feels like they are all trying to position themselves to influence government policy.

Claude will, without prompting, break the implementation into 6 phases, and write AI slop “code as English” specs for each phase, each one with glaring errors and unintelligible terse jargon. It will review them all several times with major findings every time and tons of design churn…

Then it will implement a total heap of garbage over many hours of many agents, despite it working in “lanes” and in parallel, and with regular input needed.

Just thousands and thousands of lines of junk, which auto review then plays whack-a-mole to polish and fix. All the while it is able to see the errors and edge cases, yet fails to see the key architectural blunders that led to them in the first place.

I’ve had to fully rethink my approach to LLM tasks like this. For example, for library-esque modules, I have found that isolating the problem outside of the codebase is one useful approach. Something about the lack of noise. It can land on a cleaner solution that can be retrofitted.

I also find that asking it to implement end to end in one fell swoop with a very high level plan actually saves time and creates a cleaner result.

I have found that WG client management is as simple as a couple scripts and a TSV file to match pubkeys to users.

Anyhow, cool project and I like the overall concept of respecting existing config.

This is why I made ParaSpace (https://paraspace.dev/)

With ParaSpace I can use a real terminal and my normal tooling to develop locally on fast booting isolated containers. I even get to bring my own dotfiles.