Recent submissions

Comment · 11 days ago
an inteligent summary is a user feature, that is not a matter of opinion
Comment · 11 days ago
the position on the project is both prescriptive and descriptive because the project actively prohibits AI
Comment · 1 month ago
i would argue that the way internet infrastructure in general already works as such system, local AI tools just set that as a commodity
Comment · 1 month ago
It would be so much work to follow any of this that the answer is guarateed to be a no
Comment · 1 month ago
on fish shell i never felt the need for any of this, since pressing up does a history search on what you have typed
Comment · 1 month ago
i like the decision prompts taking the whole fucking screen, so true.
Comment · 2 months ago
you can always use /rewind to go back, always work thinking "can i shorten or clear my context now? and how should i do it?"
Comment · 2 months ago
i think they call it a web browser, but it looks like it does not do that most of the time.
Comment · 2 months ago
this touches one of the reasons opus 5 came by default with 200k context and 1M gated behind usage tokens, auto-compaction by default, and it keeps telling you to clear and start from scratch all the time, AI is does a better job fresh, telling multiple agents to pipe around ideas without the human intent is the worse thing you can do
Comment · 2 months ago
that's just making a good and complete product then
Comment · 2 months ago
kinda of a conspiracy theory, but maybe this is why anthropic and other companies are looking into fingerprinting now
Comment · 2 months ago
my hypothesis is that its trying to hide the thinking process so people can't train models on the output, try to learn anything complex using AI, its basically imposible, its like its actively fighting giving you the main rationale
Comment · 3 months ago
this idea of smaller tasks is me simply asking "quick question, foo?" to my fable 5 on max effort, the model just does it, and even gives a more complex answer on misunderstanding or a actually hard question
Comment · 3 months ago
i let it run once to just generally find any bugs using all my quota, them once i get back to it i ran on effort /max to continue with explicit don't run workflows fix all yourself, it was ultra effective.

ultra becomes a mode where your model enhances your prompt them it is execution multiplied by all other agents, which have them going in all sorts of directions.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton