More

airstrike · 2026-03-05T15:59:57 1772726397

A travesty, to be sure, but not insider trading.

airstrike · 2026-03-05T14:13:46 1772720026

> Someone else downloads it as an xls and manually writes (not copy pastes) the numbers into a power point presentation and makes graphs by drawing shapes. This is then presented at some bi-monthly meeting.

I made an app that fixes this part of the problem. The rest is cultural.

airstrike · 2026-03-04T16:12:14 1772640734

if you add 2 spaces at the start of the line, you turn it into a code block

  like this

andai · 2026-03-04T16:58:26 1772643506

  | Model | DocType% | Year% | Subject% | In $/MTok |

  |----------------|----|-----|----|-------|

  | llama-70b -----| 83 |  98 | 96 | $0.72 |

  | gpt-oss-20b ---| 83 |  97 | 92 | $0.07 |

  | ministral-14b -| 84 | 100 | 90 | $0.20 |

  | gemma-4b ------| 75 |  93 | 91 | $0.04 |

  | glm-flash-30b -| 83 |  93 | 90 | $0.07 |

  | llama-1b ------| 47 |  90 | 58 | $0.10 |

azath92 · 2026-03-05T09:04:22 1772701462

thank you so much! i suffered with this, and now i never will again!

airstrike · 2026-03-04T16:10:17 1772640617

I'm hopeful they will pick up their work elsewhere and continue on this great fight for competitive open weight models.

To be honest, it's sort of what I expected governments to be funding right now, but I suppose Chinese companies are a close second.

airstrike · 2026-03-03T16:39:22 1772555962

I'm worried about my browsing to be tracked across the entire internet for the purposes of marketers to "enrich" my profile... just to sell me more and to sell that data to third-parties who can make all sorts of decisions based on a made up story about who I am, my preferences, my values and whatnot.

there's a reason I don't walk around naked either. it wouldn't hurt me, but I don't need that kind of exposure for no upside

caseyohara · 2026-03-03T17:28:25 1772558905

> third-parties who can make all sorts of decisions based on a made up story about who I am, my preferences, my values and whatnot

You're going to be presented with ads and preyed on by marketing no matter what. The "made up story about who you are" is just even more imaginary the less they know about you. You'll simply be presented with less-targeted ads.

mixmastamyk · 2026-03-03T18:50:09 1772563809

Not the point, no one benefits by having an accurate (or non) dossier built on them, up for sale. The drawbacks may be infrequent and postponed but as history confirms, quite real.

airstrike · 2026-03-02T13:19:35 1772457575

How do you do, fellow kids?

airstrike · 2026-03-02T09:07:25 1772442445

1. I think at a minimum we need a separate "Show HN" for AI posts, that people can filter out, so that users are not incentivized to spam Show HNs hoping to make it to the front page

2. Then that separate group, call it "Vibe HN", gets to decide what they find valuable through their own voting and flagging.

Some guidelines on what makes a good "Vibe HN" post would be helpful to nudge the community towards the things you're suggesting, but I think (1) cutting off self-promotion incentives given the low cost of creating software now and (2) allowing for self-moderation given the sheer number of submissions is the only tenable path

airstrike · 2026-03-01T16:17:04 1772381824

Isn't this just asking Claude to make a website? Even the style seems to suggest that. And if so, this is slop. It's not really adding to the conversation.

airstrike · 2026-03-01T02:14:08 1772331248

is your prompt literally 1-sentence?

if so, a better approach would be to ask it to first plan that entire task and give it some specific guidance

then once it has the plan, ask it to execute it, preferably by letting it call other subagents that take care of different phases of the implementation while the main loop just merges those worktrees back

it's how you should be using claude code too, btw

nl · 2026-03-01T03:05:19 1772334319

Claude Sonnet can easily one-shot that without specifically asking for plan first.

airstrike · 2026-03-01T15:22:01 1772378521

I believe you, but performance on 10-word prompts is pretty useless as a metric

nl · 2026-03-02T01:25:57 1772414757

Why? Seems like a valid requirement to me?

I build micro apps from 10-word prompts multiple times a day.

airstrike · 2026-02-27T21:09:49 1772226589

Plus a bunch of other elements that are dead giveaways.

In this day and age, I wish people would ask any model OTHER than ChatGPT to rewrite their shit. At least we'd get a different flavor of slop.