Recent submissions

Comment · 17 days ago
Comment · 1 month ago
> Opus 5.5 is the first Opus model to launch with a similar class of safeguards to Fable 5.1 on cybersecurity, biology, and distillation, all of which fall back to another model transparently.

This is where Chinese models are going to eat Anthropic's lunch.

Comment · 1 month ago
Speculating here, but I assume researchers can make a reasonable estimate of the size of closed models based on factors like training time, training speed, and the number of tokens processed.

Also, Anthropic and OpenAI probably want to keep each other on their toes so they don’t end up on the wrong side of another Opus 4.6 / GPT-5.3-Codex situation, where one lab releases a model only for the other to drop a better one hours later.

Comment · 1 month ago
Funnily enough that is against the Instagram ToS (https://help.instagram.com/termsofuse):

> 4.2 How You Can't Use Instagram. [...]

> You can’t sell, license, or purchase any account or data obtained from us or our Service, regardless of whether such data was obtained while logged-in to an Instagram account.

> This includes attempts to buy, sell, or transfer any aspect of your account (including your username); solicit, collect, or use login credentials or badges of other users; or request or collect Instagram usernames, passwords, or misappropriate access tokens.

Comment · 1 month ago

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton