From the very beginning, I kept a 'dirt notebook' only, but because of its lack of structure, some thoughts would occasionally get lost or were hard to make sense of. Then I switched to ColorNote, but it has poor formatting. Recently, I asked an LLM to divide my ColorNote notes into notes and tasks in .md format for Obsidian. The LLM, by the way, did a bad job: some notes were mistakenly identified as tasks and vice versa. Also, I had a problem with task notifications because tasks in standard Android apps like Google Tasks are easily swiped away from the phone's notification shade. As a result, I installed Notification Notes, which lets you change the behavior of notifications.
HN user
AlanAzarkin
Agreed. Those guys put together as many as four configurations that are cheaper than the Steam Machine https://www.digitalfoundry.net/features/build-your-own-steam...
1. Maybe OpenAI is preparing a lawsuit against Apple to secure an antitrust ruling that would select OpenAI models for Siri. 2. Maybe Apple is preparing a lawsuit against OpenAI to force them to disclose their developments during the discovery process.
QA is more likely needed for humans, but in the LLM era, when agents write consistent code, there is no need for it anymore in my opinion. A bad example, but still: I recently developed a TG bot without a single test, and yet, it works.
Reading about Stripe’s stablecoin ambitions is funny. X uses Stripe, but Stripe on X doesn't accept my Bybit stablecoin cryptocard for a Pro subscription, ads, or boost purchases.