Hacker Newsnew | past | comments | ask | show | jobs | submit | JSR_FDED's commentslogin

Richard Hendricks solved this decisively with middle-out compression

Safe in this context means not corrupting the database. You might lose data, but the database will live on happily as if the missing data was never written to it. You don’t get half-committed transactions, and the database isn’t in a weird state.

> even if the server crashes, only the uncommitted transactions in the WAL are lost

I'm mainly referring to this quote.

So, you can lose committed transactions too.


I’ve never found the time to dig into Tailscale so this was a good high level intro to how it works

Yes, I was curious to understand how it works and wrote it to remember.

I see no mention of API use on the website, so it’s more of a privacy option for chat. That in itself is useful (no logging, no training on your data, encryption so Proton can’t see or share your data, no ads).

Another super useful feature is Accessibility > Zoom > Use keyboard shortcuts.

The you can use Command+Option+Plus and Command+Option+Minus to zoom the entire screen.

It’s a great way to zoom in on anything regardless of whether an app supports it or not. Sometimes I help family members remotely and look through a bad camera at a document on their side, and applying this zoom makes the difference between being able to decipher the content and not.


Lots of respect for a solo founder. But you submitted this yourself, and it’s incomprehensible.

For which funding will be immediately halved by the administration

I like the 2x2 grid that describes when to fine-tune a model, when to use a frontier model, etc.

From the article it’s not clear how the scorer grades every episode - was it a frontier model that assigned the grade? How does that continue to work as the model that is being fine-tuned becomes better at the task than the frontier model?


From TFA: “This is a tongue-in-cheek attempt to demonstrate what a human can do that an LLM cannot”

From the other comments here it seems there’s another thing humans can do that LLMs can not: choose not to read the article and respond only to the headline :-)


Works surprisingly well, had a super sensitive passenger in the car the other day and zero signs of nausea.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: