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

I've watched Patrick's videos for enough years that I know he is not, but I still wonder from time to time. His voice is incredibly flat and uniform, he always uses fake backgrounds and there is extremely high use of jump cuts in his edits.

I suspect that he at least uses AI for scripts. He tends to repeat the same thing worded slightly differently a few times.

You may be surprised by how much easier it is to dump the framework/stack and just write it from scratch. I say this because I too work on compilers and have a crud app as a personal project. The first versions were a nightmare in various frameworks and since I switched to a C++ backend / vanilla .js frontend it has been incredibly easy to write.

interesting, how did you manage the db interactions in c++? and did you have a single executable for the app and the web server combined?

no database for this project - the data model has a simple text representation so it gets serialized out to a folder/file layout on disk that goes into version control. single self-contained binary: contains the web/websockets server, backend logic, parser/serialization. there is a separate component in python that sits behind an internal network connection to handle an execution sandbox.

The rate was 1% so this does not mean that "many" people are using LLMs despite claiming not to.

This comment doesn't seem to fit the discussion at all?

The discussion is not about humans using LLLs to write papers. It is about humans who agreed not to use LLVM in reviewing papers, then did exactly that.


There's a lot of irony in a defensive comment being written based on misreading / inattentive reading of a post about reviewing papers (requiring attentive reading).

It might be that paper authors required others not to use LLMs for reviewing their work. Then, by the rule of reciprocity, they shouldn't use LLMs for reviewing others work. The article is unclear on whether this implied reciprocity rule was explicitly stated or not.

It was. More details here: https://icml.cc/Conferences/2026/LLM-Policy

In particular: "Any reviewer who is an author on a paper that requires Policy A must also be willing to follow Policy A."


In addition to being a reviewer, they also submitted their own research to this journal. So it leads to the question: if they were willing to cheat on the side of review with less incentive, why wouldn’t they cheat on the side that provides more incentives?

(Meaning, your career doesn’t get boosted much for reviewing papers, but much more so for publishing papers)


As the DJ is an interface to shuffle, and the author specifically wants to listen to unshuffled music the lack of intelligence may not be entirely in the AI.


No side profile pics on the page. My only concern would be can it lay flat on a table for taking notes or does it have a camera bulge that makes it wobble?



I hate apple. Can't they just add a second bump on the other side? They're being a PITA with this wobble and it's been going on for like 15 years now (iPhone 7 forward)


FWIW, the iPad Air I bought a couple years ago has a small protrusion for the single camera lens, but does not wobble when laying flat and is not really noticeable. This latest iPad Air has a similar design.


I would imagine that the target audience has an attention span and literacy level that allows reading sixteen thousand words without too much trouble.


I like the idea that people are downvoting and not rebutting a summary because they think that an accurate summary will cause folks to not read sixteen thousand words that they would otherwise. It’s kind of agreement by dissent


I personally love the appearance of the tl;dr about a third of the way through, that is some S tier trolling.


Remarkably it has only cost a few trillion dollars to get here!


don't forget the insane costs to stay here


Reasonable hypothesis. Supported by data. Seems legit


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

Search: