Nikhil Harithas

Hello I am Chappell Roan’s bodyguard who yelled at that 11 year old girl. My name is Fujimoto Tatsuki, I'm the author of a manga called Chainsaw Man

When Salah, Mane, Coutinho and Firmino are on the pitch together, one of them scores every 23 minutes.

Absolute icon. Never forget it. #LFC

@alphatozeta8148 @opencode droid would never 😇

@opencode WTH

mcnulty: sorry i’m late, i was having sex. where are we on stringer bell kima: sorry i’m late, blame my annoying baby. gotta go find bubs bunk: despite drinking enough to kill a horse nightly, i’m never late lester: I Am Become Computer

Modernity is obsessed with instant legibility but the most interesting people and projects are not. They trust themselves and they trust us.

@Suhail We literally did this to our data repo at @FactoryAI with hooks! docs.factory.ai/cli/configurat…

It feels like someone should make a post-git-hook where it asks the AI model to look at the diff of what you changed for a merged PR and update the repo’s various readmes and other documentation to make it easier for an LLM to be able to write code and reference things faster rather than reading every single line of source code that might be relevant constantly. The agents need their own docs.

@Suhail Dm'd!

my fav type of AI person is the claude-pilled nuancemaxxer who is excited about the tech, capable to reason about risks & doesn't participate in binary culture wars where you either get gulagmogged or are supposed to be stoked about getting turned into a dyson sphere component

in every engineering org there is a secret war between the normies and the freaks

@badlogicgames @matanSF The criteria is fixed. How it decides a criteria should be skipped or not depends on the criteria in question, def tuning skipping strategy as well.

I just have a few questions, maybe they'll result in feedback. - Is the set of criteria fixed or generated ad-hoc by an LLM with some guidance? - Does an LLM decide if a criteria should be evaluated for the given codebase or is that done via heuristic? I see that some criteria are skipped in the eval, while others that don't apply to the project are evaluated E.g. most of the ones here marked as x can't really apply to the pi-mono codebase, but are still evaluated.

any open source repos you'd like us to run agent readiness reports on? will do every one for the next ~24h

@badlogicgames @matanSF We're iterating fast! Appreciate any feedback you might have!

@matanSF Some of these are debatable, I think. Regardless, super helpful.

@travisennis It isn't! Hill climbing your readiness score will help any agent succeed in your code base 😄 None of the checks are factory specific

I’d like to see this not tied to a single agent/product.

@_coenen Yea totally, I guess this is also hard because "good people" could just be bad people that other bad people feel good about 😄 Spiritually i agree with your take tho

@hurrythas I mean generally speaking people who build bad things for the world think they’re building good things for the world, so personal morality ain’t that great of a gate… But “fun” could probably be expanded to “is it something I enjoy and feel good about?”

The only questions that matter when choosing what to work on: - is it fun? - is it challenging? - is it with good people?

my competitive advantage is i'm having fun

Chainsaw Man

In the episode of The Simpsons where Homer adopts a pet lobster it is called Mr Pinchy. In the European French version of The Simpsons, Mr Pinchy is instead called Homard Simpson (homard means lobster in French)

Quite pleased it seems the players have now realised Mamardashvili is left-footed.

I don’t know what Thibaut Courtois and Ismailia Sarr have against Liverpool Football Club.