198 karma · joined December 4, 2022
if they are able to detect hallucinations, filter them out and automatically re-run, that's a huge improvement in result, even though core model didn't get new training
2004 - The Google team realizes they're sitting on a pile of user data bigger than the Eiffel tower and starts the development of MapReduce, because finding pictures of cats among a petabyte of data is harder than finding a needle in a haystack.
2005 - David Heinemeier Hansson takes a break from blogging about how much money he has to create Ruby on Rails, a framework designed for developers who want to turn coffee into web apps. A vibrant community of hipsters who can't code in PHP quickly forms around it.
2006 - Guido van Rossum wakes up from a fever dream and decides that Python needs to be harder to write. Thus, Python 3 is born. Many developers remain happily in Python 2, ignoring the encroaching darkness.
2007 - Amazon sees that people are having a hard time storing and retrieving their data, so it introduces Amazon Web Services (AWS) to the world. Developers around the globe rejoice, then promptly get lost in the labyrinthine AWS Management Console.
2009 - Ryan Dahl combines JavaScript, the most misunderstood language, with server development, the most misunderstood part of software development, and creates Node.js. Programmers are finally given the opportunity to write callback hell on the server side.
2010 - Instagram is created by Kevin Systrom and Mike Krieger, offering filters to make your bad photos look vintage. It's quickly discovered that the Valencia filter is the most efficient way to generate envy in your followers.
2011 - After creating C++, Bjarne Stroustrup decides that it isn't complex enough. He proposes C++11, an update that adds enough features to make any sane developer's head spin.
2012 - Facebook, faced with maintaining the world's most visited website, accidentally creates React, a JavaScript framework designed to solve problems no one knew existed. It quickly becomes popular among developers who are tired of enjoying their job.
2013 - John-David Dalton releases Lodash, a library that brings sanity to JavaScript developers by giving them the ability to write code that can actually be understood by humans.
2014 - Apple, tired of Objective-C's quirky syntax and all the fun developers are having with memory management, announces Swift. It's hailed as "Objective-C without the C," though it soon becomes clear that it's also "Objective-C without the fun."
2015 - A group of developers decides JavaScript hasn't been overused enough and creates Electron, a framework for building desktop apps with web technologies. The world rejoices as their RAM is consumed by hundreds of Slack threads.
2016 - Google invents TensorFlow, its AI and machine learning framework. In a surprise twist, nobody understands it. Still, cat identification accuracy skyrockets.
2017 - The Bitcoin bubble inflates to its maximum size. Suddenly, everyone and their grandmother becomes a cryptocurrency expert. Developers use their GPUs for mining rather than gaming, sparking widespread panic in the gaming community.
2018 - The world finally admits that managing state in JavaScript apps is horrifying. Dan Abramov gives us Redux, so we can be horrified about state management in a structured and predictable way.
2019 - TypeScript becomes mainstream, as JavaScript developers realize that sometimes it's nice to know if your function is going to explode before you run it.
2020 - COVID-19 hits, and developers around the world discover they can work from home, leading to a surge of productivity and innovation, until they realize Netflix has all of Friends.
2021 - Rust, a language developed by Mozilla, has its time in the sun as developers scramble to write more efficient and less buggy code. Meanwhile, a group of committed enthusiasts prepares a secret plot to rewrite everything in Rust, including this text.
reality is pretty different, fb business is doing great despite all their missteps
Have you seen this done anywhere?
Priority inversion: A situation where a higher-priority task is indirectly blocked by a lower-priority task holding a shared resource.
Mutex vs semaphore: Mutex ensures mutual exclusion for a shared resource, while semaphore controls access to a resource by multiple tasks with a counter.
Preemption: The act of interrupting and temporarily suspending a task, allowing another task to execute.
Interrupts: Signals to a CPU to temporarily stop its current task to handle an event or perform a specific function.
Interrupt latency: The time between the arrival of an interrupt and the start of the interrupt service routine (ISR). ISR vs function: ISR handles an interrupt, cannot be called directly, and must complete quickly; a function is a reusable block of code that can be called as needed.
Interrupt vs exception: Interrupts are external events requiring CPU attention, while exceptions are internal events caused by the execution of an instruction.
Hard vs soft real time: Hard real-time systems have strict deadlines that must be met, while soft real-time systems have more flexible deadlines.
CPU boot process: Initialization sequence a CPU follows upon startup, including loading firmware, running tests, and loading an operating system.
Board bring up: Process of validating and configuring new hardware to ensure correct functionality.
Memory sections: Code (executable instructions), Data (initialized variables), BSS (uninitialized variables).
TLB: Translation Lookaside Buffer, a cache for memory address translations in virtual memory systems.
Big vs little endian: Big endian stores the most significant byte first, while little endian stores the least significant byte first. 32 vs 64 bit processors: 32-bit processors have 32-bit wide registers and address spaces, while 64-bit processors have 64-bit wide registers and address spaces, allowing for larger memory and better performance.
Null pointer dereference: Undefined behavior occurs, often leading to a crash or unpredictable results.
Daily metrics can vary wildly without affecting overall result.
if we assume that a computer can be powered by 100 watts, over a day it will use 2.4 kW h, which is about 2000 Calories
GPU will consume a lot more, but we aren’t that far off in efficiency
you just did another update?