1. Full Time
Total comment counts : 38
Summary
The writer has quit their job to work on Marginalia Search full-time. They express relief at the decision to commit to their passion project, and look forward to achieving a better balance in life. The writer acknowledges the uncertainty that comes with solely relying on their project for income but is excited by the opportunity to create something that is meaningful. They reflect on how they started the project to see how far they could take it, and express surprise at the fact that they are now doing it full-time.
Top 1 Comment Summary
The article congratulates Viktor, who has left his office job to work on his search engine project. The author shares their excitement and references previous discussions and articles about the project on Hacker News.
Top 2 Comment Summary
The author of the article is testing Marginalia Search and notes the low latency of the site. They speculate this may be due to lower traffic compared to other sites, even those with zero traffic. They also mention a desire for a front-end focused write-up of the Java sourcecode. A link to the sourcecode README is provided.
2. Reddit is removing moderators that protest by taking their communities private
Total comment counts : 123
Summary
The article is an error message from a website, informing the reader that they appear to be a bot and that there is a speed limit on requests to prevent abuse. The reader is instructed to wait six seconds and try again, or to change their browser’s user agent string if they are spoofing one. Developers are reminded to make no more than one request every two seconds to avoid this error message.
Top 1 Comment Summary
The author expresses sadness and frustration about Reddit’s recent decisions, including the promotion of an unpopular app and the implementation of “new” Reddit which is also unpopular. The author wonders who will continue to use the site and criticizes the website for antagonizing its most loyal users in exchange for profit. The author concludes by expressing disappointment in Reddit’s recent actions and the negative news coming out of San Francisco.
Top 2 Comment Summary
The article discusses how individuals like Kanye West, Mark Zuckerberg, and Reddit’s Steve Huffman have fallen into the trap of trying to prove that they are special due to receiving praise and being told they are brilliant and important. Huffman, in particular, is mentioned as having been in the right place at the right time with a LAMP stack when Digg failed and now wants to show that he is special, rather than attribute his success to timing and circumstance.
3. Software effort estimation is mostly fake research (2021)
Total comment counts : 40
Summary
This is not an article, but an error message stating that the requested resource could not be found on the server and was generated by Mod_Security.
Top 1 Comment Summary
The article discusses a method for estimating a job’s duration in the field of software development. This method involves taking a guess, doubling it, doubling it again for stakeholders who are not genuine, adding a percentage for developers with less experience, subtracting a percentage for features that will be copied, adding a percentage for sick leave/time spent browsing, doubling it for unresolved questions, and dividing by the room temperature multiplied by the number of people using mechanical keyboards. The author suggests that this method provides a relatively accurate estimate, at least until the next sprint.
Top 2 Comment Summary
The article proposes that there are two types of engineers - good ones and “shit” ones. Good engineers value autonomy, while “shit” ones need more help. The author suggests that scrum masters, who are often former “shit” engineers, tend to micromanage teams with agile methodologies and do not add value to the process. Instead, teams need a mix of good and “shit” engineers, rough project estimates, and the ability to ask for help when needed. Good engineers do not need to be questioned every morning as if they are children.
4. Localrf – Nerf from casual shaky videos
Total comment counts : 16
Summary
The article describes a new algorithm for reconstructing the radiance field of a large-scale scene from a single captured video. This task is challenging for two reasons: accurate pre-estimated camera poses are usually required, which are frequently difficult to obtain from Structure-from-Motion algorithms for in-the-wild videos, and using a single, global radiance field does not scale well to larger scenes. The proposed algorithm addresses these challenges by jointly estimating camera poses with the radiance field in a progressive manner and dynamically allocating new local radiance fields trained with frames within a temporal window. The algorithm’s evaluation on datasets demonstrates its effectiveness and compares favorably with state-of-the-art approaches.
Top 1 Comment Summary
The author’s grandmother’s farm had to be sold after she died in 2012. The farm had sentimental value to the author because despite moving around as a kid, they always visited there for holidays, and it felt like home. The author has extensive videos they recorded in 2006 of the farm and is considering using them to reconstruct the farm. However, the author questions whether they should do it as they have not watched the videos in years, and they wonder who wants to remember something they can’t go back to.
Top 2 Comment Summary
The article provides information about Neural Radiance Fields (NeRF), a technique used in generating synthetic data. It includes links to an external website with a guide on NeRF and to a personal website with further information on the topic.
5. Diablo Pitch Document (1994) [pdf]
Total comment counts : 28
Summary
error
Top 1 Comment Summary
The article recounts a story about Diablo’s development. David Brevik, the game’s creator, initially wanted a pure turn-based and permadeath game like Rogue or Moria, until he implemented real-time movement in an afternoon and attacked a skeleton, creating what would become the ARPG genre.
Top 2 Comment Summary
The article states that the original version of the game Diablo took 11 people and one year to develop, while Diablo 4 has taken more than 300 people and over six years to develop, which is over 150 times more expensive than the original. The article emphasizes that the video game industry has changed dramatically.
6. Metformin shown to prevent long Covid
Total comment counts : 23
Summary
Metformin, a commonly used drug to manage type 2 diabetes, may prevent long COVID, according to a study published in The Lancet Infectious Diseases. The COVID-OUT study examined the antiviral properties of three readily available drugs: ivermectin, fluvoxamine, and metformin, for both COVID treatment and long-COVID prevention. After more than two years, metformin was the only medical intervention in the study shown to prevent long COVID. The phase 3 trial included 1,200 adult participants who had never had COVID-19 before. The trial found that metformin reduced the incidence of long COVID by about 41%.
Top 1 Comment Summary
A study found that taking a 14-day course of metformin reduced the incidence of long COVID by 39.4% compared to a placebo group. By day 300 of follow-up, 8.3% of participants reported receiving a diagnosis of long COVID. Half of the participants were unvaccinated, and the researchers noted that metformin appears to have reduced the viral load greatly, leading to speculation that the reduction in viral load may be related to the lower incidence of long COVID. Further research is needed to determine if the benefits of metformin are equal among vaccinated and unvaccinated patients.
Top 2 Comment Summary
The article refers to a men’s mental health talking group facilitated by the writer, where some attendees who are anti-vaccination continue to resist popular science and instead trust their own research. These individuals are now assigning blame to almost any other person’s illnesses as caused by “the vaccine”. The article also mentions one attendee’s unscientific belief that the banks will collapse on June 27th. It’s worth stating that the writer disagrees with these individuals and is themselves vaccinated and boosted.
7. Generating Income from Open Source
Total comment counts : 19
Summary
The author discusses the struggle for open source projects to generate revenue and suggests that individual donations are not enough for maintainers to make a living. They propose that instead of relying on donations, open source maintainers should charge companies for the value their projects provide, treating it as a business and selling a useful product. The article provides examples of successful open source businesses and their approaches to generating sustainable income, such as offering different licenses and subscription plans. The author emphasizes the importance of understanding the value that open source projects deliver to companies and charging accordingly.
Top 1 Comment Summary
The author argues that open source software has become a catchphrase and people use it to seem “cool,” but it has lost its original meaning. They believe that for-profit open source projects should have a clause obligating them to donate a percentage of profits to the open-source projects on which they rely. The author suggests that supporting core open-source projects should be a priority since they have enormous importance for millions of people. Additionally, the author points out that many companies should stick to being “for profit,” and it’s not bad to make money off of open source software as long as it doesn’t conflict with the original intention of open source.
Top 2 Comment Summary
The author mentions a new tool, called Polar, which aims to help maintainers get more funding and insights into their customer’s needs. They agree that while companies need to be converted more, the traditional sponsorship/donations model is not enough and investments need to be quantified. The author hopes that the first version of Polar will provide better insights on needs and funding for maintainers to align with their vision.
8. Illinois prohibits weapons, facial recognition on police drones
Total comment counts : 41
Summary
error
Top 1 Comment Summary
The article advocates for a federal bill to prohibit the use of armed drones by law enforcement to avoid a patchwork of state and local restrictions. The author identifies possible irresponsible arguments that may be used to oppose the bill such as the need for law enforcement to have armed drones to combat criminals and that deadly force must be used to protect officers. The article also suggests that drone operators have no fear and are thus more likely to use weapons prematurely.
Top 2 Comment Summary
The author is in favor of using CCTV cameras with facial recognition technology as a way to combat crime without having to resort to potentially dangerous interactions with the police. They acknowledge that there are concerns about potential abuses but believe that as they have no privacy in public spaces, they are prepared to accept this technology. Additionally, they believe that it is cheaper than hiring more police officers and suggest that laws and secure systems be put in place to protect the data collected. They draw the line at using lethal weapons on drones but suggest that riot control drones may have a valid purpose.
9. Louis Rossmann calls community to leave Reddit
Total comment counts : 30
Summary
error
Top 1 Comment Summary
The author agrees with Louis, who says that the Reddit management doesn’t care about users but only thinks about servers and API. The author has encountered many people in tech who openly state that they don’t care about users or employees. The author says that it’s impossible to change the Reddit management’s mind, just like it was impossible to change the minds of the people they’ve encountered in tech.
Top 2 Comment Summary
The article suggests that Reddit has not behaved in a way that makes business sense, due to VC money distorting its business model. Reddit is now willing to potentially harm itself to prevent new players from accessing data that big tech companies have already used to train their models. The author thinks this is a repugnant way of dealing with users and indicates that a business that did not generate enough value to be sustainable can be classified as a scam.
10. Annoying A/B testing mistakes
Total comment counts : 25
Summary
The article explains common mistakes to avoid while conducting A/B experiments. These mistakes include including users who aren’t affected by the change being tested, not breaking down results into subgroups, not deciding the experiment duration beforehand, not testing with a small rollout first, not monitoring counter-metrics, not accounting for seasonal changes, and poorly defined hypotheses. The article also highlights the importance of measuring the right metrics and not relying solely on metrics to determine the success of an experiment. The author shares an example of how a change that didn’t result in a metric increase was still shipped due to improved user experience.
Top 1 Comment Summary
The article discusses a disagreement with the examples provided for testing unclear hypotheses. While agreeing with the overall point, the author finds the examples lacking in precision and clarity. They specifically criticize the use of a problem statement instead of a hypothesis, as well as the introduction of multiple hypotheses within the same statement.
Top 2 Comment Summary
The article explains that determining the appropriate sample size for a trial is not an easy task, particularly without prior knowledge of effect sizes. The author recommends using a sequential testing framework and optional stopping instead of waiting for a fixed sample size. However, standard t-tests do not allow for optional stopping since Type I and II errors are only valid at a determined sample size. The author suggests safe anytime-valid statistics or Bayesian testing as alternatives. The article also provides links to resources for these suggested alternatives.