I often forward this article to my fresh software engineers but they often get put off by the layout. So I tell them beforehand not to pay too much attention to it =)
Feels like there is a similarity in a big ball of software mud and complaining about the layout of a website. Does it do the job? Yes? Cool, moving on.
This one and https://www.mindprod.com/jgloss/unmain.html should be required reading for all new programmers. Although sadly they can’t really appreciate them until actually subjected to them.
Can someone recommend an app for ipad that can read PDFs? I want to be able to bookmark using my browser but read it on my ipad. Sort of like "Save to pocket" extension.
Primarily I've been using Zotero and Notability. They each have "save to" on mobile. Zotero has a chrome plugin that requires the desktop app to be running. They both optionally support a dark mode for reading in the dark.
I like the experience of reading in Muse.app on the iPad. It's a nested whiteboarding thing, but also can act as a PDF reader. (It'll let you pull out chunks of the PDF and put it on your canvas with a link back into your document, if that fits your flow.) I often read on my phone, so this is not an option for me.
Apple Notes and Muse slow down with a lot of ink. For taking a page full of notes I'm using Notability.
I've heard good things about GoodReader, but haven't played with it in years.
I use Readdle Documents to sync PDF folders with my server PC via FTP. Free version supports PDF highlighting & simple annotations, basic file management, and automatically syncs back everything.
In theory the built in files app will work for this. However, I like goodnotes, which has good highlighting snd library support. I’ve used it since grad school for reading papers.
>Here is how the S174 change impacted some companies, based on what I found in their annual reports:
>Microsoft: $4.8B additional tax paid in 2023. The company generated a $72B profit that year, so this tax increase was manageable. It’s still a very large amount!
>Netflix: around $368M in additional tax paid – also manageable with $4.4B annual profit.
>Google: the tax change was minimal, because Google was voluntarily amortizing software development expenses for most staff, already. This was for all projects that reached “technological feasibility,” which is a milestone products pass before public release.
For finding answers, Stack Overflow may be useful, but for posting questions, it is not. I ceased posting questions a while ago due to the aggressive moderation, which often labeled my questions as duplicates or too common. I was very upset with the direction Stack Overflow had taken, as it seemed to embody the very criticisms it had once had about forums. The Stack Overflow community has lost its vibrancy and has become stagnant. I am pleased that platforms like ChatGPT are emerging as real competitors.
If they didn't moderate aggressively, they would probably end up with a bunch of duplicate questions. I really like the fact there's not a ton of dups on SO. ie, see other help sites. They also have a high standard for those questions, to the point where it will take 10+ minutes to post a good question. I don't ask questions either, mainly because I don't want to spend 30+ minutes writing a good question AND that's a good thing.
IMO, the aggressive de-duplication of questions means that people will keep updating the answers as things change, which is really nice as you don't have to worry about the responses to a question becoming out-of-date/stale.
I don't think people object to de-duping true duplicates. The problem is near duplicates, where the original question/answer didn't quite cover what you're asking. In those cases, it can be frustrating getting past the SO duplicate filter, which can sometimes be too aggressive.
Or, when searching for answers, you keep chasing your tail because the mods kept killing off the variants of the question you cared about, and you have to cobble together answers from the comments because those will often be all you get.
Or (and arguably worse), you'll find a relevant question but the answers are inundated with "like the real answer, but variation for situation X" because people are posting to the main question because they know the almost-but-actually-not dupes will all get closed out.
There still are a ton of duplicate questions. When questions get marked as duplicate they don't get merged or anything, they just lock the most recent ones (even if they are much better than the older one).
Their moderation doesn't prevent duplicate questions it just makes it more frustrating when the question you find has been marked as duplicate.
It can still happen. I've been actively using SO for 14 years and had my first question marked as a duplicate last week. The question was already 9 years old and had been thoroughly answered. Last week somebody decided to "optimize" it, which sent it to the review queue and finally led its closing.
It does seem likely that I'm an outlier . OTOH it could be that just people that have had bad experiences are posting their experiences and the silent majority without bad experiences aren't posting. It's so hard to tell in a forum format.
I think the loudest people are the ones talking about their bad experience. I've been on SO for, I think, 13 years. I've rarely had questions closed as duplicates and if one was it truly was a duplicate and I found my answer on the referenced question.
I don't see how all the people complaining about moderation don't look at all the low quality questions and see why moderation is what it is.
Personally I don't remember having gotten questions closed. However,
sometimes other people's good questions have gotten closed, making me annoyed that I didn't get to see any, or more, answers.
> For finding answers, Stack Overflow may be useful, but for posting questions, it is not.
That is mostly as it should be. The primary original intent of StackOverflow is to curate a good Q&A, not to provide an interactive support/mentoring service. Questions are (supposed to be) evaluated by what they add to the repository of questions and answers.
That being said - there has certainly been a culture of harshness in relating to newbie question askers on large parts of the SO.
Suppose you are just learning C++ language. There is a syntax error or type mismatch. The error message is 50 lines long and it does not make any sense to you. You already spent 10 minutes on it. What do you do now?
If there is an experienced C++ developer around who is happy to answer questions, great. Otherwise you are stuck with having to figure it out by using trial and errors and Google-fu. If you post it to a forum, it could take a while for anyone to respond. If you post it to stackoverflow, very likely nobody wants to look at your horrible code (natural for a beginner, you know), and your question gets downvoted.
By contrast, ChatGPT can look at your code and explain very clearly what is wrong with it within seconds.
This is just one example. And it's not only for beginners. I have found that ChatGPT can answer high-level questions in programming very well. The alternative would be searching for the Internet and sift through all the noise to find the answers as well.
Isn’t it worth considering that the reason ChatGPT can do those things is that it was trained on data from platforms like Stack Overflow? This is hard to quantify but my guess would be that without the SO data it wouldn’t be as useful.
This is why reddit and twitter both locked down their APIs, due to the data being very high quality and immensely valuable for training.
Too bad they all did it too late, no one saw ChatGPT coming. And since all the data was already scraped from Stackoverflow, it no longer has any value for OpenAI. Stackoverflow is rapidly declining in volume, so future data from it is irrelevant.
Oh wow that's interesting, thanks for sharing! I thought that the value in today's LLMs comes from the wealth of "free" crowd-sourced answers online. It seems like OpenAI is trying to create more source material specifically for "basic coding", though I wonder how cost-effective that is in practice. Or effective in general, for that matter.
Exactly. I've already hit the dreaded "knowledge cutoff" of 2021.
What happens when this training data is 10 years old? Where will we get new data?
In the meantime we'll get used to functioning with these models with assumption the "knowledge cutoff" will just be moved. But if there is little new data, how? Furthermore, how do we prevent feeding ML generated data to ML training in the first place?
