Timpy

I really like Crash Course World History. I'm older than the target demographic for this resource to be sure, but it's a great high level overview and having this under my belt helped me feel a little more oriented in every other historical thing I've dabbled in since.

https://www.youtube.com/playlist?list=PLBDA2E52FB1EF80C9

I discovered Anki 12 years ago while living in Japan. I was trying my hardest and absolutely failing to remember any of the Japanese I was studying. Maybe I was due for a learning-style renaissance for myself and Anki was just the catalyst, but it really made a positive impact on my life. More than just memorizing kanji on AnkiDroid during my commute, I just started to believe I could learn anything. I was starting to take my coding hobby more seriously at the time and hacking on Anki was a big part of that too. Thanks for all the hard work Damien and David Allison. I'm so grateful for the software you've worked on.

A lot of training data was curated in Kenya[0]. I would imagine if LLM data was curated in Japan our LLMs would sound a lot like the authors of their most popular English text books. Maybe other common Japanese idioms would leak in to the training data, like "ね" or "でしょう", ChatGPT would say "Don't you agree?" at the end of every message.

[0] https://www.theverge.com/features/23764584/ai-artificial-int...

Yes but I think it's a good virtue to signal considering the circumstances. If they paid the ransom that would signal that ransoming this company works, incentivizing more ransoms. If they refuse to pay the ransom it might signal that they care more about money than they do integrity. Taking the financial hit of the ransom, but paying it to something that signals their values, is about the best move I can imagine.

This is cool, very much the hacker ethos. But I didn't see any evidence that it can make a phone call though?

I think this guide is nice, and having a variety of articles like this is great so everybody can look at the different ideas and find what's right for them.

I would urge people to consider going a little bit further than this guide, consider not using your phone as a reading device. Imagine deciding to sit down with a physical book, but keeping your phone nestled on the opposite page as you read. It would be a lot nicer to read without interruption, without being exposed to notifications at all times. Sure there are going to be use cases where the phone is more convenient, but I think sacrificing convenience is worth it.

This is something I want to see in the world. Do you have a public repo? I'm currently doing third party application development for the Yoto, and I've done a lot of hacking on MP3s. If you're open source I'd be interested in helping, or at the very least chatting about the project.

I keep feeling like there's a huge disconnect between the owners/CTOs/managers with how useful they think LLMs _are supposed to be_, vs the people working on the product and how useful they think LLMs _actually are_. The article describes Harper Reed as a "longtime programmer", so maybe he falsifies my theory? From Wikipedia:

>Harper Reed is an American entrepreneur

Ah, that's a more realistic indicator of his biases. Either there's some misunderstanding, or he's incorrect, or he's being dishonest; it's my job to make sure the code that I ship is correct.

This is totally orthogonal to the issue but I think the best fix possible is to block the YouTube home page. I have gained value from algorithm-curated feeds in the past but it's no longer a net positive in my life. I recommend checking out News Feed Eradicator[0], Distraction Free YouTube[1], and set up some extremely aggressive uBlock Origin rules.

[0] https://addons.mozilla.org/en-US/firefox/addon/news-feed-era...

[1] https://addons.mozilla.org/en-US/firefox/addon/df-youtube/

The models outlined in the white paper have a training step that uses reinforcement learning _without human feedback_. They're referring to this as "outcome-based RL". These models (DeepSeek-R1, OpenAI o1/o3, etc) rely on the "chain of thought" process to get a correct answer, then they summarize it so you don't have to read the entire chain of thought. DeepSeek-R1 shows the chain of thought and the answer, OpenAI hides the chain of thought and only shows the answer. The paper is measuring how often the summary conflicts with the chain of thought, which is something you wouldn't be able to see if you were using an OpenAI model. As another commenter pointed out, this kind of feels like a jab at OpenAI for hiding the chain of thought.

The "chain of thought" is still just a vector of tokens. RL (without-human-feedback) is capable of generating novel vectors that wouldn't align with anything in its training data. If you train them for too long with RL they eventually learn to game the reward mechanism and the outcome becomes useless. Letting the user see the entire vector of tokens (and not just the tokens that are tagged as summary) will prevent situations where an answer may look or feel right, but it used some nonsense along the way. The article and paper are not asserting that seeing all the tokens will give insight to the internal process of the LLM.

Just a friendly heads up, for the Japanese exercises the video starts just a bit too late and the word you're listening for is cut off. This might only be pertinent for languages where vocab words are appearing at the beginning of sentences, French and Spanish didn't have "gap words" at their sentence start for the few exercises I tried.

This is a cool app, I would have enjoyed this when I was grinding on Japanese back in the day.

"The concept of running shoes as we understand them today, specifically designed to improve running efficiency and comfort, did not emerge until the late 19th and early 20th centuries."

It's a neat idea imagining what running footwear was like throughout history. Was Pheidippides barefoot when he purportedly ran from Marathon to Athens? Did Indigenous Americans debate which style of moccasins was better for your long term foot health? Did warriors develop plantar fasciitis from poorly designed boots?

I love these easy-to-illustrate examples where a request seems simple, but adds a lot of trouble for the person implementing it. "It's just adding a new skin for their footwear, how hard could that be?"

Back in the Windows 7 days I installed Skype on my parents computer before moving abroad, their user experience was basically like receiving a phone call. Even though they weren't tech savvy we never had any issues. I would call them, and if they were home and near the computer, they could answer it and we'd be video chatting.

A year or so ago I found this to be impossible, there was no application for desktop that was as simple as receiving a phone call. My father has no smart phone. I sent him a zoom link via email but he couldn't log on to the family computer without getting blasted with UI updates, terms of service changes, "Do you want to use OneDrive?", "Here's what's new in Chrome", "Try asking Copilot anything!", etc. From his perspective the computer never worked the same way twice. I wish we had regulations that prevented buying out competition.

You're right, it's possible and common to be recorded without somebody pointing a phone camera at you. That doesn't negate my point, right now the norm is that in many situations you _can_ record to your personal device without informed consent but you shouldn't. That will change if these types of cameras are everywhere.

Right now people have to point their phone cameras at at the thing they're photographing or recording, it's a very clear visual signal to others and there are cultural norms for this behavior. If a person is doing this in a common tourist destination that's more acceptable than pointing your camera at somebody else's children without asking them first. Imagine how uncomfortable it would be for somebody to hold their phone at eye level and point it at you the entire time they're having a conversation with you, even if they say they're not recording or anything like that. Having a distaste for smart glasses is pretty consistent with the status quo.

I wasn't even thinking about paywalls, the first thing I did was check to see if cookie banners and "Sign in with Google" popups went away. There's so many user-unfriendly things that you constantly deal with, any amount of browsing is just a bad experience without putting up defenses like this.

> Ask any artist to explain how color works, and they’ll launch into a treatise about how the Three Primary Colors—red, blue, and yellow—form a color wheel

I don't think the author asked _any_ artists how color works before writing this line.

I didn't read that to mean a literal two minutes. Based on his style of writing that was quoted in the article and the nature of his mistake I don't think he's reporting an in-depth postmortem on the situation, although I didn't bother going to twitter to take a closer look. He probably didn't even realize his mistake within two minutes.

When I took my graduate level operating systems class we had to do a bunch of coursework in C. Beej's Guides are the reason I passed that class. Thank's Beej!

I think there's a fair argument to be made that computing is the most complex synthetic thing humans have ever conceived all on our own. If we're talking about making systems in the broadest sense, then any category of problems that we create systems for could conceptually be managed on paper and pencil. Let's say "number of things that could go wrong multiplied by the difficulty a layperson would have in fixing such a thing" is what we mean by complexity, a computer would be more complex than a filing cabinet every single time.

There are trade offs for the complexity though, and well managed complexity could disappear behind the interface of a computer. When this is done well it feels seamless, and when it's done poorly it's painful. So maybe I'm arguing the meaning of complexity isn't 1:1 with the meaning of complicated. At the end of the day "did moving this to a computer make it better?" is the question to answer, and a lot of times the answer is no. QR menus at restaurants is my favorite punching bag for this but any home appliance with bluetooth or wifi is an easy target.

For me bash is a weird exception to the rule "code is easier to write than it is read". I'm never doing anything crazy in a shell a script but it's easy to just ask ChatGPT to whip it up for me and I will read it to verify. Perhaps anybody who can't validate the output is just as likely to write an accidentally-destructive script the old fashioned way.

I agree that there's not much point in writing actual literal semantically correct code on a piece of paper, but I don't think that's at all the point of the article. The author talks about understanding the problem space, getting their ideas organized, and (if I may charitably interpret the article a bit) even go as far as scratching out some ideas of what the code might look like.

I noticed that too! This could be a fantastic learning tool, my music professor described how to crescendo into the tension and decrescendo into the release but being able to experiment with this tool could help you get an idea of what you want to do before your fingers have the right touch to do it.

It sounds like that's exactly what I was looking for. Thank you!

I really wish I could just put a native neovim or vim text editor inside VSCode and call it a day. I absolutely love the lightning fast experience of editing with vim. When I know exactly what I want to do, I feel like I just ripped a line of spice and piloted a trade vehicle across Frank Herbert's universe.

But a lot of my job is exploring a code base and trying to figure out what the hell is going on, and I have to concede that setting up vim to do this kind of broad exploration isn't that great. Being able to click around at random and juggle tabs in the UI instead of storing buffers in my head is nice. I know vim can solve all of this, but I just found myself gravitating towards VSCode when I'm crunched for time and can't be bothered with vim.

My career is in full stack web development but I program in C in my master's degree coursework and as a hobby. Every time I have to peel back a decade's worth of CSS to move a button on a webapp I daydream of moving to a career in C. Is the grass actually greener on the other side?

Unfortunately I know people who drink diet coke like it's water. If somebody ate bacon at that volume they wouldn't live long enough for cancer to get them.

I remember the first time I played Minecraft and I was in awe at how expansive the play world felt. Without thinking too much about it, I had the feeling that if I set off in any direction I would discover infinitely new things. After enough playtime I saw the repeating patterns and eventually it felt so small again.

To some extent I agree. I taught guitar while I was in undergrad for my music performance degree, and having a guitar in my lap and reviewing the absolute basics with kids was way better than making pizza or pouring beers to make a paycheck like the other guys in my studio were doing. It gave me an edge that I'm grateful for, and being able to practice between lessons was insanely valuable.

But I have to reiterate that for me it translated into unhappiness when I got to the playing/performing/writing/networking loop. Those between-lesson practice hours became directed at making more money and not making myself better in a way that fulfilled me. So I kept grinding on my programming career until I had the work/life balance that I needed and now I don't feel too drained to enjoy my passions when I get home. It sounds like we have different goals though, because my definition of "taking the music gig forward and advancing" is probably unconventional. I'm not telling anybody what to do, I'm just saying that I had the same experience the article describes and it was a great growing experience to learn from.