cehrlich

As a maintainer of a different library, I think there’s something here. A revised version of this tool that also gets fed the docs and asked to find inaccuracies could be great. Even if false positives and false negatives are let’s say 20% each, it would still be better than before as final decisions are made by a human.

Yeah, I don't think the 12" MacBook was anywhere near the most controversial, because it was always clear that it was a niche product. I had one and loved it. Those who it wasn't right for didn't buy it.

The 2016-2019 Pros on the other hand replaced a machine that was universally loved. there were entire podcasts dedicated to complaining about them for four years.

Not sure which version was more popular, but I bought a red/blue switch on launch day. And anecdotally I'd say I've seen more of those than the grey one over the years.

Macs used to have good font rendering on Low-DPI displays (I would say the best, but I suppose that’s a matter of opinion)

Then Mojave switched to Metal and removed subpixel AA, and now it’s the worst.

Thread from when it happened: https://news.ycombinator.com/item?id=17476873

- Raycast for clipboard history and window management (it uses Rectangle internally)

- TripMode for reducing data usage while tethering

- KeyboardCleanTool for disabling the keyboard when wiping it down

cracking everything wide open for 5 minutes doesn't have too bad of an effect on temperature in my experience, and gives you good air for a few hours.

CO2 Monitor to let me know when I need to let some new air into my office. Mine is a TFA Dostmann 31.5009, but any model with an NDIR sensor should be fine, there are many choices around 70-150 EUR.

Same but Astro on Vercel's free tier. If my traffic suddenly spikes 1000000x, my site might go down temporarily. That's ok.

Paying $400/year to get some static html based on markdown for a personal website seems insane to me.

I got hired from there about a year ago

I'd suggest to not put a bunch of 4chan memes in your product demos.

Shouldn't be too difficult. The main thing you'd need is a dictionary that's as close as possible to covering the entire language (ie a digital-first one rather than a digital representation of a paper dictionary), and at least one but ideally several frequency lists. Ideally in the JSON format that Yomichan uses, as that's what the script for generating the data currently expects, but if not then the data can always be massaged. Then you just need to replace the specifics in the app (hardcoded frequency list names etc) with the stuff for your language.

When language learners learn vocab, there are two main processes: 1. Use a premade list of the top x words. Pro: they are guaranteed to be common words. Con: Once you make it past 2000 or so, they might not show up depending on what sort of niche things you read/watch/talk about 2. Look up random words as you come across them, and learn those. Pro: These are words you really saw. Con: You don't know how common that word is, maybe this is the only time in your life that you'll see it

I made a web app that lets you note down words that you come across and might want to learn, and then generates a learning order of those words based on a variety of frequency lists, as well as linking offsite for sample sentences etc. It allowed me to pass the JLPT N1 with just 6k known words (people usually need 8-10k)

vocab.c-ehrlich.dev

The exciting thing to me is that if there is enough demand, someone can make a keyboard module without having to make a whole laptop.

Thinkpad-style? Sure! Ortholinear? Why not!

Given the title I was expecting some kind of surprise, but everything works exactly as you would expect.

According to many commenters in the linked thread, it's fixed in MacOS 13.2.1. Not great that it happened at all, but good to see that it was a software issue.

If Copilot and ChatGPT cost the same and I could only have one, I'd pick Copilot.

T3 stack (Next.js, tRPC, Prisma with MySQL, NextAuth, either Tailwind or a component library depending on the project), deployed to Vercel. Maybe swap Prisma for Kysely if cold starts become an issue.

Disclaimer: I maintain an init tool for this stack so I'm somewhat biased

At age 31 I was was working as a part-time university lecturer in a non-stem field plus constantly struggling to find any additional freelance work, couldn't really see a realistic path forward in that career. Had also just become a father, so being broke felt even worse than before. I spent about 18 months learning web dev (just free resources online), and successfully made the career change. Required a lot of sacrifices in that time period though as well as a bit of luck, and I don't think my path is representative of the average person who tries this.

CS50x and CS50web are fantastic. web is a bit outdated but it doesn't matter because the projects are so good.

I got a bit frustrated at YouTube dev videos being aimed primarily at entry-level developers, so I tried to record some more mid-level content. There is definitely room for improvement on the production values, but hopefully the content itself makes up for it. Any feedback appreciated :)

tRPC is awesome, but it will teach you about typescript wizardry and how to write a library, not how to write an application.

I need to use PHP at work sometimes. After working in it for a while I still don't love the language and would never pick it for a project myself, but it's overall decent a few quirks aside, and I think a lot of people have built up a habit of disparaging it without actually knowing much about it.

That being said, one issue I've found is that while I can use VSCode for every other language, using PHP feels terrible if I'm not using a JetBrains IDE. I tried installing the PHP extension pack, but it was still leagues behind. Is there any guide someone can recommend for making PHP a first-class language in VSCode, or am I stuck in phpStorm?

I use Framer on some projects and I love its declarative approach to creating nice animations. It allows someone like me who doesn't know much about the topic to create really nice experiences.

But one thing that has been confusing me is its performance. The same simple animation might be smooth 60fps 9 times out of 10, but chug significantly the 10th time, on the same device etc. Any advice on how to profile this?

This looks great, and I remember seeing Jalup a few years ago. I'm alright at the language now (N1 a year ago, 50+ books read, etc) and so deep into Anki at this point that I doubt I'll ever transition away, but I could see recommending this to someone who wants an easier solution.

Yeah, I lived in Austria and travel across Europe fairly often, the only time I buy bottled water is in the airport or when I accidentally leave my refillable bottle of tap water at home. The quality of the tap water is equal to or better than the bottled water.

How many stars were you at when this began to happen? I maintain a repo with 8k stars and so far we've not gotten a single PR that felt like it was in bad faith - even with the hacktober fest tag.

Yeah, I don't understand negative comments. I don't care if it uses the most state of the art whatever mapping, fact is I used to live on the square where the main part of the video is shot and I was shocked by just how close to reality everything looks. I think it's the most immersive environment I've seen in a game.

The human models aren't as convincing, but there's also dozens of them on screen at once, the developers are still bound by what a pc/console can do.

Of course it's _possible_. If you go deep enough, React is just vanilla JS after all. The question is how fast a team can build and maintain something with each method.

Yeah, I think people who pretend that modern FE is possible without React or another similar framework haven't built a modern frontend and dealt with all the demands for interactivity that users have nowadays and the complexity this brings regarding rendering logic, state, etc.

There's still a time and place for sending HTML and maybe a few tiny JS files down the wire, but for anything "application-like" it's just not feasible, unless your business is so big and performance matters so much that hundreds/thousands of engineering hours are a valid tradeoff to achieve it.

Regarding data loading and performance, there is a lot of exciting stuff in the near future - Remix loaders, Next.js RFC, React 'use' RFC, Quik, Astro, better bundling/build tools, etc.

Are salaries in the UK really that low? I live in the EU, my salary is similar to the OP and I just started working as a developer _this year_, with 5 years of a fairly unrelated career prior to that.