Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Writing toy software is a joy

blog.jsbarretto.com

Post reply on HN
Print view View original

91–100 of 318 posts

Re: Writing toy software is a joy

#91

Post by sailorganymede »

dakiol wrote: post #39

> Perhaps you’re a user of LLMs. I get it, they’re neat tools. They’re useful for certain kinds of learning. But I might suggest resisting the temptation to use them for projects like this. Knowledge is not supposed to be fed to you on a plate Am I the only one using LLMs as if they were a search engine? So before LLMs I was searching on Google things like "pros cons mysql mongodb". I would read the official document…

Nope, I do this too (most of the time.) I don’t like working on code I don’t understand. I have started to ask it to use a client API I’ve written to figure out how clients would work with the stuff I write though. It’s great.

Re: Writing toy software is a joy

#92

Post by captainbland »

dakiol wrote: post #39

> Perhaps you’re a user of LLMs. I get it, they’re neat tools. They’re useful for certain kinds of learning. But I might suggest resisting the temptation to use them for projects like this. Knowledge is not supposed to be fed to you on a plate Am I the only one using LLMs as if they were a search engine? So before LLMs I was searching on Google things like "pros cons mysql mongodb". I would read the official document…

> Am I the only one using LLMs as if they were a search engine?

I quite like the search-first LLMs like Gemini and Copilot for this reason. They give you links which you can use to verify the output and seem to be less prone to directing you to SEO spam than Google Search/Bing.

Although I do partly think that search engines today are delivering poor performance compared to what they delivered historically so LLMs are benefiting from that.

Re: Writing toy software is a joy

#93

Post by eliasdorneles »

I quite enjoy the spirit of the article, and I believe joy in programming has become even more important in the AI agent coding age we're leaving.

However, am I the only one finding those time estimates way too short?

I'm not the fastest programmer on Earth, but I'm not the slowest either, and I think most of those projects would take me a lot more time than those estimates, specially if I'd be working only 2-3 hours per day.

I feel most of those projects would take significant time researching and learning about the subjects, before even starting to code.

Example: recently I replaced my Pelican blog by my own hacky static site generator written in Odin, working 2-3h per day, it took me 2 weeks -- and this is a simpler project than many on that list.

Re: Writing toy software is a joy

#94

Post by didip »

FirmwareBurner wrote: post #43
didip wrote: post #20

Big time! Almost all of my interesting knowledge came from toy projects I made to solve my own problems (or zero problem solved but it was fun so I did it anyway). For example, I wanted to know if I can make 1 Kubernetes cluster span multi region, multi cloud. So I slapped TailScale for networking, replaced etcd with multi-region PostgreSQL Aurora, and span the Kubelets between my Raspberry Pi, Digital Ocean VMs, and…

>I wanted to know if I can make 1 Kubernetes cluster span multi region, multi cloud Can I ask what you had running on that cluster?

As of right now? Nothing important. If I want to use it day-to-day, then I have to care about SLA, then I have to be the sysadmin of that thing.

Not exactly fun :)

What's interesting is that the TailScale network allows the local daemon be exposed on the internet with small settings change. Pretty revolutionary for personal/home IoT projects.

Re: Writing toy software is a joy

#95

Post by Karrot_Kream »

hathawsh wrote: post #73
dwh452 wrote: post #38

What's sad is how difficult it is to write software today. In the old days your dad could buy a C64 and cobble together an application. It should be vastly easier to do the same kind of thing with vastly better building blocks today. Why can't some Grandma drag and drop some widgets and have a recipe manager with sharing features amongst her friends and family?

I think the main difficulty is deployment. Grandma wants that recipe manager to be available to her family 24x7. How can she deploy it easily for free or very low cost? If there were a modern Hypercard, I think the key to its success would be making deployment extremely simple, reliable, and safe.

There are platforms like Observable and Repl.It that just let you deploy code/data pretty quickly.

Re: Writing toy software is a joy

#96

Post by fragmede »

hathawsh wrote: post #73
dwh452 wrote: post #38

What's sad is how difficult it is to write software today. In the old days your dad could buy a C64 and cobble together an application. It should be vastly easier to do the same kind of thing with vastly better building blocks today. Why can't some Grandma drag and drop some widgets and have a recipe manager with sharing features amongst her friends and family?

I think the main difficulty is deployment. Grandma wants that recipe manager to be available to her family 24x7. How can she deploy it easily for free or very low cost? If there were a modern Hypercard, I think the key to its success would be making deployment extremely simple, reliable, and safe.

NextJS with Neon on Vercel has a capable free tier, and there's enough training data that LLMs are decent at it. If Grandma is that interested in building an app, I'm sure she'd love to spend a few hours with a grandkid to set things up and then being taught how to vibe code (and also how to call said grand kid for help).

Re: Writing toy software is a joy

#97

Post by alanbernstein »

I agree with the sentiment in theory, but the time pressures of adult life make this difficult.

> What I cannot create, I do not understand

I suspect Mr Feynman would have phrased this a bit differently if he had first-hand experience with the modern front-end technology treadmill.

Re: Writing toy software is a joy

#98

Post by fragmede »

dwh452 wrote: post #38

What's sad is how difficult it is to write software today. In the old days your dad could buy a C64 and cobble together an application. It should be vastly easier to do the same kind of thing with vastly better building blocks today. Why can't some Grandma drag and drop some widgets and have a recipe manager with sharing features amongst her friends and family?

How hard did you look? WordPress has a few recipe maker plugins if you didn't want to code anything. Just install one and password protect the whole thing, and then teach (and write instructions for) Grandma to use it.

In the age of powerful computers, you can use Hypercard on an emulated Mac, you can use any number of hypercard-clones out there. She can just use Google slides. etc.

Re: Writing toy software is a joy

#99

Post by simonw »

thrtythreeforty wrote: post #79
simonw wrote: post #29

Earlier quoted context omitted.

This is also a great way to stand out among other candidates in the resume-review stage of a hiring process. The number one question I have as a hiring manager is always "can this person actually write code". A GitHub profile with 2-3 projects with multiple commits to each one answers that question really well. I won't filter out people who don't have that - lots of great engineers have never published code before -…

The converse is actually a negative signal for me. If you list a GitHub profile on your resume, but that GitHub profile is a ghost town consisting exclusively of "test repo please ignore" and "freshman-year assignments I was required to fork from professor's repo," that tells me you don't have an above-average understanding of why you add something to your resume. If you don't have an impressive GitHub profile, that'…

It partly depends on the level I'm hiring for. A junior candidate with a GitHub profile full of obvious tutorial projects still helps me skip putting them through FizzBuzz. A senior candidate with a ghost town could reflect poorly on them.
Post reply on HN