Win10 iot ent runs on even 3rd gen i3s very well...
HN user
winrid
I manage databases "at scale", and also run some services:
fastcomments.com sidewaysdata.com watch.ly dirtforever.net govscent.org
https://github.com/winrid
Chat with me: winrid aaattt gmail
Your computer is fast and can display hundreds of thousands of divs a second. :)
I've been building watch.ly[0] which intercepts all network /fs traffic on the machine originating from the agent and pings you for approval.
[0] https://watch.ly
It's embedded. Single machine. But it runs their forum/ticketing thing so good enough for that I guess :)
Probably worth saying it's a fork of orientdb from like 2010.
I'm not sure I'd say "React Bad", I'd rather say "React Not Really Super Necessary and More Often Than Not Leads to a Worst Experience at Maybe a Slightly Lower Initial Cost" :)
I did an experiment recently, I actually wanted to launch a dashboard and picked React because I knew I could throw React + Tailwind + popular_component_and_charts_frameworks at Opus and get a decent result. But actually, I realized a lot of fairly simple button clicks had like a 100ms-200ms delay, and I just did not like it. I ported it to Java/Javalin with J2HTML and kept tailwind, and it's very snappy and works great. It's currently 96k lines of java ^_^
To write good React you have to follow conventions, too. You can't get away from conventions.
People keep touting react etc but I swear every mobile ordering app I use lags like hell, just from recent example, and these can easily be built and maintained without a huge framework to "manage state" or "connect state to the view".
you just agree on a set of conventions for components and state management, it's not that hard. This is basically how Rails works...
yes it's still a framework, duh. it's just less abstraction than react etc.
Excellent! We can use this to start to throw executives in jail and maybe society won't erode as fast.
My fun ancedotal story about US healthcare:
Kaiser transported me 4hrs in a ambulance to get surgery and I think I just paid the $200 ER fee
But then they would never cover an MRI for an issue with the same leg.
It's like we're almost done with this project guys, come on.
oh for sure! I think in some ways LLMs have raised the bar in terms of what you an expect from software (once we get past the hurdle of increased bugs, but that seems to be getting better)
Yes, because now your competitors do the same. The winners: inference providers
Pretty proud that after six years fastcomments is still only about 30kb
I should probably work again on getting that down some :)
post-ban I've even had Opus 4.8 say it cannot let me continue (I was analyzing a nodejs heap dump)
So basically if you have queries that are hard on the query planner, that constant fan out has higher CPU cost than the alternatives.
Today I tested GLM 5.2 by giving it an example stylesheet and told it to change the background color of a submit button.
It then hallucinated the submit button class...
Your ISP will eventually require a new protocol that can hold your digital ID in every packet to access certain content and you'll have to upgrade/migrate your OS if you want to view p0rn etc.
It's not about the kids.
It's so Facebook can prove to advertisers the views are not bots/AI.
Next we'll be putting smart people in jail for being too dangerous.
You mean what companies used to do before we decided to import cheap labor instead?
This is AI right?
Friendly reminder this is lobbied into existence by advertising companies since the advent of AI they aren't sure what is real traffic anymore, tying all traffic to a government ID solves that.
None of my LLM created rust code has unsafe, or unwrap and must pass clippy checks..
Python isn't even that bad, I have python desktop apps that use 50mb of memory. More than needed but not 1-2gb realm people complain about with electron stuff.
At the end of the day it's a battle of competence of who runs your company.
Sure you might be able to have PMs deal with the easy stuff for a short while.
But will the company that can hire more competent people pull ahead? Probably.
I can tell you one thing: I haven't worked with a PM for like 5 years.
Which does technically distribute wealth to thousands of people
I had a feeling this was all related to someone trying to run a database on networked disks.
every fucking time
The teeth ones "stretch" to grab different sizes, too, but they won't support as large a range as a chain or band one for sure.
It's possible it has been fixed in the last couple years but for a while it was the case.
Working on sidewaysdata.com a lot :)
Coming soon is a full automated training system to "certify" people to time events.
It'll spin up a VM with our timing software and an emulator of the hardware of your chosing and you kind of play a game to get certified (you deal with real radio traffic, real world like scenarios where shit goes wrong)... This way we don't have to train people AT events from zero.