I've been using Claude Code for weeks now and I've found it to be fantastic at writing / debugging tests unit / integ tests that don't require external context. Still needs some guidance of course but it's been a huge productivity improvement.
Unfortunately, I can't say the same for other types of tests like E2E tests. It makes sense why: CC doesn't have access to all of the context to determine what's going wrong. During an E2E test suite run it can't pause to look into the console, view whats on the page, look at the backend logs, etc. I tried writing some utilities to snapshot and log the state of parts of the page when a test failed and this did help a bit, but not enough to get the same productivity boost for other types of tests.
Has anyone had any luck with this? Any strategies to share?
If you find yourself bottlenecked on IO performance, OrbStack is the best. Also a fan of the fact that it has a very similar VM function to WSL2. Also supports using either Rosetta or Qemu for emulation of Intel architecture, and has some kind of Kubernetes integration that I didn't try.
Last I tried Podman Desktop, it was OK, but had worse performance than Docker and was a little rougher around the edges. Still, not bad.
Colima is probably fine, but I had issues that were showstoppers for me. Trying to get it to consistently start up correctly in a CI environment seemed oddly challenging, and it doesn't seem to support IPv6 at all. YMMV.
I did evals of most of the popular ones for my company and my opinion is OrbStack is best performance and most reliable. The Kubernetes integration is also very nice and resetting it takes about ~10s in case you get in a pickle.
OrbStack has been great. I'm not affiliated with them but a happy customer. They've had issues where an update broke them in very _weird_ ways, but I've been very impressed with their support engineers on this.
I've even had a situation where the support engineer jumped on a call with me to see the repro as it was a very specific/hard to share a simple repro situation.
Big fan of colima. It's pretty fast (in some cases not on parity with Orbstack), but it's free and doesn't have annoying Electron GUI or any crap like that.
Orbstack would be my next choice, but costs money for comercial user and my employer is a tight arse.
Tangential question - what are people using their homelab for / what are some interesting or useful projects you've spun up on them? I've been thinking about setting one up but not 100% sure I'd find use out of it :)
At the risk of leaking info to any toes I step on w.r.t my home environment,
- home assistant
- Network Video Recorder
- Jellyfin
- network management such as ubiquiti or omada etc
- vault warden/1pass/other secrets servers
- tailscale or wire guard server
- build server/k8s test environment
- private artifactory or mirror (especially useful if you're using the same distro on a bunch of devices but don't want to overload the actual mirror+improves download times)
- torrents (someone's gotta seed Wikipedia)
- onsite backups
- bastion into your home network (see also: wire guard)
- some people even use it for their router
You could also take a look at the tteck scripts, there's a bunch of cool stuff in there
I'm just going to repeat a bunch of what hughesjj has already said, but anyway:
OPNSense (as my household's internet interface), Unifi Controller (as my household's primary wifi), Jellyfin, Wireguard, Pi-hole, LMS[0], Frigate NVR (migrating off ZoneMinder, awaiting delivery of a Coral TPU to finalise this), couchdb (as Noteself[1] back-end), nginx (serving a handful of sites for my own entertainment), Mailu[2], Calibre[3], various other in-flight experiments (which Home Assistant will soon become, Bitmagnet DHT scraper).
Most of the above are docker instances hosted on a small number of VMs hosted on two (or sometimes three) physical machines running proxmox.
I have Proxmox on my older (11th gen) Intel NUC - it runs my Unifi network controller for my WiFi APs, an Unbound DNS cache (because my router doesn't support DNS over TLS or DNS over HTTPS), a NAS (using ZFS on some hard drives in a USB 3.2 Gen 2, 6-drive disk chassis - the ZFS managed directly by Proxmox and then a Debian container just doing Samba).
That's all for now but I've just installed Home Assistant but haven't set that up yet. I also intend to try out Jellyfin as a media server and Frigate as a video recorder when I get some cheap cameras.
I started with a home lab a few months back. Its basically an old miniPC running Proxmox virtualization and LCX containers. Mostly it helps me learn various technology I have not much experience with in my work. I run about 20+ services on the home lab. Some of them include:
Open WebUI which can connect to the OpenAI api or a local Ollama LLM. You can also connect various tools to the LLM like a calculator or web search to augment them. The AI has helped me learn how to configure and debug stuff. Like I got step-ca to setup a local certificate authority and give certificates to my various internal services. I played around with configuring Caddy and Nginx along with ACME to the the step-ca. The LLM was even helping be debug my config files.
I'm also using Hoarder for bookmarking and it can use AI to automatically tag your bookmarks. It can even backup the webpages.
I've been using Mealie to clip and save online recipes.
I'm running Uptime Kuma to check if my computers and services are up and if they are down, I'll get a notification.
- home assistant (just a few currently, more soon)
- paperless:absolutely awesome document management system
- immich: image management with automatic synchro of my mobile taken images (ML features)
- tailscale
- StirlingPDF: simple tools for all things PDF
Working on https://www.brief.news - a completely personalized daily newsletter on the topics you're interested in. We've just launched the ability to add custom topics, so you can create a newsletter on anything now!
The nice (ish) thing about Apple is that since they price everything stupidly high, I have some faith they're making a profit.
I don't think Alexa makes money. I'm pretty sure the hardware is sold at a loss, with the intention it gets you to buy shit through Amazon with it. Which - obviously - no. Nobody is shopping via voice, that was always a stupid idea.
I tried the same but I have multiple people in my house so about 20% of the time when I say “turn off the lights” it responds with “who is talking” and goes through a voice calibration exercise.
It’s super infuriating because I guess they want to personalize to the speaker, even when it doesn’t matter. And I suppose I sound similarly to others in my household.
We’re working toward exactly this with https://www.brief.news! We’ve got the ability to create custom topics in beta (I’m currently using it to follow Starship updates), happy to add you to the beta just shoot me an email (in profile).
We are actively looking for feedback on what SDK to develop next. Quite a few people have voiced interest in Python so far. This will make it more likely that we might tackle this soonish. We'll keep you posted.
Upnext | 100% Remote | Full Time / Contract | Software Eng / Design / ML
At Upnext, we are passionate about solving information overload. Every day we get bombarded with content from social networks, news sites, blogs, messages, etc. It’s hard to keep up and it’s even harder to find the content that really matters to you. It takes time and energy to sift through the noise and find what really matters. Our latest app helps you stay up to date on the topics and news that you care about by aggregating updates into a single place. Using our own AI models we’re building in deep personalization from the beginning so our users will always have the most important updates about topics they care about. We have open roles for:
I'm building https://www.brief.news, an AI powered newsletter that condenses tens of thousands of news articles into a daily briefing of the top stories, we support 30 topics today and are adding the ability to add your own!
Stack is a combination of TypeScript (Next / Node) + Python with a pretty simple deployment setup right now (GHA -> Container -> Cloud Run).
A lot! We’re actively reducing that though by training our own specialized models. We’re seeing equal or better performance with curated datasets at > 10x cost reduction.
This is pretty cool. Just a heads up: there's a french newsletter company I was subscribed to that is using brief.eco, brief.me and brief.science. Ironically, their main selling point is summarized news but by humans.
This looks awesome - might I suggest splitting the headlines on the homepage into a punchy title and subtitle? The wordiness of them makes it difficult for me to parse them for the topic quickly
Thanks! We got a couple different formats available, check out the top stories format which is close to what you suggest. Would love to hear your thoughts on it! We’re considering making that the default.
Unfortunately, I can't say the same for other types of tests like E2E tests. It makes sense why: CC doesn't have access to all of the context to determine what's going wrong. During an E2E test suite run it can't pause to look into the console, view whats on the page, look at the backend logs, etc. I tried writing some utilities to snapshot and log the state of parts of the page when a test failed and this did help a bit, but not enough to get the same productivity boost for other types of tests.
Has anyone had any luck with this? Any strategies to share?