mwpmaybe

Interesting. I use `--greedy` for casks but not regular formulae.

So much of GNU et al is already in Homebrew that I wonder how close you are to being able to support Intel Macs using the Linux/amd64 branch (or backend or whatever you call it) of Homebrew or something close to it, possibly after installing a few base libraries and utilities from MacPorts.

ETA: I've got Homebrew 7 building packages from source on an Intel Mac running Monterey; nothing yet required from MacPorts and a pretty minimal patch to the installer. I'd probably want to add legacy-support from MacPorts and update the macOS build environment to add it as an extra library on Intel Macs if I were to continue.

That being said, I'm sure y'all talked about continuing to support Intel Macs as source-based and ultimately decided against it, so it's unlikely that this will be interesting to the team. But let me know if I'm wrong and I'll open a couple PRs for further discussion.

Ahh, interesting. I have an issue today where my GitHub agentic workflows keep blowing through the context window that my local models can support; it ends up spending most of its runtime compacting over and over again. I've done all the usual stuff to slim down the fixed context and prevent it from slurping up too many files but it's still a recurring issue. Now I'm wondering if I can put engrim in the gh-aw container (or in the DinD sidecar it uses on ARC) and teach the headless claude harness how to use it, possibly even: "keep persistent context in engrim and use `/clear` instead of `/compact`". I'll play around with it. Thank you!

I hope you don't mind me asking you some ELI5 questions because I'd like to understand how this fits in with everything else.

I generally discourage my agents from storing memories locally because they don't travel; I'd much rather have a small in-repo document—whether it's an architectural decision or TODO or session log entry or runbook or environmental quirk or something else—that the prompt/CLAUDE.md/whatever can guide agents to grep and sed. Then I can dispatch work to a local agent, or a local agent running in a sandbox, or an agent running in a provider's cloud environment, or a GitHub agentic workflow running a headless harness, etc. and they all have access to it. Another benefit is that memory updates get PRs and reviews like everything else.

So what am I missing by not using something like engrim or gbrain?

Add a PoE switch and PoE splitters for IP-KVMs that don't already have PoE-in?

I went with Sipeed NanoKVM (PCIe) units across my homelab as well and I've also been happy with them. For a while it's been the best value option (not to mention the most consistently available option) cf. GL.iNet, PiKVM, and JetKVM. The PoE versions are great in a rack and the integrated ATX control is fully-featured (including the little power switch icon in the web UI turning green when the system is powered on). I set up an isolated OoBM VLAN with no Internet access and any switch ports assigned to it are isolated by default as well.

Have you played around with OpenBMC or any of the open-sourced BMC firmware alternatives?

ETA: specifically on older server/workstation boards instead of the stock firwmare

The memory controller is integrated into the CPU, so the motherboard chipset is irrelevant. There are some OEM-only v3/v4 parts with dual memory controllers, but the E5-2620 v4 is not one of them.

There are some OEM-only v3/v4 parts with dual memory controllers (because of a RAM supply crunch at the time, funnily enough), but the E5-2620 v4 is not one of them. The classic example is the very popular 12-core E5-2678 v3.

They've been stable for me but I'd be lying if I said I'd only heard good things.

Looks like prices on the E810-XXVDA2 have come up since the last time I looked while prices on the ConnectX-6 Lx have come down, so that'd be a good option!

You can't punch down twinax and the connectors are too big for keystones.

> Nowadays I'm not actually sure if DACs are cheaper than fibre now

Yep still are, although fiber transceivers are surprisingly affordable, and of course you can usually re-use jumpers when you upgrade transceivers so TCO over a couple generations might be comparable.

> I'm getting a really bad taste in my mouth for 10GBASE-T RJ45 SFP+ modules mostly due to the god awful heat they pump out

It's truly bananas especially in a homelab environment. Nearly every time I've thought to myself "oh I can just do a copper gbic for this" it's been the wrong thing to do.

If you start doing bonded links with DACs or if you have a bunch of servers, the cable management situation gets ugly in a hurry, and the usual solutions like patch panels and keystones aren't applicable. Source: my basement

I've started buying E810s even for 10G links. PCIe Gen4, lower power draw, RDMA support, generally backward-compatible with SFP+ DACs and transceivers, and relatively inexpensive. Not nearly as dirt-cheap as the X520s but not crazy expensive (last I looked, at least). As I gradually replace switches over the next few years I can start taking advantage of 25G.

I use UniFi for most of my home network so It Just Works™, but I've thought about mixing in Mikrotik for e.g. the compute rack so I can play around with 100G+ links and more esoteric stuff like VXLAN.

I've started buying Intel E810s for most purposes, even for 10G links. (SFP28 ports are generally backward-compatible with SFP+ DACs and transceivers.) The ones you can get on eBay for cheap typically run Dell firmware but it's serviceable. An E810-XXVDA2 is Gen4 x8; as long as the host slot can physically accept the card connector you only need Gen4 x1 electrical for a single 10G link or Gen4 x2 for dual 10G or single 25G.

My homelab has a 10G fabric (switched) for NFS, iSCSI, NVMe-OF, etc. and a 25G fabric (a mix of back-to-back and switched) for clustering (Ceph, DRDB, ZFS replication, migrating VMs).

I spun up some iSCSI-backed SQL Server a few months ago and 10G couldn't keep up with the workload, so I dropped in a pair of 100G ConnectX-4 cards with iSER (iSCSI Extensions for RDMA) support for that particular use-case.

Just because your uplink is less than 10G doesn't mean the rest of your network can't be a bit more capable. :)

No. It's twin-axial. Think coaxial, but more axial.

Couldn't have said it better myself.

Holy carp.

Is that a squeaker, or a squirter?

Yes, I want these for my kids so very badly. They have Yotos (similar to Tonie) for bedtime, and iPads for school work, but those are not ideal for a number of reasons. I want them to be able to experience music like I was able to with an FM+cassette walkman clone in the '80s and early '90s, or with my Nomads and iPods in the late '90s and early aughts. Hopefully someone here can suggest something!

ETA: OK, there are quite a few highly-rated options on Amazon, so I just need to solve the "putting music on there" problem and the "dropping it and immediately destroying it problem".

Unfortunately you had to run Windows NT, which was a leap for most folks and had poor support for games and some other software written for Windows 9x (e.g. software that required DOS mode and wasn't compatible with NTVDM). Windows 2000 (Pro) was a bit more approachable, and then of course Windows XP (Pro) smoothed out most of the remaining wrinkles.

I ran Slackware on my BP6 while I was in college. Of course CONFIG_SMP wasn't set in the default kernel config at the time so you had to build your own. Great for running bind, apache, sendmail, etc., and of course NetQuake servers. :)

> The Celeron 300A was the one most folks would go after for this.

Yes but they got hard to find in a hurry once word spread. I had two 366s in my BP6 overclocked to 550 but IIRC I had to buy a few to find two that were stable at this frequency.

> There was a whole cottage industry of folks modding these CPUs as a small side hustle for people who were not comfortable with soldering onto CPU pins if you wanted to put these into a SMP system.

No modifications necessary with the BP6.

They just renamed everything... the Pro Plus series is basically Latitude.

It's... not great. It's a dual-boot laptop that I take out into the field so I'd like to encrypt the Windows and Linux volumes with BitLocker and LUKS respectively, and ideally I would leave Secure Boot enabled for that extra bit of security. Ultimately I'll need to decide whether to disable Secure Boot or patch the kernel to let me override lockdown mode. I know SuSE has implemented it but I don't know if their patch series will apply cleanly to a mainline Ubuntu kernel.

I scored an "open box" Dell Pro 14 Plus on eBay for about 600USD. It's got two USB-A ports, two USB4 ports, built-in Ethernet, 4G LTE, upgradeable storage, and decent Linux support (on Ubuntu LTS at least). The battery is field-replaceable but the 32GB RAM is soldered. (There are other models with upgradeable RAM.)

It's got a middling display (the 2-in-1 display is better) and a somewhat dated Hawk Point SoC, but it's fine for running to a client's site for imaging or network troubleshooting or what have you. I still don't think it's going to last very long, but it's a nice complement to the MacBook I use for client dev work and it didn't break the bank.

> Power management on Macbooks is unbeatable in my experience, both Windows and Linux have really serious issues dealing with sleep and low power modes.

I've been dealing with this recently. Linux won't hibernate if you have Secure Boot enabled, even if your swap is encrypted. So I either have to leave my laptop plugged in all the time or remember to shut it down before unplugging it so it doesn't completely drain its battery while sleeping.

Yes, it's a race to the bottom for most. Lots of plastic and soldered, non-upgradeable components, and a lack of ports. The used market is tricky to navigate thanks to the proliferation of model names and numbers, so little to no salvation there either. Spend a bit more for something that will last a bit longer (MacBook) and/or be upgradeable (frame.work), or accept a cheaper model (e.g. Dell Pro) that will likely need to be replaced in a couple years.