Have been using a "rubber duck" for decades to talk things through when there is no human available. By talking it out, it uses different neural pathways to help solve problems faster. My first rubber duck was given to me when I joined Nortel Networks. I did not understand why they gave one to me. They said it would help me to think through things faster, I was skeptical until I tried it for the first time months later. My current "rubber duck" is tux penguin that my wife gave me decades ago and it has sat on my home office desk ever since.
I think "through writing" / "write to think/feel" a lot (usually in a mix of 1st/2nd person, 5-6 pages of text per day on avg.). Think: stream-of-consciousness writing, journaling, but also just keeping a screen with disappearing text notes at hand (append-only) where I ramble.
That being said, not many things can compete with talking at your dog for an hour or so.
That is excellent news, been running ollama locally since 2024 testing out running local models to compare them against the big frontier models and for times when I want to keep my conversations local and private.
From the article:
"The project’s binary distributions didn’t include the required MIT license notice for the llama.cpp code they were shipping. This isn’t a matter of open-source etiquette, the MIT license has exactly one major requirement: include the copyright notice. Ollama didn’t."
I'm always puzzled by how often I see this happen in the open source community. Can the llama.cpp take legal action against Ollama for this? What's the point of these licensing requirements if they aren't (or can't be) enforced meaningfully?
Note: I'm not blaming the llama.cpp team for inaction here, I just want to see them get some justice!
Have used it in my homelab for the past 2 years, can recommend it. Easy to run in docker and helps both get better search results and keeps your search history local and in your control.
Like the author, CarPlay is table stakes for any vehicle purchase. I literally installed a Sony screen receiver in my 2016 F150, that was given to my by my father after he lost his license due to his age. That was a fun weekend with the dash apart reliving my younger days when I did car audio installations. The truck lasted 1 week before I had the dash apart to get CarPlay back. The native Ford infotainment system was so crap.
Same stack here — proxmox as the hypervisor, packer for the golden images,
opentofu to deploy, ansible to configure, whether the vm runs bare services or
hosts containers. The separation is what keeps it low-effort: image build is one
job, provisioning another, config a third, so a change only ever touches one
layer. Few minutes a week, rollback through captured state or pinned versions.
On your synology holdout: I run two of them plus a custom truenas, and I treat
the storage layer as the deliberate boundary of the declarative model rather
than fighting to pull it in. Topology is truenas on fast ssd backing the running
vms, synology one as primary, synology two as the backup target for both truenas
and the primary. The synologys are imperative islands — dsm doesn't want to be
config-managed — so I codify the consumption side (the exports, and the vms that
mount them) and treat the boxes themselves as data, not infra. Truenas is the
exception, since the api gets you closer to declarable.
Are you leaving the synology fully hands-off, or driving any of it through the api?
2. storage, RS818+ (16GB): 4x12TB RAID0 (everything not easily replaced is backed up)
3. san, an RS2418+ (64GB): 8x1TB SSD RAID10, and 4x5TB RAID10
I can't picture too much of a reason to drive it through the API, or at least not yet - still getting it all finalized. But I also managed to figure out IPv6 with Xfinity across my VLANs and SLAAC.
I have 2 Synology boxes, both DS1618+. The first one is my media storage, the other one is a backup target for both the media storage and other devices on the network. I also don't use the API, these are just "dumb" storage boxes that exist to store. No real configuration here from my perspective.
A long time ago a wise person said to me is it better to be kind or right. I live with that decision matrix every time I engage with a human. At the end of the day the humans carry significant ego and emotions that for me to be right, most of the time, they have to be wrong. Arguing turns into a zero sum game with only losers on both sides. Until I learned this lesson, I won many arguments but lost the human. This article bears this out from another person learning the same lessons.
This rings so very true. Have been in the software business for too many decades and having inherited other people's code many many times, there is always the decision about what to do going forward. Leave it running, refactor it gently or rip and replace. Each has it's own set of traps that need to be carefully considered.
Happy to see that Fable is back, was sad to see it go after getting to use it for only 1.5 days before. Add on top sonnet 5 availability and coding is looking fine again.
