zenoprax

This is my experience as well. I have spent many hours trying to eliminate this kind of contamination and the only solution I've found is to re-write from scratch every time. If the result isn't to your liking just reset its context and provide a more detailed prompt.

My thoughts exactly.

I recently experimented with some image generation models and I was surprised at how much context/prompting is needed to get something that doesn't look Lovecraftian. There was seemingly no downside to adding more and more detail up to the limit. Prose sits between code and images in terms of fuzziness and I wonder if promptmaxxing is the winning strategy for technical documentation.

I've been thinking that fine-tuning a model to write differently shouldn't be that hard (tedious, perhaps) and if I had the resources I would consider doing it myself.

Do you have any experience with multi-node operation with Incus? I like the simplicity of Proxmox for moving VMs and LXCs around but I haven't settled on a solution for OCI. K3S is an option but it's a big jump. I'm moving towards Fedora CoreOS VMs at the moment for live migration and security isolation.

What is your workload?

I briefly tried GLM 5.3 Flash and the first thing I noticed was how slow it was followed quickly by its single-mindedness. With DS 4.1 Flash I find that it understands that not all the detail in my prompt is important. Sometimes I like to give a quick example of what I'm looking for and it can take the hint whereas GLM will seemingly reshape its worldview around this tiny detail.

DS just handles ambiguity and my bursts of detail with much more grace. I'm only using "max effort" and the CoT is reasonable compared to GLM.

Air-gapped Dropbox for a pre-rsync, pre-ssh world.

The timeless debate of separating art from the artist.

In another project (literally: "Untitled") of theirs they are explicit about motivation:

> This is to fight against the trend of crappy github-powered websites that plague the internet.

I looked into kata but it doesn't support Podman (my preferred baseline). I assume that touching a kernel requires some sort of rootful interface but I don't know enough about how seccomp and the other layers interact to know if it's the right balance.

gVisor felt a bit more "hacky" initially (at least compared to krun) but I will look into it a bit more as I saw someone else integrating it into their own setup.

My goal is to get 80% of the isolation for 20% inconvenience. Right now it feels like I'm getting 30-50% isolation for 10% inconvenience — I'm willing to tolerate a bit more friction if I can get a bigger jump in security.

I was just testing this and it's not clear that it works out of the box. `krun` shows a different kernel than with `crun` but it doesn't reflect the dropped capabilities in the same way. I'm probably holding it wrong but I'm not sure what to look for at the moment.

This makes sense to me but if I continue with the "fancy testing" analogy it seems like there is no accounting for interactions beyond the system itself: it has perfect unit and integration testing but only for a myopic amount of global state.

I am curious about the limits of what one can expect this language to solve and therefore the types of programs/domains to which it is more/less suited. Why is determining quorum the canonical use case?