amitprasad

Our system somewhat predates ublk; at the time we wrote it, FUSE was the most reasonable option. Moving over to ublk would require re-architecting around block devices rather than filesystems, but is indeed something we're looking at!

Relevant post: https://modal.com/blog/dollars-per-token-considered-harmful

(disclaimer: I work with the author)

I’m being a little reductive and ignored Guix purely based on adoption; but you’re completely right. I believe the core argument is that there is a (relatively small) category of these tools of which Nix is the most well-known representative.

Heard on everything else, thank you!

Just because you or I may invest effort into deep-thinking, it does not mean that others will.

I'm not worried about this at Modal, but I am worried about this in the greater OSS community. How can I reasonably trust that the tools I'm using are built in a sound manner, when the barrier to producing good-looking bad code is so low

To be clear, I am also having the most fun I've had when it comes to side-projects and even more exploratory things at work. I don't derive all my joy from "Good Code" -- that's silly! I would much rather ship tangible products and features and/or tackle things at home that I wouldn't otherwise.

On the other hand, the other responsibilities of being an engineer have become quite a bit less appealing.

It's hard to say. Perhaps LLMs of tomorrow will become capable enough to fix the mistakes of LLMs today. If so, great -- I'm worried about nothing.

If not, we could see that LLMs of tomorrow struggle to keep up with the bloat of today. The "interest on tech debt" is a huge unknown metric w.r.t. agents.

I would be extremely happy to be proven wrong! I love using agents for exploratory prototypes as well as "rote" work, but have yet to see them really pay off when dealing with existing tech debt.

I find that the flaws of agentic workflows tend to be in the vein of "repeating past mistakes", looking at previous debt-riddled files and making an equivalently debt-riddled refactor, despite it looking better on the surface. A tunnel-vision problem of sorts

Right: Having "Good Code" is an investment into future velocity.

IMO we shouldn't strive to make an entire codebase pristine, but building anything on shaky foundations is a recipe for disaster.

Perhaps the frontier models of 2026H2 may be good enough to start compacting and cleaning up entire codebases, but with the trajectory of how frontier labs suggest workflows for coding agents, combined with increasing context window capabilities, I don't see this being a priority or a design goal.

There's an opportunity-cost here. I use agents to be more productive. As many have noted, "Good Code" doesn't rank highly compared to actually shipping a product.

The tragedy, for me, is that the bar has been lowered. What I consider to be "good enough" has gone down simply because I'm not the one writing the code itself, and feel less attachment to it, as it were.

If you're not looking for GPU snapshotting the ecosystem is relatively mature. Specifically, CPU-only VM-based snapshotting techniques are pretty well understood. However, if you need GPUs, this is a notoriously hard problem. IIRC Fly also was planning on using gVisor (EDIT: cloud-hypervisor) for their GPU cloud, but abandoned the effort [1].

Kata runs atop many things, but is a little awkward because it creates a "pod" (VM) inside which it creates 1+ containers (runc/gVisor). Firecracker is also awkward because GPU support is pretty hard / impossible.

[1] https://fly.io/blog/wrong-about-gpu/

Generic data stores often don’t have this luxury — if you’re designing a system in which data is relatively opaque, you’re often forced to work with bytes. (e.g. rocksdb, etc)

author here -- agreed on all fronts. Mentioned this in the other comment but I approached the topic from a relatively narrow perspective (I was working on a specific project at the time)

I think it's worth including these things in a future update to the post, but I didn't have the time / need to explore it back then.

In the meantime, I'd point to the following post on Unicode that remains very nice to read >20 years later: https://www.joelonsoftware.com/2003/10/08/the-absolute-minim...

Unexpected seeing this posted here.

I wrote this post mostly out of interest for a personal project and thus it's not actually a very holistic exploration of the topic. May revisit and update it in the future :)

How is the 795s7? What do you put it to work on?

Just sprung for one at a good price.

This is notoriously hard when you start to involve GPUs

This is heavily language runtime dependent — there’s nothing that fundamentally stops you from doing anything during the phase between jumping to an entry point and the main()

Not so much the confusion of what little endian is, but how we tend to represent it in notation. Of course this confusion was back when I was first learning things in high school, but I imagine I’m not alone in it

I think I got stuck in the same rut that I learned address space in whilst writing that diagram. I would tend to agree with you that your model makes much more sense to the student.

Related: In notation, one thing that I used to struggle with is how addresses (e.g. 0xAB_CD) actually have the bit representation of [0xCD, 0xAB]. Wonder if there's a common way to address that?

You’re right, and I knew this back in February when I wrote most of this post. I must have revised it down incorrectly before posting; will correct. Bit of a facepalm from my side.

Ah I should have taken the time to verify; It might also have something to do with the way I was compiling / cross-compiling for RISC-V!

More generally, I'm not surprised at the symtab bloat from statically-linking given the absolute size increase of the binary.

Ah yes, you’re completely correct :)

General idea was just to highlight some of the dangers of vector registers. I believe the same is true of ymm (256) to a lesser extent.

Note that, especially on certain Intel architectures, using AVX512 instructions _at all_ can result in the whole processor downclocking, and thus ending up resulting in inconsistent / slower overall performance.

https://stackoverflow.com/questions/56852812/simd-instructio...

Feels like there’s three different ways that JSON objects are represented in KEON.

It might be more concise, e.g., in the Newtype case, but the sacrifice seems to be quite a bit more cognitive complexity, which I would personally value over conciseness

Also relevant: LLM in a flash: Efficient Large Language Model Inference with Limited Memory

Apple seems to be gearing up for significant advances in on-device inference using this LLMs

https://arxiv.org/abs/2312.11514

Hey Marcus! Super cool to see another fellow Minecraft-jumpstarted developer. I got into programming way back in 2013 with Minecraft mods and then Bukkit/Spigot plugins.

> I decided to look for open Developer positions, to work with a team of experienced developers so I can learn even faster.

Seeing as you're in Europe and not in the U.S. (with exorbitant tuition costs), I would actually recommend against skipping out on College. I can confidently say that while experience in the industry is great (having worked part-time at a startup, as well as various internships during College), I think the knowledge gained from University classes is often underrated.

Not only that, but College is definitely a life experience that I would recommend not to skip lightly.

Regardless, this is super impressive work!