uberduper · HackerTrans

23 giorni fadiscuss

I get about 20 minutes of work from my 5h limit with the $20 plan. It wouldn't bother me as much if codex would continue after the token bucket refills instead of waiting for me to show up and tell it to continue. I don't jump to the $100 plan because I would be in the exact same situation.

uberduper

mese scorsodiscuss

Gimme an rtx spark in the new framework 13 pro.

uberduper

mese scorsodiscuss

My take from working at a Big Corp is that individuals using coding agents can increase velocity substantially and produce good quality work, assuming they are proficient with the tools. But it falls apart quickly when you have a team trying to work together.

imo we either need to centralize the agent and and submit plan, spec, reference doc MRs rather than submitting code changes. Or develop SCM systems/workflows that incorporate plan/spec/reference/prompt metadata with code so intent can be factored into merges.

uberduper

mese scorsodiscuss

Nix isn't involved in my container images. I just take the dependencies and env vars from the flake and generate a dockerfile.

Guess I need to try out dockerTools. That looks really convenient. Thanks!

uberduper

mese scorsodiscuss

I do the same. Codex manages a per project flake.nix and uses `nix develop` for all testing. nix-direnv for my own convenience. I generally have it generate dockerfiles or other deployment assets at some point.

Codex is way better at nix than I am.

uberduper

2 mesi fadiscuss

The distinction is scale. "AI Datacenters" are a new level of scale with new levels of power consumption and heat generation. Sure you could run regular compute and w/e in them but it's not practical to build these mega sites for regular compute. GPU Compute / AI workloads require network/interconnect bandwidth and latencies where distance matters so you're forced to solve problems you wouldn't otherwise have to. Those problems are mostly solved with money.

uberduper

2 mesi fadiscuss

No. Agendas like, "I need to push my ideas for promotion credits."

uberduper

2 mesi fadiscuss

I'm building the same stuff I've always built. Just faster and with less dependence on others. Not having to argue with devs that have their own agendas has been my biggest benefit from coding agents.

uberduper

2 mesi fadiscuss

TBF this was the case prior to the firmware change. It wasn't a bait and switch. It just wasn't obvious to someone buying a printer they thought worked with open source slicers.

uberduper

2 mesi fadiscuss

Did you install the network connector Orca slicer prompted you to download? It's a closed source blob that runs on your PC which I'm presuming you haven't air-gapped as well.

uberduper

2 mesi fadiscuss

`initcall_blacklist` is a thing.

uberduper

2 mesi fadiscuss

amd gpus compete but they lack the interconnect. NVLink performance is a huge deal for training.

uberduper

2 mesi fadiscuss

There's a few dimensions you can look at for gpu load. Probably the easiest indirect metric to watch for gpu load is power usage.

But if you really care about this, you should actually profile your application. nsight systems makes this pretty simple to do. Dunno how many actually care about having a TUI.

uberduper

3 mesi fadiscuss

I previously worked at a managed database as a service company. On more than one occasion during my time there, a junior engineer deleted a customers database and at least one time one of our most senior dbas made it unrecoverable. Never got such straight forward confessions out of them.

uberduper

3 mesi fadiscuss

I did some work on an agent that was supposed to demonstrate a learning pipeline. I figured having it fix broken linux servers with some contrived failures would make for a good example if it getting stuck, having to get some assistance to progress, and then having a better capability for handling that class of failure in the future.

I couldn't come up with a single failure mode the agent with a gpt5.x model behind it couldn't one shot. I created socket overruns.. dangling file descriptors.. badly configured systemd units.. busted route tables.. "failed" volume mounts..

Had to start creating failures of internal services the models couldn't have been trained on and it was still hard to have scenarios it couldn't one shot.

uberduper

3 mesi fadiscuss

Do people really want codex to have control over their computer and apps?

I'm still paranoid about keeping things securely sandboxed.

uberduper

3 mesi fadiscuss

Well they're liquid cooled, so there's some extra handling required. Then there's the sky high failure rate.

uberduper

3 mesi fadiscuss

A dozen people couldn't handle the number of daily nvidia compute tray RMAs one of these datacenters produces.

uberduper

3 mesi fadiscuss

Appreciate you making my point for me.

uberduper

3 mesi fadiscuss

I don't know anything about this particular site, but I presume it's one of the new mega gpu sites.

I'm seeing many people in the comments with an early 2000's era concept of datacenters. The scale of these new sites is mind boggling. Take your idea of a typical datacenter building. Make it 4x bigger. Then put 4 of them together into a cluster. Then imagine 10 of those clusters at the site.