Hacker News Reader

volf_on MiMo v2.6

I've got a working recipe to run this model on Dual DGX Spark: https://github.com/volfco/spark-vllm-docker/blob/main/recipe...

Averages ~25-35tok/s which isn't bad for a first attempt.

volf_on Ox-Alpha Is GLM?

same business model as crack. best way to get people hooked is to make the first hits free.

volf_on Ox-Alpha Is GLM?

It's possible

volf_on Ox-Alpha Is GLM?

Yeah. It could be. The Z.ai DC latency is still ~1.2s faster than whomever is serving this model.

volf_on Ox-Alpha Is GLM?

GLM 5.3 and all previous models don't have a vision encoder and can only accept text. Ox-Alpha can accept video and images, so unless Z-ai added a pretty good vision encoder for this model, I don't think so.

My money is on Moonshot and this being Kimi K3.5. The measured tps and latency is in-line with K3's tps and latency from Moonshot.

MiniMax M3.5 is also possible (but the MiniiMax provider is a lot more performant than the lab behind ox-alpha, so less likely).

volf_on Qwen3.8-2.4T

A ~5TB model.

volf_on Hygon Reveals 512-Thread CPU and AI GPU to Rival Intel Xeon and Nvidia

Not mentioned in the article, but this is a x86 architecture because Hygon has a license for Zen core IP from AMD.

volf_on Annoying and alarming things about OpenCode

I switched from OpenCode to Pi and there was a big improvement in terms of tool calling performance and I find the experience less buggy.

OpenCode has also seemed to have disappeared from https://openrouter.ai/apps/category/coding

volf_on Fable extended until 19 July

Keep using Fable. Don't think about using GPT 5.6 Sol

volf_on Ask HN: Developers still enjoying development after AI?

No.

the amount of coworkers who have said "<ai tool> said X,Y,Z". Then they make basic mistakes with what they need to do and it becomes my problem.

volf_on Discord will require a face scan or ID for full access next month

atproto is a really good attempt at solving this issue

volf_on Discord will require a face scan or ID for full access next month

do as we say, not as we do

volf_on Problems with D-Bus on the Linux desktop

The world is filled with "good enough" solutions. If it has 80% of the functionally needed, it will be good enough for most use cases.

> the most meritorious solution never rises to the top

You know why? Because it was the best solution at the time that covered everything they needed. You can only complain about it because the initial problem was solved and people were able to move on to something else.

Does it work for what we need it to do? Yes? Ok. Now that we have settled this problem, we can continue on developing.

Meritocracy is only relevant to someone who has ideas but no desire or ability to actually structure the idea in such a way that would work in the world we find ourselves in today.

> Maybe D-Bus has something similar?

So you just commented here to say nothing but "dbus sucks because it isn't the best. I wonder if there are better options?".

volf_on I got an Nvidia GH200 server for €7.5k on Reddit and converted it to a desktop

That's awesome.

These are the best kinds of posts

volf_on Everyone in Seattle hates AI

because autocorrect and predictive text doesn't help when half your job is revisions

volf_on Aldous Huxley predicts Adderall and champions alternative therapies

Exactly.

Adderall has no positive relationship to my mental efficiency. It can in fact be a negative once your passed the 8 hour windows where its still in your system.

At the end of the day, it makes it easier to not bounce between different things. It doesn't help me be smarter. It helps me drive to work without needing to listen to music and be on my phone.

Modafinil... maybe.

volf_on ADS-B Exposed

"Elon Musk" as a filterable option.

A++. No Notes.

volf_on Algorithms for Modern Processor Architectures

A commercial enterprise is dropping support for older cpu architectures in their newer OSs so they can improve the average performance of the deployed software?

Don't see how that's controversial. It's something that doesn't matter to their customers or their business.

volf_on Telephone Line Rural Outside Plant

Also check out https://www.youtube.com/@ConnectionsMuseum

The channel is about the fully functional Central Office turned Museum in Seattle

volf_on Telephone Line Rural Outside Plant

Yes. Dynamic as the IP is renewed everytime your device resets

volf_on Telephone Line Rural Outside Plant

https://www.youtube.com/@ConnectionsMuseum

There's a fully functional system deployed in Seattle as a Museum

volf_on A TikTok ban would deal a blow to creators, businesses and the American economy

You clearly don't use TikTok. There is quite a lot more than just dancing for 15 seconds.

> Any time one hears someone mention the phrase "content creator", one should, by default, execute s/content creator/manipulative parasite/ in their mind.

volf_on A TikTok ban would deal a blow to creators, businesses and the American economy

TikTok has one of the lowest bars to monetization out of all the major social media networks.

There is a large amount of creators on the platform that live off their content and eCommerce enabled through the platform. So it disappearing overnight would severely impact people who have a majority of their livelihood through the app.

Facebook. Twitter. YouTube. They all have short form video, but TikTok has a lot of features that just don't exist on other platforms.

volf_on Consider removing autogenerated files from tarballs

You could have `./configure` print a message saying to do exactly what you said.

volf_on Ask HN: Who wants to be hired? (February 2024)

Location: Seattle, WA

Remote: Yes

Willing to relocate: Maybe, for the right gig.

Technologies: Python, Rust, Kubernetes (Helm, ArgoCD, CNPG, AWS ACK), AWS, AWS EKS, AWS CDK, Terraform, Grafana LGTM (Loki, Grafana, Tempo, Mimir), GitLab, PostgreSQL, OpenTelemetry, System Architecture, Hybrid Cloud, DevOps/SRE stuff, and more!

Resume: https://www.linkedin.com/in/volf/

Email: hn [at] m1.volf.co

Been doing DevOps/SRE for the past 7-8 years. I've found myself working well in positions where I can work closely with other engineers to build out Cloud/Distributed Infrastructure and develop software in a way that makes it not painful to run in the cloud.

I can help take small MVP applications & infrastructure, and help turn it into something that can scale up as the business does. I'm comfortable working with any stack- just depends on how much time it'll take me to brush up on the language or design.

volf_on Ask HN: Who is hiring? (January 2024)

Any chance for remote, or is it all on site?

volf_on Noteable.io is shutting down

Cool. How about you open source the application? Let the community take over if they want to.

volf_on Amazon EFS now supports up to 250k IOPS per file system

I'm aware.

Buuttt... you're trying to compare apples (a rack in a DC) to oranges (an AWS Native Solution that spans multiple DCs). And that's before you get into all the AWS bullshit that fucking sucks, but it sucks more to do it yourself.

A Rack in a DC isn't a solution that's useful to people who are in AWS.

volf_on Amazon EFS now supports up to 250k IOPS per file system

You use EFS as a shared folder that you can share between a number of different workloads. If you want a POSIX compatible shared filesystem in the cloud, you're going to pay for it.

For example. I setup Developer Workspaces that can mount an EFS share to their linux box, and anything they put in there will be accessible from Kubernetes Jobs they kick off, and from their Jupyterhub workspace.

I can either pay AWS to do it for me, or I can figure out how to get a 250k IOPS GlusterFS server to work across multiple AZs in a region. I think the math maths out to around the same cost at the end of the day

volf_on Is distributed computing dying, or just fading into the background?

https://aihorde.net/ is one that I've been running off and on.