HNHacker News
TopNewBestAskShowJobs

rubiquity

4,949 karma · joined June 1, 2013

Interested in improving the developer experience from the bottom up (databases, distributed systems, virtualization, etc.)

Currently building the best managed Postgres service at Snowflake, by way of Crunchy Data. Formerly at AWS working on (likely) the largest consensus and replicated WAL systems in the world as well as macOS virtualization just for funsies.

Twitter @bitsandhops

submissionscomments
rubiquity··on Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
At 8-bit quantization (q8_0) I get 20 tokens per second on a Radeon R9700.
rubiquity··on SpaceX says it has agreement to acquire Cursor for $60B
Wouldn’t Cursor agreeing to such a deal be almost ironclad proof they are subsidizing tokens/inference out the ass? There’s wide speculation all the large revenue growing companies right now are selling inference at break even or a loss.
rubiquity··on Framework Laptop 13 Pro
Are the mainboards and upgrade kits available for purchase now or just the whole laptop?

edit: I think I found it: https://frame.work/products/laptop13pro-mainboard-intel-ultr...

rubiquity··on Vercel April 2026 security incident
He doesn't work at Vercel but he is the type to never pass up any opportunity to chase clout.
rubiquity··on Tesla tells HW3 owner to 'be patient' after 7 years of waiting for FSD
I have both HW3 (2021 Y) and HW4 (2025 3). FSD in the HW4 is a delight. FSD in HW3 phantom brakes constantly both back when FSD was a pile of C++ and now with the "Lite" driving model. I don't see how Tesla can ever make FSD suitable on HW3 given the hardware (<200 TOPS).
rubiquity··on Qwen3.6-35B-A3B: Agentic coding power, now open to all
You are right that it is an environment variable, and that's how I have it set in my nix config. Thanks for correcting that.

Unfortunately llama.cpp is somewhat notorious for having lackluster docs. Most of the CLI tools don't even tell you what they are for.

rubiquity··on Qwen3.6-35B-A3B: Agentic coding power, now open to all
Have you tried running llama.cpp with Unified Memory Access[1] so your iGPU can seamlessly grab some of the RAM? The environment variable is prefixed with CUDA but this is not CUDA specific. It made a pretty significant difference (> 40% tg/s) on my Ryzen 7840U laptop.

1 - https://github.com/ggml-org/llama.cpp/blob/master/docs/build...

rubiquity··on Qwen3.6-35B-A3B: Agentic coding power, now open to all
Not sure why you're being downvoted, I guess it's because how your reply is worded. Anyway, Qwen3.7 35B-A3B should have intelligence on par with a 10.25B parameter model so yes Qwen3.5 27B is going to outperform it still in terms of quality of output, especially for long horizon tasks.
rubiquity··on Qwen3.6-35B-A3B: Agentic coding power, now open to all
Could be on a bike path where bikes are on the left and pedestrians to the right.
rubiquity··on Copilot is 'for entertainment purposes only', per Microsoft's terms of use
All Microsoft services are for entertainment purposes only, as in you’d have to be absolutely crazy enough to use before a Microsoft sales rep has taken your execs to a steakhouse and strip club.
rubiquity··on Why Switzerland has 25 Gbit internet and America doesn't
The optionality of consuming services from places other than internet titans for one would be nice.
rubiquity··on Vulnerability research is cooked
I was distracted by the picture of the ingredients to a Final Ward being at the top of the page.
rubiquity··on Hypura – A storage-tier-aware LLM inference scheduler for Apple Silicon
I did not know that. Thanks for sharing!
rubiquity··on Hypura – A storage-tier-aware LLM inference scheduler for Apple Silicon
llama.cpp and llama-swap do this better than Ollama and with far more control.
rubiquity··on Qwen3.5 122B and 35B models offer Sonnet 4.5 performance on local computers
This is great to hear! Out of curiosity, which brand did you go with? I tend to stick to Sapphire but the prices are within $200 of each other.
rubiquity··on I'm helping my dog vibe code games
Dogs are undefeated at reinforcement learning.
rubiquity··on End of an era for me: no more self-hosted git
The scrapers should use some discretion. There are some rather obvious optimizations. Content that is not changing is less likely to change in the future.
rubiquity··on Super Bowl Ad for Ring Cameras Touted AI Surveillance Network
If you’d like the opposite of this story: I was reversing out of a parking spot. I had moved about 2 feet and a drunk driver hit and ran my vehicle. The driver was charged by police because driving at 40mph and hitting cars in a shopping center in broad daylight got a lot of attention.

I gave the video footage from my car to local police and insurance company and the insurance companies ruled I was at fault because I was in reverse.

rubiquity··on Apple's slow AI pace becomes a strength as market grows weary of spending
Local AI sounds nice but most of Apple’s PCs and other devices don’t come with enough RAM for a decent price needed for good model performance and macOS itself is incredibly bloated.
rubiquity··on OpenAI signs $38B cloud computing deal with Amazon
I find it more likely that the entire "second" level of software companies are in OpenAI's cross hairs more so than Google. Salesforce, ServiceNow, Intuit, DocuSign, Adobe, Workday, Atlassian, and countless others are easier to pick off than Google.
rubiquity··on AWS outage shows internet users 'at mercy' of too few providers, experts say
Networking leaving the cloud provider (or even just to another zone on the same cloud) is $0.02 GB. That adds up real fast.
rubiquity··on Doing Rails Wrong
The entire CNCF landscape
rubiquity··on Doing Rails Wrong
It was real over a decade ago. This is sheer stupidity and selfishness of bringing hobbies into their jobs. If it makes Front-End devs feel better, the "DevOps" world isn't much better.
rubiquity··on Amazon fined $2.5B for using deceptive methods to sign up consumers for Prime
It should be one click.
rubiquity··on How AWS S3 serves 1 petabyte per second on top of slow HDDs
Rest assured STUMPY was replaced with another home grown protocol! Though I think a stream oriented protocol is a better match for large scale services like S3 storage than a synchronous protocol like HTTP.
rubiquity··on How AWS S3 serves 1 petabyte per second on top of slow HDDs
I used to work on the backing service for S3's Index and the daily humps in our graphs from lifecycle running were immense!
rubiquity··on How AWS S3 serves 1 petabyte per second on top of slow HDDs
I've always felt it's probably a wrapper around the Amazon EFS due to the similar pricing and that S3 One Zone has "Directory" buckets, a very file system-y idea.
rubiquity··on How AWS S3 serves 1 petabyte per second on top of slow HDDs
> So is any of S3 powered by SSD's?

S3’s KeyMap Index uses SSDs. I also wouldn’t be surprised if at this point SSDs are somewhere along the read path for caching hot objects or in the new one zone product.

rubiquity··on Zig feels more practical than Rust for real-world CLI tools
This blog is atrocious from an ad standpoint and the recent flood of posts feels promotional and intentionally controversial. The articles are also devoid of any interesting perspectives. Are people actually reading this?
rubiquity··on How to make the Framework Desktop run even quieter
Gotta give it the old Noctua and blow on that thang. Their fans are seriously awesome. I didn’t believe the hype at first until a Ryzen build was 5-15C cooler, and a lot quieter.
← PreviousPage 2 of 34Next →