The invisible guardrails are a test run for the invisible enshittification. Just wait til they start dialing down ability to better absorb peak demand or simply to have more profitable inference
HN user
cmdrk
Lead SRE, MIT
email: username at protonmail.com
IIRC, this is the same lab who created Glaze.
Do you think this will be a trend for larger companies as well?
The decadal move to all-cloud-all-the-time killed off in-house hardware teams while the C-suite chased their OpEx dreams.
It would be interesting if we come full circle on this.
My observation is that the true believers really don't want to think of models as an inert pile of weights. There's some mysticism attached to imagining it's the ship's computer from Star Trek, HAL-9000 or C-3PO. A file loaded into memory and executed over is just so... _pedestrian_.
100% the same problem here.
I have to show people extremely zoomed-in screenshots of how $VENDOR default monospaced fonts get rendered compared to Terminus at the correct size in order for them to understand my pain. The hinting is just blurry bleh.
These days, because I am also old, I want a comparatively large pixel-perfect font. I've yet to find a good one but haven't looked much beyond Terminus honestly. Maybe I can render it an acceptable integer multiple without it being too large?
Bellringing, specifically change ringing. It’s a type of church bell ringing that is rather algorithmic in nature. Tends to attract mathy types. Religion not required or expected!
If you have English-style tower bells near you, it's worth checking out, even if only to listen.
Their only original idea was Metaverse
surely you must be joking.
too expensive for our internal server needs; not the right fit for our datacenter partners/customers
You and me both. They're doing neat stuff, but I wonder how many other potential customers feel that way too.
What is Oxide's market? It feels a bit like advanced alien technology that is ultimately a little too weird and expensive for most enterprises to adopt.
bcantrill gave a great talk many years ago about compute-data locality. would be nice to know if those ideas panned out for some customers, but it seems the world has by-and-large continued to schlep data back and forth.
it's too bad too. The concepts behind Manta were such a great idea. I still want tools that combine traditional unix pipes with services that can map-reduce over a big farm of hyperconverged compute/storage. I'm somewhat surprised that the kubernetes/cncf-adjacent world didn't reinvent it.
Random assortment of projects as time allows with the $JOB.
- Prototyping a cute little SSH-based sorta-BBS, inspired by the Spring '83 protocol, but terminal-centric rather than web-based. It's called Winter '78, and if we get another Great Blizzard this year, I'll be able to make some progress on it!
- Another prototype, for an experimental HPC-ish batch system. Using distributed Erlang for the control plane, and doing a lot of the heavy lifting with systemd transient units. Very much inspired by HTCondor as well as Joyent's (RIP to a real one) Manta.
CephFS implements a (fully?) POSIX filesystem while it seems that TernFS makes tradeoffs by losing permissions and mutability for further scale.
Their docs mention they have a custom kernel module, which I suppose is (today) shipped out of tree. Ceph is in-tree and also has a FUSE implementation.
The docs mention that TernFS also has its own S3 gateway, while RADOSGW is fully separate from CephFS.
Does their training corpus respect copyrights or do you have to follow their opt out procedure to keep them from consuming your data? Assuming it’s the latter, it’s open-er but still not quite there.
I feel that the article draws a false equivalence between skepticism and doomsaying. If anything, thinking AI is as dangerous as a nuclear weapon signals a true believer.
just wait for the AI summary
Not to pick on you, but there are always posts like this in every Erlang thread. One is not strictly superior to the other, and the BEAM community benefits from the variety IMO.
I wouldn't mind the Erlang-dominated front page coming back :)
Containerization is amazingly great for scientific computing. I don’t ever want to go back to doing the make && make install dance and praying I’ve got my dependency ducks in a row.
This sounds super cool! Any public code you can share?
Broadly: Still forging ahead building a game server framework in Erlang. I've shifted my attention away from Godot integration (which AFAIK is still working) and toward LÖVE and Lua. Godot is great, but having to write GDScript on the client and Erlang on the backend has caused me many headaches in my game logic. My current goal is to have a beautiful, concurrent, Erlang-based control plane with Lua-based game logic running on both the server and the client.
To that end, I've most recently been hacking on Robert Virding's Luerl (https://github.com/rvirding/luerl), working to adapt the Lua test suite to chase down some small compatibility issues between PUC Lua and Luerl. While Lua is a lovely language, it would also be swell to get Fennel working under Luerl. I wrote a game for the LÖVE jam a few months ago in Fennel and it was a pleasant way to dip my toes into lisp-likes.
I've also been adding things to control plane software, Overworld, here and there: https://github.com/saltysystems/overworld Happily all of the Protobuf and ENet stuff that I've already built nicely carries over into the LÖVE world.
How does this compare to something like BYOND? I realize it’s dated now but conceptually there are some similarities.
My understanding is that Luerl can remove various modules from being used, like preventing use of the IO library for instance.
I would say BEAM itself is not particularly sandboxed, and certainly clusters of Erlang nodes assume a lot of shared trust.
“My job safe from obsolescence by new technology”, says man directly threatened by new technology
Godot has a nice netcode story with ENet, WebSocket and WebRTC integration. If you just want a off-the-shelf product instead of bespoke Erlangy stuff, I think there are a lot of good options out there like Godot's native high-level networking, products like Nakama, etc.
Server backend framework: https://github.com/saltysystems/overworld
Chat server implementation (Erlang): https://github.com/saltysystems/chat_server_example
Chat client implementation (Godot): https://github.com/saltysystems/chat-client-example
I'm still tinkering on my multiplayer game server framework when not grinding away at my day job.
The idea is essentially: An Erlang-based control plane, supporting ENet and WebSocket connection modes, with Protobuf for messages. Erlang has an excellent concurrency story, and I think it's a great fit for game servers. I've wrapped up a bunch of this work into behaviors on the Erlang side, such that developers can target the "gen_zone" behavior (for instance) to implement a tick-based game server. I'd like to expand that into other types of games, such as turn-based games.
I've also got a Godot plugin for generating a client library based on your protobuf schema. The plugin handles session stuff, exposes functions for client-to-server messages, and emits signals for server-to-client messages.
These days I'm working on integrating Luerl (Lua in Erlang) and Love2D support. I want to be able to take advantage of Erlang on the back-end while writing the majority of game logic in Lua. Further down the road I want to explore hot reloading parts of the Lua game state on the client/server, perhaps with an in-game editor, to develop the game "inside-out", in a way.
Erlang is my favorite language but getting a job writing Erlang feels impossible. I make it a habit to ctrl-F every Who’s Hiring? thread and find Elixir occasionally and Erlang never.
as a Lisp-curious person, Fennel was a gateway drug for me
Indeed, but the hallowed Job Creators have the means to influence the people in power to make the taxes go away.
Anything going on these days in Chicago?
Location: Chicago, IL, US
Remote: Who doesn't say yes to this?
Willing to relocate: within the US
Programming Languages (in order of preference): Erlang, Elixir, Bourne Shell, Python, Go, Javascript
Platforms: FreeBSD, Linux, BEAM, Kubernetes, OpenStack/KVM, AWS, GCP
DevSecOps: Puppet, Ansible, GHA, GitLab, Crowdstrike, Rapid7
Résumé/CV: https://www.linkedin.com/in/lincoln-bryant-8b770148/
Email: cmdrk@proton.me
Long-time technical team lead with over a decade of experience in a broad variety of technologies (too many to list! I've put some of my favorites above). Looking for new and interesting distributed systems work, preferably using the BEAM platform (Erlang, Elixir, Gleam, etc).
it's quite okay to be a fan of the cultural output of a society, and hardly boot licking.