kstenerud
Weak processor, 16gb ram, but good enough for browsing and running daily tasks, and it's simple enough to connect to more powerful machines when the need arises. And the docking station has HDMI and displayport. Hardware H.264 codec, and also HEVC and VP9, but alas only with the windows driver.
My big rig desktop is now in the basement, sleeping until I need to run LLMs or diffusion or a heavy compile. I also sometimes use it for Steam gaming, connecting to it via moonlight (which works surprisingly well, even over WIFI).
Yes, we're going to quickly come to a head on the issue of robots on sidewalks, much like we have with e-scooters, or Google's ill-fated bikes.
The robots open a new possibility on an existing system (sidewalks), which is going to lead to problems from both sides until the legislature comes up with new rules. This is how free and democratic societies work.
Q put it best:
"This is the man you are today. And you should be happy! You have a real heart beating in your chest, and you get to live out the rest of your life in safety: running tests, making analyses, and carrying reports to your superiors."
The difference was that rather than turn it into soap opera porn, the characters actually tried to behave professionally (most of the time).
Really? So doing things like learning to play an instrument or mastering chess are pointless?
> University (really all education beyond teaching people to speak) will be pointless because there’s no work to do and because growing your human capital doesn’t change your life outcomes: you get the same UBI whether you’re illiterate or a PhD.
Why must a university education be only a career generator? What about a well-rounded education, with a specialization in the things that interest you?
> If Claude Code has existed when I was starting out, nobody would have cared about Austral or my other open-source projects
Claude code is here, and yet I'm still writing open source code that people appreciate.
> As a teenager I tried to install Linux, ran into all kinds of problems, and went on the IRC chatrooms to get help. And in the process I made friends. Today, I would just ask Claude. AI disintermediates everyone, so that we no longer need each other.
Your friends don't need to all be techies. There are many other ways to make friends.
> What even is the point of cities in a world without work?
Community? Actually physically going outside and meeting people?
> The BATNA to hanging out keeps getting better, cheaper, and more convenient. When people’s sensorium becomes an infinite firehose of AI-generated superstimulus, why would they hang out with anyone?
Human connection? Experiencing things in the world together? Physical contact?
There were a few small issues with its work (it's not the greatest at getting layouts right, and one of the features it implemented feels clunky and will need tuning), but the architectural refactors it did worked perfectly and I'm happy with the performance improvements.
I haven't yet checked out the research tasks it did, but usually it does a decent job. I could have just asked it to research and implement, but I find it best to manually verify its conclusions first.
"You may bring one 8.5x11 double-sided cheat sheet into the exam. You may write anything you want on it. And I promise you that if you prepare such a sheet, you'll never even look at it during the exam."
And he was right.
I'm the exact opposite. I want the choices and possibilities so that I can run with them. When the space is limited, it can still be fun and interesting, but only until you've plumbed the depths and everything you do becomes derivative, repetitive, and boring.
Growing up, my favorite toy was a huge bucket filled with random lego pieces. Some were basic bricks, some were from the technical set, some were from themed sets (cars, space ships, etc). I'd spend hours coming up with new designs, the original box designs long forgotten.
I wrote this because I'm unhappy with the offerings out there, including Emby, Plex, and Jellyfin. Also too many bugs that have been sitting for years with no solution. For Jellyfin in particular:
- SyncPlay is buggy, doesn't support voice, and seems to be unmaintained now.
- Way too many CVEs, mostly FFmpeg injections (CVE-2026-35033, CVE-2026-35032, CVE-2026-48793, CVE-2025-31499).
- The session key leaking in URLs.
- Upgrades breaking the database.
- Needless transcodes and wrong "can play this" guesses.
- Subtitles that load late or drift after resuming.
Differences from Plex:
- Watching your own films away from home isn't paywalled, and never will be.
- Hardware encoding is free, and always will be.
- No cloud account required unless you want my convenient discovery service.
- If you don't want to use a cloud account at all, you can set up your own discovery server on your own VPS for watching away from home even through routers and shit (also free).
- No ads, no storefront, no telemetry.
- Watch Together with voice. Multiple people can watch a synced video, and talk to each other (I've tested this with me in Germany and the other person in Australia, and it worked beautifully).
- Hardened against corrupt video files and issues like NFS failures (getting stuck forever in a syscall). Everything vulnerable is separated and isolated.
- Open source. The server is AGPL and the web clients are MPL.
- Your files are left alone. Media is read-only by default. The server reads the .nfo metadata files you already have, from Kodi for example.
What's not there yet:
- Native apps. I'm only doing a web client for the time being until all the features are in place. It's pretty fast, though.
- Music. This comes later.
Our evolutionary heritage is that shared belief and social unity (even when it's wrong) trumps independent thinking for group survival in a hostile landscape.
- I've addressed the most common complaints and unfixed issues in all of the competitors
- "Watch together" mode because my gf is stuck in another country for the next few months. Includes voice chat and will soon have tv casting to watch on the big screen and chat on the phone.
- all the hole punching tricks to get access from remote if you want it (or just keep it local only if you'd rather not)
- more features I'll talk about later - out of time atm
Basically: easy self hosting, includes a discovery server you can run on your VPS, all the usual stuff you expect from this kind of software, and free forever.
The only thing I'd charge for is my own discovery service which gives you convenience because you don't have to host your own. Any feature that can be hosted yourself remains free forever.
That was the case for the majority of the post-WW2 order, but is no longer. We're witnessing the birth of a new world order. Likely the new order will have an international law of a sort because it's good for trade, but it won't necessarily be the same as it was.
Underneath it all, only power decides.
Whenever a new technology makes what used to be hard easy, all the non-experts pile in and produce shit for awhile, until they're finally forced to admit that software engineering is hard (or die trying).
The same will happen with LLMs in a couple of years.
Most V2 rewrites fall under a similar category.
AI doesn't preclude this. In fact, it can help accelerate parts of it.
He's describing the typical big project lifecycle:
- Examine the landscape
- User research (how they use existing software, what their frustrations are, etc)
- Brainstorming
- Early ideas and prototypes
- Refinement, user feedback
- Solidify the vision and high level process design
- Choose technologies
- Design & architecture
- Plan out phases
- Build phases, then test them with users
LLMs are great at research, and great at prototypes. Once you have your design, they're good at coding as well. They're also good at distilling user feedback.
The authors are just being lazy.
Of course you can get the same kind of thing in binary. I wrote a drop-in binary JSON replacement because it's easy to write a binary one that's twice as fast as simjson and yyjson. The important thing is to never give the drop-in replacement any extras that break roundtrip compatibility.
Was a fun little project to write, and quite useful for me: https://github.com/kstenerud/bonjson
Yes, there were genuinely helpful people there. I strove to be helpful as well. But my most common experience there was:
- People asking why you're doing that (without answering the question).
- People telling you that you shouldn't do that (without answering the question or providing an alternative).
- People closing your question as off-topic or duplicate without actually reading and understanding what you wrote.
- People editing your question and changing the meaning (leading others to innocently conclude off-topic or duplicate). Or removing actually important details.
- People giving just plain wrong answers that get upvoted, with comments from more senior people begging them to change or remove it.
Or in the words of Brick Top: "If I throw a dog a bone, I don't want to know if it tastes good or not."
And yes, every so often a golden answer by an amazing person.
Now I can just ask a few LLMs, comparing and contrasting their answers. Even better: I can interrogate them on their answer.
Stack Overflow had a community of sorts, but it wasn't anywhere near like a physical community.
We may work online, but we exist in the flesh.
The built-in "sandboxes" these companies provide are laughable.
It's very easy to argue with stawman arguments.
Every time the LLM throws jargon around, you call it. "What do you mean by gated wedge?" You call its bullshit, check what it's saying against your understanding of the overall system, and keep it on the straight and narrow.
It's a lot like supervising a junior dev who happens to be very quick at absorbing lots of info, but not so great at the big picture.
Until I fully understand what's going on, the PR doesn't move and my interrogation of the LLM doesn't end. My interaction is littered with "Explain X" and "How does this square with Y?" and "What if Z happens?"
The interrogation is the point, without me having to wade through hundreds of lines of irrelevant code to get at the meat of the matter.
As a test, I built a sandbox with only the host-side filtering proxy allowed for networking. 99% of traffic was HTTP. No QUIC at all.
npm, pip, apt, go, curl and git-over-HTTPS all worked on the standard proxy environment variables alone. No mirrors or other coaxing needed.
DNS is disallowed through the chokepoint, but that's no problem because the proxy resolves host-side anyway.
I used two agents: One with network access to collect the evidence, and one with everything except the model endpoint cut off, which did the analysis.
The second agent's entire input was attacker-authored. So PHP droppers, obfuscated loaders, database rows, filenames, blah blah.
In this case I'm more worried about hostile input attacking the agent, and I need to contain the damage. My sandboxing solution does that by restricting access to the source data, making it read-only. The work dir can only transfer data via patch and apply (like a git workflow), so even my workspace can't be modified until I approve each change. And then restricted network means that any compromise ain't going noplace.
The second agent couldn't even install PHP or contact any CVE site to check if it was looking at a known attack, and that was by design. All it could do is write up a report about what it observed, not make assumptions about what it is. I could then take its (much smaller) clean output and pass that to a third agent with network access.
This is forensic work, so of course not your median dev's bread & butter. But the attack surface is only just starting to be plumbed. Compromising input can turn your agent into their agent, planting things as easily as planting worms was back in the early internet days when people connected without a firewall.