Ah, hadn’t heard that, and I see there’s a TechCrunch article about it: https://techcrunch.com/2026/03/22/cursor-admits-its-new-coding-model-was-built-on-top-of-moonshot-ais-kimi/
Only 1/4th of the training was from Kimi, though, even though the model id evidently says it’s Kimi. 🙄
Nvidia’s middle finger to China is the reason they’ve developed open weight models like Qwen 3.6 27B that have modest compute requirements with performance catching up to frontier models. They have plenty of motivation to give Nvidia the middle finger right back, and I fully anticipate that a couple years from now, we will see Chinese open models on par with frontier models while also having inexpensive Chinese hardware to run them.
A pillory would be a good start. Then all the Amazon drivers can line up and dump their piss bottles in his face.
It Is Difficult to Get a Man to Understand Something When His Salary Depends Upon His Not Understanding It
I’d imagine this applies even more to billionaires who make the average person’s salary every minute.
And this is why Copilot failed. It’s bad enough watching a human do dumb shit on a computer, but having to watch a dipshit AI fumble around on your machine just takes it to a whole new level.
Yeah, whatever bias and censorship Micro$lop’s search index has, except supposedly without the tracking. It’s a step up from Google and Bing, but not the perfect solution.
I wouldn’t know because I mainly use DuckDuckGo and haven’t had a reason to use Google in a long time.
I think software subscriptions are a scam, but I don’t mind buying a perpetual license that is only good up to a certain version with additional fees for newer versions. It’s also fair to charge a recurring fee for something that has recurring hosting costs like a VPN, cloud storage, etc.
If they weren’t such dipshits, the “lifetime pass” should have been a perpetual license you can keep using as long as you want, but charge an optional fee for newer versions if you want to upgrade and get more features. They should also have offered a hosted service to make your instance available to others and charge a monthly fee for that. I think people would’ve been fine with all that.
Not necessarily the best example in this post, but in general, I find I get downvoted a lot when I make a good faith comment in .ml posts where I think I’m agreeing and am trying to understand the topic further. Getting 20 downvotes kind of kills my motivation to engage any further, though, so I usually just delete my comment , shrug, and move on. I’ve gotten to the point where I try to avoid wasting my time and energy commenting on .ml posts in the first place.
Maybe it’s just a different perspective on what a downvote means? To me, I’m generous with upvotes and withholding an upvote means I don’t find it interesting or disagree. I use downvotes sparingly for spam, trolling, comments made in bad faith, etc.
A downvote in my mind roughly translates to “fuck off”, so if a group gangs up and tells everyone with a slightly different worldview to fuck off, then they eventually will and said group will be all by themselves in their own echo chamber. If that’s the goal, then fair enough.
Good point, although on the local scale you mention, wildlife could still be impacted. Hopefully, the overall impact on the ecosystem will be monitored and studied before expanding these data centers more broadly.
I guess anyone who uses ShitLocker is shit out of LUKS.
Thinking about the decapitated heads in her freezer makes her smile.
Addiction drives quarterly profits, though.
I run k3s on a single node and it’s not really that much more overhead than Docker Compose if you understand k8s. I mostly have a deployment.yaml, service.yaml, ingress.yaml, and network-policy.yaml for each service that I’ve copy / pasted and updated. Here are some of the benefits over Docker Compose for my setup:
Has a built-in Traefik reverse proxy / ingress controller so I can access my services by domain name instead of by port, like http://jellyfin.lan/, http://forgejo.lan/ (using local dns on my OpenWRT router)
I use the Calico CNI so I can have network policies for each service to allow them to access only what they need. If a service doesn’t need internet access, it doesn’t get it.
I use Bitnami Sealed Secrets to store my secrets in YAML files that can be safely stored in git
ConfigMaps make it easy to manage configuration files
Easier to have separate YAML files for each service while sharing a network between them. Services connect to each other like http://forgejo.forgejo.svc.cluster.local/
Of course, if you’re looking to load balance across multiple machines, k3s makes even more sense.
Edit:
k8s is the clear industry standard for container orchestration at this point, so if you want something beyond Compose, a lightweight k8s distribution like k3s is an obvious choice.
deleted by creator
Never really understood why people drink aspartame water. Soda was originally intended to be a sugary treat and an experience to be enjoyed every once in a while, not a daily source of hydration.
I run Mint on a HTPC currently, but it might be worth switching to a different distro for this.
We are planning to join the Plasma release schedule starting with Plasma 6.7 (in June), which will make it possible for distributions to ship it in their repositories.
I’m imagining it will only be available in rolling release distros for a while.
Correction: 99% of CEOs are planning to use AI as an excuse to layoff employees to juice quarterly profits. They have zero accountability and golden parachutes, so no skin off of their asses.