Is this a significant problem? It seems like there are 15 solutions to every AI “text format” problem, but all are merely structure over feeding the AI informal English, structure which I believe the “transformer” model is specifically able to strip to a common form (which is how training it on the entire internet worked).
AI does benefit from learning to use skills, and being instructed with good structures (like “save [specific important kind of detail] in memories” and “plan before implementing anything”). But none of these memory solutions seem better than any other, they’re basically just different folders. Your CLAUDE.md can say “see AGENTS.md” and AGENTS.md can say “write memories to .agent/memories/“, and I’d be interested in a measurement, but I assume it performs almost the exact same as native CLAUDE.md and .claude/memories.
One thing I’ve noticed AI does much better than me, is not have OCD over the “perfect” solution.
Software developers need to realize that their app is not the only one that will be running on the user's computer. A single app taking 300 MB of RAM? No problem. 20 apps? That's already 6 GB of RAM.
And keep in mind we're talking about a damn chat client, not some complex video editing or CAD app.
> In our Linux test it opens in under a second and uses about 200 MB of idle RAM
Is it just me or does anyone else feel that 200 MB is a lot for a native chat app written in Rust, that's essentially a remote interface for the WhatsApp on your phone?
I remember back in the day, apps like Miranda IM (and early versions of Trillian) used only digits of RAM, even when you had multiple plugins talking several protocols at the same time, whilst having a ton of advanced features not found in WhatsApp. Yes I get that times have changed and software has become bloated etc, but 200 MB still feels a bit too much.
The entire section "Kept you waiting, huh?" is about this and some of the background is really interesting explaining why things are as they are. Here's just the last bit:
> Luckily for us, Sven was up to the task of dealing with all this nonsense. It just took slightly longer than expected! While the patchset to enable Thunderbolt was sent upstream first (another massive win for us!), we have also pulled it into our downstream kernel and are now shipping it to users!
> But there is a catch. Currently, only USB3, XDomain and Thunderbolt Networking are supported. You can plug in USB devices or connect two Thunderbolt computers together and make them ping each other, but that’s about it. PCIe and DisplayPort tunnelling — required for most Thunderbolt docks — are not supported. Yet. Watch this space.
> An enormous amount of gratitude is owed to Sven, who has stuck with this thankless chore for nearly 5 years in between all of his other incredible work.
Hey, author here! It’s just for fun :). I think I went too hard on the satire and people thought it was something other than a good time, but it’s just a good time.
There was a very obvious and unnerving shift in style that happened early in the article where it went from human to LLM. I find such shifts really glaring. Agents, in their current state, simply cannot produce quality prose and are not good enough to write blog posts targeting a human audience.
I found the topic fascinating, but eventually I started skimming and then gave up on it.
Because permissive licensing, and free software in general, rightfully comes with a certain cachet, and it turns out some people want the benefits of that reputation, but don’t actually believe in the freedom “to use, copy, modify, and distribute this software for any purpose.” Hence the treadmill from permissive licensing to copyleft, and then later from copyleft to “source available”…
“But software which OpenBSD uses and redistributes must be free to all (be they people or companies), for any purpose they wish to use it, including modification, use, peeing on, or even integration into baby mulching machines or atomic bombs to be dropped on Australia.”
Motor winding is an amusing problem. Everybody making motors has a powered winder. It's handling the wire ends that can be either manual or automatic. If you look at videos of winders, automating the wire end handling is most of the machinery.
Here's an overview of winding automation:
Full auto, robotic transfer.[1]
Full auto, manual transfer [2].
Semi auto. [3].
Manual, powered. [4]
Manual, unpowered motor overhaul.[5] (Country girl in China who fixes motors and engines.)
Those winding machines are all from the same company, NIDE. You can buy as much automation as you want, depending on your labor costs. Option [4] is cost-effective if you can have one machine tender per machine.
I feel uncomfortable asking in a thread like this but why would a CEO talk about his private life publically?
I know others have said this is a carefully written statement, but in my culture I would feel it would be rude not to keep it private? Sorry if this sounds autistic.
OpenAI is a bit more scattered brained and spread across multiple products.
Anthropic is way more focused. That helps in a way that builds upon another: their product tends to be better and the people's perception of it as well.
But "much less"? OpenAI needs a month or two of good choices, progress & announcements, and that flips. On the other hand, Deepseek could blow both out of the water unexpectedly, ending the argument about who has the superior tech.
There's also the Elon factor. Anthropic has a good relationship with Elon (to the point where he discounted the compute sold to them by quite a lot) while OpenAI/Sama is the devil to him.
It looks like you need to do a lot of leg work to introduce Rust functions into Clojure namespaces. I guess there is no way around it and this is not a criticism of clj.rs. But it doesn't seem to be as seamless as on the JVM.
I find TUI scales better for multiple concurrent agents across the same and different codebases, especially using tools such as herdr. I can fire off an agent, which runs sub-agents. Then check-in periodically to re-steer as needed or be interrupt driven for completions of tasks. I can't see how vscode/GUI is better for this?
People who oppose the use of LLMs have the right to do what they want with their software just like anyone else does. It’s not “childish” or “toxic” for them to express their opinion by changing the license on their own software or writing about it on their own blog.
Furthermore, if the GPL is really as unenforceable as you say, then there’s no reason why their license change is “punitive”. Either the GPL works, in which case the Asahi Linux maintainers have the right to restrict use of their software, or it doesn’t and there’s no reason why anyone should be upset by a gesture of free speech.
The Asahi Linux maintainers have the opinion that LLMs are harming their ecosystem, and if they want to prevent people from using LLMs on their software, then they should be able to do so without being unfairly attacked.
Auto mode may be significantly safer in the niche case where you want the model to have free reign, in a fully unopinionated environment, where it's allowed to make any changes and make any external calls.
Many of the problems when it comes to the manual approval feel like they stem from the expectation of using auto mode in the first place. After all, what's the point of improving the manual approval and safety matching if auto mode is the end all be all, and what's the point of training the model to output commands in a manner that can be parsed by a human if auto mode is always meant to be used? It's very rare for me to not see "command could not be statically analyzed", because the model keeps outputting bash golf.
- MacBooks are still among the nicest if not the nicest laptops hardware-wise, so some people like to run Linux on it.
- People have older Macs that they like to repurpose. E.g., we still have an M1 Mini that was used by our daughter before she got handed down a MacBook from my wife. It's now collecting dust, but I might install Linux to use it as a headless server.
What makes you think that persuasion isn't capped at some lower level than world-ending? There's a lot of incentive for ads to be as persuasive as possible, but they are so far from ending the world that it'd be science fiction for an ad to be that effective.
If you're just talking about convincing a few individuals to do experiments, then there are plenty of people who fall for scams everyday.
EDIT: I think it's 2 bit? If so, would be curious to see the quality of the output. Not great, I think.