Not sure if this has been noticed, but the pricing is “$299 (U.S.) or $49.83 (U.S.) per month for six months,” which means paying in installments is 2 cents cheaper than paying the whole price upfront. 49.83*6=298.98. So you’re not just getting a zero interest loan, you’re getting a discount.
I think this is what makes this so ridiculous. They put “room temperature” in the title and while it might not have been intentionally nefarious, it definitely has the effect that most readers assume it is about superconductors. There is nothing special about a room temperature semi conductor, your CPU, RAM, and storage drives have worked at room temperature for decades…
And if not for that word, the article would be completely unexciting. You found that you can make magnets out of multiple materials? We already knew that. If the materials aren’t abundant/cheap and easier to manufacture then this isn’t a story. And neither of those claims are tested or verified in this. So this isn’t a big deal.
> There is nothing special about a room temperature semi conductor, your CPU, RAM, and storage drives have worked at room temperature for decades…
This is a magnetic semiconductor (and antiferromagnetic). All non-experimental magnetic semiconductors require cryogenic temperatures. If this pans out (works, cheap-ish to produce), it could mean significantly faster memory, with significantly less energy usage and significantly less waste heat (and thus, even less energy usage).
Obviously it's very far from "panning out", but "room temperature" is not a given here. It's not normal, and it would be a huge deal.
"Whereas traditional electronics are based on control of charge carriers (n- or p-type), practical magnetic semiconductors would also allow control of quantum spin state (up or down). "
Magnets out of multiple materials, or magnetic semiconductors at room temperature?
Wikipedia seems to say it is new, assuming this one has "robust" coexistence of the properties the framing would be important and not just thrown in there to trick people into thinking it was about superconductors:
> To date, GaMnAs remains the only semiconductor material with robust coexistence of ferromagnetism persisting up to rather high Curie temperatures around 100–200 K.
If there is an important combination of material properties that previously was only available at cryogenic temps, a room temp version is significant since you don't need cryogenic cooling to take advantage of it.
A very honest take. They tried something, it showed a very small improvement, and they acknowledged it. Acknowledging that the cause appears to be a lack of engagement agrees with my own experience. You can give students great tools but that doesn’t make them interested in math. Khanmigo might be a good tutor but it won’t help unless you’re chaining kids to a desk and not letting them leave until they finish some quota of problems.
I wondered how he didn't see this coming from a mile away, but in looking up more about Sal Khan himself, the surprising thing is that he never had any 'normal' teaching experience, and that explains everything. In a normal classroom, good kids don't need teachers. You could give them a book (or some Khan Academy videos, or a reliable LLM) and they'd learn the material far faster than you can teach it. Now let's talk about the other 95% of kids...
Those other 95% are going to generally have either a lack of motivation or a lack of intelligence. And for these kids a good teacher can do a lot, but actually teaching the material itself is a relatively small part of it. It's about giving them that motivation, or adjusting the material for slower students. And most importantly, it's about doing all of this without making it apparent that you're doing this, since keeping it feeling 'natural' is an important part of making it feel interesting and also helping to build a better rapport with the kids. That rapport/connection building is critical, and completely outside the domain of LLMs.
A lot of academically successful people recieve everything but a "normal" education. They are the ones who have developed the ability to self study from a young age and have the drive. As cliche as it is, a lot of the academically successful students I see are the ones who are passionately curious. While many of the Asian kids are obviously more inclined to be more motivated (due to their parents), this motivation makes them passionate at the very least. Some other students simply love the thrill of being better than their classmates. From my observation, a lot of those kids love going to schools where they are challenged and all their peers are equally if not better than them. This cycle is ONE of the MANY reasons why they are so successful.
There are not that many super educated autodictats. If you take that good kid out from the classroom and put them into a library with no teacher and no peers in it, their education is likely to just stop or severely slow down.
If I could rephrase this, I wonder if you're also saying in a way that 5% of people are naturally self-driven but for 95% of people, the relationship matters more than what they're doing. If they have a connection with the teacher, then they'll learn.
Have there been cases where kids in the 95% did end up being self-driven about a particular topic that they had an interest in? Or did they always need to be given motivation?
It probably isn't a good tutor, I'm using frontier SOTA models and whenever I ask them to explain anything they don't consider my question and don't display any understanding but just spam out unrelated simplistic metaphors, can only imagine how much worse the models the kids had access to were, LLMs don't have understanding and can't fake it either, there's no conception of how a memory might be formed or something might be understood, without that qualia it can't be a good teacher it can only output text that looks like it might be a good explanation until you read it
A good tutor builds an understanding of the learners current mental model then tries to broaden it with targeted materials (text, diagrams, videos), selected exercises, the Socratic method ... All LLMs can do is produce text that statistically looks like tutoring. AI has a lot of uses but tutoring is not a great fit. I think a well written textbook even would be better than AI because the author can have a general understanding of where the target audience is usually at and where they're expected to end up.
> whenever I ask them to explain anything they don't consider my question and don't display any understanding
You are absolutely doing it wrong.
My experience is completely different. It doesn’t take a frontier model to give a good explanation (likely plagiarized from a textbook or a “blank for dummies” article).
Try a bit harder to tell it your background and the kind of analogies you want and don’t want.
Possibly this is topic-dependent. Most of my queries are about science and math.
My strongest subjects where I showed the most interest were based on like 25% natural interest and 75% how I vibed with the teacher and they vibed with me, atleast in High School.
University is definately more interest based but my success was still very influenced by the quality of instructor. The materials and methods are important but not nearly as much as the quality of instructor.
"...unless you’re chaining kids to a desk and not letting them leave until they finish some quota of problems."
Uhm, hey, could we talk for a quick minute after this meeting? No, no worries - it's just that the other teachers and I were wondering if we could have a quick chat about what "engagement" means :)
Humor aside, your point is very valid - interest in a topic is often a prerequisite to someone engaging with a topic. Even stuff like "I hate this class, but I gotta do well if I want to become a doctor someday" is based on an interest in that longer-term goal.
I’m not advocating for the practice of chaining kids to a desk…
I’m trying to point out that a tool might be helpful for you to practice, but if it isn’t used then the tool can’t help. Maybe if you were forced to use the tool then you could see the benefit, but as you noticed, the act of “forcing” something may feel undesirable from a moral perspective.
Sorry, I didn't mean to say that you actually want to chain kids to desks. I thought the juxtaposition of the two statements was funny in a "clearly you did not mean this, so it's funny that one can intentionally mis-read it in this way".
And - thank you for your posts! They're interesting and insightful!
One reason is that the llama.cpp team (GGML) has strict requirements that a human must understand the code they are contributing. If a project is fully vibe coded they can’t contribute. So a lot of projects where an AI went and coded a bunch of custom kernels to increase speed are left to their own devices.
I think this is a fine behavior. We can have upstream purists that are strict gatekeepers but don’t get in the way of downstream forks. Debian has some this in the Linux landscape for a long time, and it has enabled Ubuntu, Mint, etc. to flourish without compromising themselves.
> A proper code review usually takes something like one hour per 200-400 LOC and you should be spending at least that much time on code review alone.
Not only is this not enforceable (how do you enforce how long someone spent working on a codebase on their own local machine?) the metric is severely off which instantly makes me question the competence of the llama.cpp dev team. You can easily review 10-100x that in an hour, even if you're being super pedantic about it.
I also just ran _one_ of their files (with include deps) through Astra and it detected >100 vulnerabilities/correctness errors (with over 10 outright UB/memory corruption issues). It's actually outright shocking.
> question the competence of the llama.cpp dev team
Following them for years, they seem extremely well put-together, and have excellent judgment. They are using the same policy as Linux and Debian (in my words, the speed of light is human understanding and judgment). Whether it is reasonable is a different question from enforcement, which typically comes down to "this seems fishy, explain your reasoning".
As for code review, the rule of thumb I've used for decades is: it takes about as long to review and understand as it does to write. Your 100x metric is completely outside of anything I've seen in any hobby or professional project, ever.
I'd like to see specific files you scanned and specific vulnerabilities cited.
If you can thoroughly and accurately review 100 * 400 = 40,000 lines of CUDA kernel code per hour, I know roughly 1,000 people who would love to hire you right now.
What you have quoted is a sentence elaborating on the requirements listed in the document. This is provided to help you better understand why the requirements exist and the goal they are trying to accomplish.
The reason you SHOULD take that time to read the output is because you must read it to understand it.
And the way this is enforced is explicitly called out in the document (and again in more detail in the linked AGENTS.md): the maintainers may ask you to explain it.
Frankly, I don't believe you. I'm half decent at writing CUDA directly (a holdover from a project a few years ago and it is a nice skill to have), the degree to which these are optimized is unlike 99.9% of all other code out there and even a tiny slip-up is either going to kill your results, your performance or both and if you're lucky only in some edge case. Understanding this code is hard work. I made a couple of minor edits to some .cu files in llama.cpp yesterday because I have a pretty weird setup which they obviously did not anticipate and it took a couple of hours to get it 'just so'.
> You can easily review 10-100x that in an hour, even if you're being super pedantic about it.
I think spending 9-18 seconds per line of code is a good baseline for review. There are diffs where you mostly moved a lot that can go quicker. There are also diffs where I spend 20 minutes thinking about 5 lines of code.
> What you really mean is, the core team there doesn't want to lose control.
It's 100% this. They basically produce vague guidelines such that only the core maintainers are allowed to use LLMs, under the guise of "well of course we understand the code" and no one else is. It's also completely unenforceable, how are they going to prove whether someone understands the code or not? Even if they show sufficient evidence/understanding the maintainers can simply sabotage them and accuse them of using an LLM to explain the code. No one wins here.
Agreed- this is the same problem we have with trusted admins or devs who have elevated privileges on their networks. We have to trust that the admins won't use their power to steal company secrets or misuse company resources. If you don't trust the admins, then they can't fix things on your network and there is no point in having them.
If you want an agent to act on its own, like pushing to a git repo, managing dependencies, building and testing, etc., then you have to trust it as much as any other privileged user.
If you don't want to trust it, then you're just forcing yourself into the reverse centaur role, where the agent edits some code, but then has to stop and ask you to push the changes or build the software again and run the unit tests.
I suppose there is a principled way of doing things like "I trust you do do basic commits but I will handle merge conflicts" and "you can build modules in this directory but you can't build outside of it" but this is just a lot of effort that most orgs won't bother with.
Even then if the agent goes rogue and decides to do the merges you can’t stop it if it has any kind of access. This goes back to the OP’s point - agents can’t be 100% constrained.
You can absolutely run an agent as a limited-privilege user that only has write privileges for specific files and only has execute privileges for certain files. If it is running as a limited-privilege user it can work on code in it's own copy of the repo and make commits and send pull requests, but it can't do the merge. The problem is that nobody wants to go through the effort to set up all these permissions and nobody wants to take the time to review everything and perform all the manual actions.
Some shops are now generating tens or even hundreds of PRs a day with relatively little involvement. That volume is simply beyond what anyone can reasonably review.
Not bad. I’d appreciate a much longer time limit. Got really rushed at the end. But this was crazy hard on my phone, I was straight guessing most of the time. I think I could be a little better with more time on a desktop screen.
Another idea would be games that were all pictures of a single category- like 10 AI and 10 real pictures of basketballs, or food on plates, etc.
But I think that your experience is actually the useful one that proves a point. Most people consume content on their phone (e.g instagram) and would usually just scroll by photos. They wouldn’t deeply analyze them.
I 100% agree with you. It looks like multiple people (including me) were typing comments about how good the Selena Gomez Oreos were at the same time as you.
The post seems to have found a strange niche of HN, or maybe the Selena Gomez Oreos had wider appeal than I thought.
I really liked Selena Gomez's Oreos, but I would prefer that Oreo brought them back as "horchata flavored" or "cinnamon cream flavored" so I didn't have awkward conversations every time I accidentally said, "I like Selena Gomez flavor" and people would ask how I knew what Selena Gomez tasted like.
Honest question/suggestion for the HN audience- Since Qwen released the weights for Qwen3.8 2.4T-A95B and we already have the staring point of Qwen3.6 35B-A3B, couldn't someone distill the bigger model and make a "pseudo" Qwen3.8 35B-A3B? Sure, it wouldn't be an official Qwen release but couldn't someone improve on Qwen 3.6 and get the thing everyone is asking for?
I am calling this a suggestion for the audience because I don't have the will/resources to do this.
> Base / architecture: Qwen/Qwen3.6-35B-A3B (Qwen3_5MoeForCausalLM, 256 experts, ~3B active). The "3.8" in the name refers to the teacher, not the base.
Not endorsement, haven't run it myself, just found the link.
reply