I wanted to switch to Rust for new embedded project. I was looking for native RTOS, and embassy came up. But i always feel like it's hacky to me. What i want is RTOS that is similar to FreeRTOS or Zephyr.
It's not about features, i believe i am not that advance user. It's just different from the RTOS i used to work on. Most of exist C RTOS work quite similar(FreeRTOS, Zephyr, ThreadX), the difference is mostly in API and Software support arround the core kernel.
But Embassy, with different task switching mechanism, mean i need to learn another new concept, understanding the pros/cons. That is not what i want to do when busy adapting to Rust.
In my experience unless you're really at the edge you can basically just treat it like regular threads, and you'll probably run into fewer concurrency issues than you would with threads. Probably the bigger problem is just the rough edges in async rust, though they are getting filed down over time.
Personally I'm excited about these sorts of experiments. We all learn in different ways, and these sorts of techniques allow us to create "on-demand" syllabuses and lessons that fit our learning style and learning level.
It's not perfect, but I'm optimistic this will be a useful way to teach/learn in the future.
And, to be clear, I think this will be best utilized within a group/community setting. I don't think it will replace teachers or classrooms.
> Nvidia selling more GPU's at the cost of it's datacenter business is pretty close to Kodak selling digital cameras at the cost of film.
This assumes a) AI is a zero-sum game, and b) we're actually talking about on-prem AI will replace cloud-based AI. I think neither statements are true.
AI is like compute: we'll need all sorts of it, in various sizes, everywhere. I'm sure Nvidia whats to own all the workloads.
On the other hand, I do think open weight, like open source, will win in general.
It will take about 35 years for digital camera sales to reach Kodaks profit margins.
The market growing and companies raising the economy are not zero-sum. Nvidia/Kodak killing the golden goose is zero-sum. When the difference is your valuation crashes that's zero sum for the company, just not for everyone else.
My only concern is if we can limit the economic impact from lowering investments and causing a 40% market collapse circa 2008/9.
It's a good write up, but it's lacking some details, the most important one is: which Claude model was used?
The second issue is: what was tooling and the prompt approach?
(To be clear, I have no problem with the premise of the write up. But without some details like this, it's sort of like saying "I had a bad board on my deck, and my tape measure wasn't able to help me remove the nails. What a bad tape measure."
Opus 4.whatever (it was last week) via a command line interface in the IntelliJ Claude plugin.
The series of prompts weren't particularly interesting or innovative on my part: a paste in of the user report then a few back and forths on fixing it, me reviewing the changes and coming up with the final answer.
It might feel interesting, but it's sort of the crux of the issue. Average or below average prompts will produce average below-average results. The model can't make up for that.
Not saying every problem can or should be solved but AI, but mastery of the tools is kind of important when evaluating the tools. It's like complaining that vi or emacs is slow to use because of the bindings are complicated.
It's more like asking what editor and keyboard layout they use. Highly relevant to the user but you should simply assume someone describing work is using a setup for it they find productive. If you decide to dismiss their output it wouldn't be over these details.
> Music is about connecting to human emotions, not poor facsimiles of it
"art is in the eye of the beholder."
I listen to a lot of EDM, which can be very mechanical, but I personally have strong emotional connection to. I personally would welcome AI-generated music as an alternative to human-made.
To be clear: I do agree a "human-verified" system would be great, but I don't think it would be black and white. And I would guess that eventually AI music will be better than a lot of human made music.
You will probably be able to listen to machine-generated music on most major platforms. I just hope there’s one which excludes all of that.
Personally I think it’s a bit like cultural junk food: it has the appearance of real food, but leaves one hungry afterward. Which really isn’t all that surprising – music isn’t just some random collection of patterns, it’s intimately tied to real culture. Current AI software is only ever going to copy and regurgitate human culture, not make meaningful creations from scratch.
I agree with the other poster. I think it's very difficult to guide this missile so that it blows up AI-generated music and doesn't blow up EDM.
My own taste in music is pretty junk-food-y I guess. Electronic music and not the pretentious kind. Dubstep, electro. Give me something that goes wub-wub. Incidentally, I think this experience mostly isn't one about human connection? Like, there is some circuit in my brain that likes that sound and wants to be tickled.
I can play classical piano to a mediocre standard. I listen to it and enjoy it occasionally. But, honestly, what I feel like my spirit needs is something that goes wub-wub and I think that space is densely seeded enough that maybe we can scale back human involvement in producing it.
Sure that’s fair, and maybe my ideal platform doesn’t really work for electronic music and works better for singer-songwriters that perform live. Which is fine - I just like the idea of a platform that guarantees that a real person made this music.
I don’t mean electronic music writ large, I mean AI generated music.
Electronic music is probably my favorite genre, broadly. But there’s a human behind the machine, not a random collection of patterns. To use a concrete example: NIN is about 1000% more interesting because of who Trent Reznor is, and not because it’s merely good music.
This disconnect is much more of an issue with say, country or bluegrass or jazz. To divorce those from the musician and their cultural context is to miss the whole point.
Typing “make me an electronic guitar song about a bad breakup” is a whole lot different than learning to play the instrument yourself and conveying your own emotions from your real experience into a song that you write and perform.
If you can’t see how these are fundamentally different things, I don’t know what to tell you.
I think a lot of songs about bad break-ups were written by talented musicians without actual experience of a bad break-up, riffing on the corpus of songs they'd heard about bad break-ups.
Like, a lot of times you're just engaging with someone's desire to have made a song, and what they felt about some songs that someone else made.
Even if you generate the soundtrack, if you set the lyrics it can be great music. At least for one person, who chose the lyrics. I set some poems on music and I absolutely love the results.
Do you really think that's the only way to do it? I spend hours refining the output, slicing things up, redoing certain parts, tossing it into the DAW and compressing, adding effects, etc. I mean, if you think one shot prompts are the only way to use AI to make music fine, but you're being intentionally obtuse.
I made (what would eventually get called) EDM in high school and a lot of what I enjoyed was dismissed as “not real music”. It’s not a musician playing it, but a computer! Unless a guitarist was plucking strings or a pianist hitting the keys, it wasn’t “real”.
Doesn’t matter how carefully crafted it was: it’s only real if you couldn’t hit “play”. Sorry, Mike Oldfield. Hate to break it to you that you’re a fake musician.
I agree with you. I do enjoy some live musicians jamming on a stage, but for a lot of the genres I frequently listen to, I’d have no way of knowing if a song was written by human or by AI. If it’s good, it’s good.
I learned music production by first learning how to produce house. Because of this, I can confidently say that house is not music. You literally just copy/paste over and over. Just because a human makes something, doesn't make it good or "music" necessarily.
Their approach of "manually tagging songs that are good 'non-AI' and not good 'AI'" is a questionable approach from an engineering and product perspective.
They would benefit much more from a have a better recommendation and ranking algorithm that carefully monitors all metrics, recommends high-performers, and excludes unpopular content from the feed.
You can use the fact it was AI-generated as a signal, but it is just a signal among other criterias, not an outright ban.
Essentially, explore songs and artists, exploit winners.
Lot of people abort a song/artist/creator/etc.
Lot of fake listens.
People don't like the song of an album
Creator is historically with a low score
etc...
Then you judge the popularity of a song, an album, a creator, a playlist, etc not its creation method. Exactly like a genre type. It's not because you don't like country music that everybody should be forbidden to listen to country music.
It's good for them too, the more streams they do, the more money they get, and the more their audience is engaged. If the person doesn’t like “AI-generated genre” then just downrank it heavily on its recommandation feed, like YouTube or TikTok does.
People engage with music in different ways. Some people focus on lyrics and the connection they make. Other people largely ignore lyrics and focus on the rhythms and patterns. Both are valid and both are very human. We do not need gatekeepers coming in and judging what is music and what is not music.
House music can bring me as much joy as listening to Bach performed by a skilled ensemble. It depends on where I am at mentally. Both are valid forms of human expression.
I went to see an orchestra performance. They literally just only made sound by dragging their bows across the strings on their "instruments". I can confidently say that is not music.
and it's so low effort too. like they never try dragging their instruments across the bows. where has all the creativity gone now that we have these "official protocols" of performance and musical quality!
It's pretty clear you didn't get beyond the very basics of music production or house music if you think this. I'd love to hear your masterfully produced house track to prove me wrong, though.
I believe we've already crossed the quality threshold for being equal to or "better" (subjective) than human.
For EDM, check out the AI artist "Vibfy". Especially the song "I Hear You" as it has the best mastering so far. The melody and vocals of all the songs are fire, but in some of the earlier songs the mastering is sub-par with strange volume changes and muddy beats.
There is an AI folk band called "We're all f*cked" that is incredibly good and indistinguishable from actual humans.
YouTube is already recommending tons of AI-generated electronic music. It's pretty easy to spot: It will have millions of views and the "artist's" discography will only go back to 2025 or 2026. It's usually just singles. The cover art is vague atmospheric AI-generated. The "artist" will have no presence online save for a couple instagram account with a couple posts that are also AI-generated.
I'm not saying it doesn't exist at all, I'm saying electronic music is significantly harder for AI to do than a conventional love ballad or a rock song.
AI can't do "robotic" math music, it's best at sappy generic emotive stuff. (I guess this isn't at all a surprise for those that know how the musical sausage is made.)
I occasionally listen to Drum n Bass & House music on YouTube. Just a few years ago a number of compilations, artists and playlists were made with the tagging of "PS2 Ambiance" or "PS2 D&B to Relax to", etc.
Most of these were either:
1. Legitimate music from the PS2 era games.
2. Music which could of easily been included in the games of the era.
I constantly see "PS2 Jungle Mix" and its AI, or entirely by one guy (which I assume is AI?).
This type of music is my background music whilst playing a game, or just chilling because it has no lyrics to be distracted by. It doesn't really matter all that much if its AI or not, but I still find myself filtering by uploads from a few years ago to avoid the slop.
the statement isn't "GLM 5.2 has large token usage", it's "GLM 5.2 has large token usage vs modern Opus".
I haven't used it, but this wouldn't surprise me. I see ~30% lower token usage for better results with Opus 4.8 vs 4.6 (and i had great results with 4.6)
I'm comparing with GPT5.5 on Codex and it's not even a competition. GLM takes way longer and eats a lot of tokens getting work done, it's easy to rack up a big bill on openrouter. I tried the $20 plan from ollama, too, and ate through half a month of budget in a few hours and blew my daily limit twice and still had to get codex to complete it -- which it did with only 10% of my monthly limit remaining.
GLM is promising but it's pretty costly, all things considered.
I've been using GLM 5.2 extensively for the last few days. It is slower, and the lack of multimodality is a bummer.
But, it produces solid results for a fraction of the price. Worth checking out if you have the time.
One of my goto "tests" of a new frontier models is having it rebuild a programming language from scratch. For GLM 5.2 I had it rebuild the old Rebol language in Rust:
The core is open source and always will be. Crates are licensed the usual for rust: Apache2 and MIT. This also includes the relay servers.
In addition we provide services that any commercial deployment using iroh will probably find essential: observability and a custom non rate limited relay network, as well as priority access to the engineering team.
iroh is an open source library. The relay servers are open source too but number0 runs public, rate limited, relay servers that can be used by everyone. The commercial offerings are for dedicated relay servers and more insight into your network.
reply