HN user
meet.hn/city/48.906387,2.4452231/Bobigny
Socials: - linkedin.com/in/cyrilmhansen
---
The KV-cache memory usage also seems remarkably frugal, even at the full context length. That could make this model particularly useful in multi-agent coding workflows.
I wish KV-cache memory usage and related optimizations were discussed more clearly in new model announcements and demos.
Do you think advanced users should consider AVIF/AVIF2 along JPEGXL for long term pictures archival ?
Which kind of encode settings do you suggest for conversion from high resolution RAWs or JPEGs ?
Does anyone know how hard it would be to do the same with other tablets like Allwinner A733 ones ? Allwinner is supposed to also have some Linux support.
Have you thought about building a RISC-V “fantasy computer” core for the MiSTer FPGA platform? https://github.com/MiSTer-devel/Wiki_MiSTer/wiki
From a software-complexity standpoint, something like 64 MiB of RAM possibly even 32 MiB for a single-tasking system seems sufficient.
Projects such as PC/GEOS show that a full GUI OS written largely in assembly can live comfortably within just a few MiB: https://github.com/bluewaysw/pcgeos
At this point, re-targeting the stack to RISC-V is mostly an engineering effort rather than a research problem - small AI coding assistants could likely handle much of the porting work over a few months.
It appears that only the 4o text interface has been removed. Advanced Voice Mode is still branded as 4o, although it has been gradually evolving over the past few months. I suspect that voice mode is what most users are actually attached to.
These are objects from a time when people were willing to watch machines work, not just get instant output.
Today, even if small DIY plotters were cheap to build, they’d mostly live in the “art / hobby” space: most users won’t wait several minutes for a page when a laser printer does it in seconds.
That said, it would be great if a simple, well-documented DIY standard (protocol + format) emerged that hobby plotters could implement and that common tools (Inkscape, CAD, etc.) could support out of the box.
The brand suffered from energy price hikes, felt particularly sharply after 2022, and its marketing could clearly be improved. Only now, after more than two or three decades, are new designs finally appearing on the roadmap.
These glasses were once ubiquitous in public middle-school cafeterias, so the emotional attachment runs deep across generations.
30 years ago, without an SSD, my Pentium box booted to the desktop in 60 seconds — that’s roughly 5 billion CPU cycles at 75 MHz. Today, with blazing-fast SSDs and CPUs running at 4+ GHz, a typical PC boots in around 10 seconds — that’s 50+ billion cycles per core.
One of my teachers used to say: "In computing, 3 seconds is an eternity." These days, that’s enough time for 20 billion AVX-512 instructions.
It’s hard to accept that anything not truly compute-intensive needs more than that. Realistically, we should be able to hit 300 ms latency across the entire UX — and yet we don’t.
Wirth’s Law still hits hard in 2025. It's like the ghost of your first CS prof whispering "I told you so" every time an app eats 500MB to display a list of items.
We were supposed to use better tools to build better systems. Instead, we used faster hardware to make it acceptable to ship ever-more bloated layers of abstraction. Everything depends on everything else, and no one knows what any of it does, just that it “works on my machine.” Until it doesn't.
It’s not just about performance — it’s about comprehensibility. You used to be able to hold a system in your head. Now? Good luck tracing anything across 8 layers of indirection, six config files, a microservice mesh and a runtime whose lifecycle even the maintainers don’t fully understand.
I find myself drawn to projects like Red[1], MIR[2], or even Metamath[3] — not because they’re production-ready silver bullets, but because they remind me what it’s like to work on systems that are conceptually finite. With MIR, you get a JIT compiler backend that’s tiny and knowable — and that still punches way above its weight. There’s elegance in understanding where every byte and cycle goes.
The rebound effect of Moore’s Law is real: more resources led to more indirection, which led to more tools, which led to more churn. And now we’re entering the AI era, where tools can generate “working” code faster than we can understand what it’s really doing.
And sure, it feels productive — but something subtle gets lost when we stop thinking through the system as a whole.
We’ve outsourced understanding to the machine. Now we just hope it’s right.
The irony is that GCC improved so much since then that now the 48h may be reduced to 30h on the exact same hardware.
Is that kind of setup still usable for some kind of desktop computing or only for command line stuff ?
128MB RAM sounds huge for the early 90s - win 3.1 and word / excel of the time could fly with much less. Is the lack of hardware floating point support an issue to run modern apps ?
The speed difference with current systems is mind boggling. The original A1200 CPU is 2,000 to 5,000 times slower than a random N100 setup. one second wait nowadays means one hour delay on the A1200. This shows how much software bloat accumulated.
I also lack an internal monologue and have strong aphantasia, so the idea that I might not be conscious made me a bit uneasy—it just felt wrong, somehow. For now, the best I can say is that my worldview, which includes self-consciousness, is abstract. I can put it into words, but most of the time, it doesn’t feel necessary.
One always writes for the potential readers. Even if the human readers are becoming rare, the opportunity to be read forever by all these AIs can only boost our ego :-)
I can only recommend this game to its target audience. It's quite addictive and offers an experience I found to be quite valuable.
The English version has been crowd-translated into French, German, and Spanish, with more languages in progress, making the content more accessible.
According to https://www.bnf.fr/fr/depot-legal-pour-quels-documents, pure ebooks are not exempt; they are just archived directly from the distribution platforms by sampling.
What scares me is that some mixed works (PDF/on-demand prints) are not archived because the authors think the system is too clunky to deal with, which is really a shame for books like https://laurent.claessens-donadello.eu/pdf/lefrido.pdf, for example.
The first step to solve a problem is to measure it. Do you know of a windows program that can measure the latency of the UI of other windows apps ?
What really drives me mad is the latency of some file selection dialogs for example which can take like 10 seconds.
this is a proprietary FPGA design. The effective performance is limited by FPGA technology for now. Maybe additionnal design work would be required for an ASIC targeting latest foundry nodes.
the author considers an "operating system" as the ecosystem where the user live and find a convenient application "store". Same store = same OS. Different store = different OS.
So with this reasoning maybe soon Windows = Linux but Android != Linux.
Also the language wars are related to the platform/OS wars but are much more diverse.
the conclusion feels a little abrupt to me. What about the differences between web 1.0, 2.0, and 3.0? What about the failed W3C standards? What about the limitations of web apps compared to native mobile apps? What about the closed gardens that Facebook/Meta and newer social networks are trying to build? And lastly, countries are restricting web access to their own jurisdiction. The wars are just beginning.
This kind of publicity will only fuel the hype and the interest of ambitious people.
These "advanced AI systems" are only high end GPU systems from the shelves, or even M2 macbooks.
The software and data they handle are private. Even in legal terms, I can't see how this could be possible. Even DNA engineering could not be limited by law in practice...
Nice service idea. If you operate in the EU, the site definitely requires more legalese as this deals with real personal data : cookies popup, GDPR contact, site editor, company juridiction, the whole shebang.
Before the web and global geek culture, magazines were the only way to get a feeling of "belonging" to some kind of tech community, sometimes the community of some computer brand owners, sharing the same interest for the compatible software & hardware products. Heavily priced online board systems then milked the same needs providing downloads which replaced free disks offered with magazines.
This is not exactly in Europe, this online shop is China based and not recognized by the health care system in France.
News addiction and constant interruptions are indeed very bad for focus.
A long time ago, like in the 90s, non professionnal traders on the stock market were the first to have this kind of problem. Depending on their trading style some cope by only reading the news during the week end.
For general news, there is a lot of content that follows this weekly pattern. But for folks that really want to focus on some project, maybe need to spend some months in the jungle or focus for weeks... Maybe it would be interesting to have monthly summaries, something that can help you catch up months in mere hours.
I personally would feel very unconfortable to completely unplug for many weeks without some way to catch up.
The blog describes the technical principes but also how the optimization techniques are difficult to set up in the CI build and largely unfamiliar even to experts.
This shows how "performance focus" has become a niche. Large dev tools like Clang/LLVM set the defauts for performance opt and very few people wander beyond them.
Bisqwit's nesemu1 source is 940 lines at https://bisqwit.iki.fi/jutut/kuvat/programming_examples/nese... It also uses SDL.
NES emulation is a great low level programming project and the tradition will probably continue for some time.
I strongly agree with the point the problem was that compilers weren't using these addressing modes at all
at least in the 80s microcomputer compilers were very primitive compared to what we have now which maintained a strong need for ASM. Dev tools used to be very expensive and proprietary too.
GCC started to slowly changes that starting by 1987.
So there was a time when software started to be mainly compiled high level language but using stupid compilers and CPU designers had to live with that.