HN user
E-mail: kashyapc@fedoraproject.org
Not to be that guy, but the article has 14 (!) occurrences of the word "cyber". It's nauseating.
As usual, this is OpenAI trying to give themselves a backhanded compliment: "look, how dangerous our models are!"
I'll wait for someone more thoughtful than ClosedAI to comment on this complex topic.
Thanks for elaborating. I totally agree with you. I wasn't implying that funding the LHC itself was misguided, for the good reasons you articulate so well.
I know what you mean. I was recently at lunch at a (non-physics) tech conference and an ex-physicist (now a systems developer) sitting next to me responded the same.
I think (physicists in the room, please correct me if I'm making assumptions here) this is a classic "falsification" trap. Just because you can propose a hypothesis / prediction that is "falsiable" doesn't make it a worthwhile one to pursue. "Hypotheses" are dime a dozen. The art and wisdom is in choosing what to pursue.
I see; thanks for expanding on it so honestly.
felt a strong sense of frustration at how modern research in particle physics is producing papers that are all identical to each other and have no true scientific impact
Your comment above makes me think of Sabine Hossenfelder[1]. I'm sure you must've heard of her. I know she's somewhat of a "polarizing" figure. As a former "insider", do you think her core point stands? (Which is, the particle physics field has largely nothing to show for after spending billions of public money. Some particle physicists have even predicted so-called "unparticles", which almost sounds like trolling!)
I see, fair point. Sorry for taking a dig at you. Please know that I do appreciate a lot of work that you do. I was just worried for a moment when just reading that bit.
Thank you; it's an educating read for me, as someone who doesn't dwell in this space, but cares about FOSS in its true spirit.
No mention of the fact that Ollama is about 1000x easier to use.
The point of the article is not to expound on how user-friendly "Ollama" is. It's about exposing the deception and shameful moral low ground they took.
You seem to see a "pretty strong case" from a bombastic press release.
Don't get me wrong, I do know the reality has changed. Even Greg K-H, the Linux stable maintainer, did recently note[1] that it's not funny any more:
"Months ago, we were getting what we called 'AI slop,' AI-generated security reports that were obviously wrong or low quality," he said. "It was kind of funny. It didn't really worry us."
... "Something happened a month ago, and the world switched. Now we have real reports." It's not just Linux, he continued. "All open source projects have real reports that are made with AI, but they're good, and they're real." Security teams across major open source projects talk informally and frequently, he noted, and everyone is seeing the same shift. "All open source security teams are hitting this right now."
---
I agree that an antidote to the obnoxious hype is to pay attention to the actual capabilities and data. But let's not get too carried away.
[1] https://www.theregister.com/2026/03/26/greg_kroahhartman_ai_...
I recently heard a new (to me) excuse:
When in the middle of a group text-chat, someone replied with AI-generated blather. It was dead-clear with the usual sterile vocab, structured buzzphrases, and other LLM "tells".
I politely called him out and asked to use his own voice. In public he insisted that it was his voice and that he used AI only for "formatting". But in private he admits that he created a "gem to assist with multicultural comms", which generated the text. He claims he did it because "not everyone can take the native American English well". A load of bovine manure. I nicely told him to cut this crap and just write as it comes to him. (Basic spell- and grammar-check is fine.)
Well said. Some people are misparsing your core point here.
Skrebbel is largely referring to the OSS projects that need people to do consitent grunt work like shipping predictable releases, stable branch maintenance, backporting security fixes, etc. This is the kind of work maintains that the internet's infrastructure.
A bit like the Nebraska guy from the famous XKCD, dependecy: https://xkcd.com/2347/
I've found this fails on my 16 GB P550 Megrez with swap disabled but works quickly and uses maybe 50 or 100 MB of swap if I enable it.
I see, I don't have a Megrez at my desk, only in the build system. I only have P550 as my "workhorse".
PS: I made a typo above - the P550 I was referring to was the SiFive "HiFive Premier P550". But based on your HN profile text, you must've guessed it as much :)
I'm tickled pink to read this! I very much support this move. HN is one of the few internet forums I use. It'd be awful to see this riddled by bot spew.
This rule will atleast partly stem the danger of HN getting turned into what dang calls a "scorched earth" situation.
They haven't produced any chips.
Great point; I only know about MIPS legacy vaguely. As you imply, don't listen to the "hype-sters" but pay attention to what silicon is being produced.
Please skim the thread :) We've already discussed it twice. Fedora "mandates" native builds.
Build time on target hardware matters when you're re-building an entire Linux distribution (25000+ packages) every six months.
A couple of corrections (the blog-post is by a colleague, but I'm not speaking for Marcin! :))
First, we do have a recent 'binutils' build[1] with test-suites in 67 minutes (it was on Milk-V "Megrez") in the Fedora RISC-V build system. This is a non-trivial improvement over the 143-minute build time reported in the blog.
Second, the current fastest development machine is not Banana Pi BPI-F3. If we consider what is reasonably accessible today, it is SiFive "HiFive P550" (P550 for short) and an upcoming UltraRISC "DP1000", we have access to an eval board. And as noted elsewhere in this thread, in "several months" some RVA23-based machines should be available. (RVA23 == the latest ISA spec).
FWIW, our FOSDEM talk from earlier this year, "Fedora on RISC-V: state of the arch"[1], gives an overview of the hardware situation. It also has a couple of related poorman's benchmarks (an 'xz' compression test and a 'binutils' build without the test-suite on the above two boards -- that's what I could manage with the time I had).
Edit: Marcin's RISC-V test was done on StarFive "Vision Five 2". This small board has its strengths (upstreamed drivers), but it is not known for its speed!
[1] https://riscv-koji.fedoraproject.org/koji/taskinfo?taskID=91...
[2] Slides: https://fosdem.org/2026/events/attachments/SQGLW7-fedora-on-...
On benchmarks, for more precision details, I recommend the RISC-V Vector (RVV) benchmarks[1], maintained by Olaf Bernsten. He only covers the Vector stuff, but with great depth.
Arm had 40 years to be where it is today. RISC-V is 15 years old. Some more patience is warranted.
Assuming they will keep their word, later this year Tenstorrent is supposed to ship their RVA23-based server development platform[1]. They announced[2] it at the last year's NA RISC-V Summit. Let's see.
The ball is in the court of hardware vendors to cook some high-end silicon.
[1] https://tenstorrent.com/ip/risc-v-cpu
[2] https://static.sched.com/hosted_files/riscvsummit2025/e2/Unl...
Near as I know, Fedora prefers native compilation for the builds.
Your question made me look up Arm's history in Fedora and came up on this 2012 LWN thread[1]. There's some discussion against cross-compilation already back then.
Everyone is actually underestimating stickiness. The near billion users OpenAI has is actually a real moat and might translate into decent chunk of revenue.
Maybe you're overestimating their "moat" and stickiness. The dust is still settling on this madness and "OpenAI"[1] creates a lot of noise in the market.
These LLMs are being rapidly commoditized, very soon they will become as "boring" as virtual machines or containers. Altman has the exceptional skill to dupe people into giving their money to him. The "infinite money glitch" that he has been exploiting isn't really infinite.
I just hope there'll be a breakthrough with truly transparent LLMs that will stabilize this madness. As I've griped[2] two years ago, I find OpenAI too scummy, and it is unlikely that they will "win" with their sleazy ways.
[1] Air quotes because of their persistent abuse of the word "open"
Pretty sure he's not talking about the physics of stochastic chaos!
The context gives us the clue: he's using it as a metaphor to refer to AI companies unloading this wretched behavior on OSS.
what the people doing him kindness got in return was that they got to do something kind for someone
Yes, there is something to that but he gives too much credit[1] to his own "willingness to be open".
I don't know if KK has taken it too far or not, but the way he expresses it all is not inspiring. It carries an air of pretentious wisdom.
Yeah, while the article is beautifully and poetically written (he's a reputed magazine editor after all), it comes across naive to extrapolate his good fortune to "you just have to be open enough to let the miracle happen to you". (I know, I'm simplifying here a bit.)
Obviously, there's some truth to it, but there are many unspoken variables that worked in his favor that he doesn't bother to acknowledge them. Some other comments also touched on it.
I'm not being cynical here. I myself have had incredibly good fortune in experiencing the kindness of strangers, both in the East and the West, and I do my best to reciprocate. But I'm acutely aware of how invisible factors that are not in my control helped facilitate some of the good fortune that came my way. I can't merrily attribute it all to my own "openness to experience"!
KK inhales his own good fortune too deeply.
Why should we aim to reward pointless, Sisyphean tasks at the expense of actual achievement?
Of course that would be ridiculous. You're trivializing the author's point. I'm not sure you've actually read the article in full. The author admits the difficulty in measuring it and that we may have to rely on "non-scientific" measurements.
Many of the tech robber barons and VCs (who call themselves "angels") carry the air of "my winnings are entirely of my own making". They rarely acknowledge the role of good fortune (in various aspects) in any meaningful way.
They inhale their success too deeply, as Michael Sandel memorably puts it.
The article on "effortocracy"[1] is pretty very well done. Quoting the end of the article:
"... if you take anything away from this, it is to recognise that if meritocracy is based on achievement only, then we must be sure not to confuse it with effortocracy when it comes to its moral weight."
Related reading: The Tyranny of Merit, by Michael Sandel (I was hoping the article would reference this, and it does.)
The person you're quoting has a point. Everyone is losing their minds about this. Not everyone needs to be on top of AI developmemts all the time. I don't mean you ignore LLMs, just don't chase every fad.
The classic line (which I've quoted a few times here) by Charles Mackay from 1841 comes to mind:
"Men, it has been well said, think in herds; it will be seen that they go mad in herds, while they only recover their senses slowly, and one by one.
"[...] In reading The History of Nations, we find that, like individuals, they have their whims and their peculiarities, their seasons of excitement and recklessness, when they care not what they do. We find that whole communities suddenly fix their minds upon one object and go mad in its pursuit; that millions of people become simultaneously impressed with one delusion, and run after it, till their attention is caught by some new folly more captivating than the first."
— Extraordinary Popular Delusions and the Madness of Crowds
seemed to me like Op.1 b-flat minor was the one that would be overplayed
I don't watch much movies, but haven't seen Op. 9 No. 1 often in many places. I hope it remains that way! No 2 is wildly popular for its lovely long opening melody.
preludes (raindrop etc) for example and they have softer key and louder key parts and I want to blast the softer side without overblasting and distortion
Yes, the preludes are lovely. But please -- you don't need to "blast" this music. This is not rock :-) Yes, there are a lot of "dynamics" (soft and loud and some gradations: pianissimo, forte, etc), but don't overthink it.
Just use a reasonably high-end speaker (e.g. I use an old, Bose "SoundTouch 20") and pick one of the recent recordings from Deutsche Grammophon. I'm currently listening to the nocturne interpretations by Kun-Woo Paik.
My top book of the year:
Fryderyk Chopin: A Life and Times, Alan Walker
A riveting read by a legendary musicologist and biographer. Walker spent about ten years researching this. It is 700 pages, which seems daunting but he makes this authoritative bio absolutely enjoyable. It's also a "corrective biography", it dispels a lot of myths. This book is one of the best examples of accessible writing with flair. What a writer!
Throughout the book, Walker tastefully quotes musical phrases (in notation) from Chopin's works to situate them in context. I often paused reading and put on the track on a given page (nocturnes, mazurkas, preludes, etc). It made the reading experience incredibly rich and fun. Other things I enjoyed: Chopin's letters to his friends and family, life in aristocratic salons of Paris, London, Warsaw, and more—Chopin had unparalleled access. Of course, there's also a lot of gut-wrenching stuff. As the book's blurb says, it really is for both the casual music lover and the professional pianist.
If you haven't discovered them yet, give a listen to Chopin's nocturnes. But please, give them an attentive listen and play them on a high-quality audio system. Here[1] is one of his finest nocturnes (it is less famous than the "happier" nocturne that follows it, Op. 9 No. 2).
[1] Nocturne in B-flat minor, Op. 9 No. 1: https://www.youtube.com/watch?v=ThMGf07UBHQ
Yeah, I often write locally in Emacs Org-mode, generate the HTML from it, and then copy/paste that into GDocs.