OK, so the "Storing data in the network ... " title made me remember something.
If you transmit a message to Mars, say a rover command sequence, and the outgoing buffer is deleted on the sending side (the original code is preserved, but the transmission-encoded sequence doesn't stick around), then that data, for 20-90 minutes, exists nowhere _except_ space. It's just random-looking electrical fluctuations that are propagating through whatever is out there until it hits a conducting piece of metal millions of miles away and energizes a cap bank enough to be measured by a digital circuit and reconstructed into data.
So, if you calculate the data rate (9600 baud, even), and set up a loopback/echo transmitter on Mars, you could store ~4 MB "in space". If you're using lasers, it's >100x as much.
During NASA's Deep Space Optical Comms demo (https://www.nasa.gov/mission/deep-space-optical-communicatio...), they transmitted video at 267 Mbps from 16 million kilometers away. That's 1.78 GiB stored in space while in transit (assuming 53.3 seconds light-speed delay).
The furthest they did was 8.3 Mbps at 400 million km which is around ~1.38 GiB in transit.
You could totally do that with the mirror on the moon. (Retroreflector + optical data transmission).
The moon is approximately (it varies) 1.3 light seconds away, i.e. a 2.6 second round trip, and optical links can have very high data rates. You could fit quite a lot of data on there! (Edit: although maybe the data rate won't be so high at these distances)
People of Earth. I AM LRRR, RULER OF THE PLANET OMICRON PERSEI 8! We will raise your planet's temperature by one million degrees a day, for five days, unless we see McNeal at 9pm tomorrow - 8 central!
Before I consumed calories over days to figure out syntax. Now, a language model exhausts those calories away in seconds. Eventually we will advance too far into the future that the tail end of humanity will forget how to make pants.
That's only true in classical electrodynamics, as it happens. If you're in a very strong B-field like you must find near a compact object you'll get nonlinear QED effects.
I have a few blog posts which have received only about ~250 upvotes across different communities, plus a GitHub project with just 30 stars.
Still, both of these were really interesting to my future colleagues (not the recruiter) who interviewed me in the last round of the interviews which landed me my current job. They had read them ahead of time and it really shaped the technical part of the interview.
maybe not the recruiter but the hiring manager or prospective colleagues who'll interview you later?
not the number of stars, but I like looking what people have done online ie GitHub/blog. I feel like it is a nice thing to talk about.
I know it's an unpopular opinion these days cause everyone wants work life balance and not work beyond the office but it's always nice to see projects you've worked on it does show some interest. also while one can fake GitHub activity it's hard to fake well thought out and cared for projects.
it's easier to fake metrics from your previous jobs like I saved X amount of money for the company or had Y efficiency gains.
I hired many many people and never once I cared about GitHub stars. Not even sure what signal it suppose to be. Are there jobs for tech influencers or something?
-> Rich developer spends $15k to run a model slightly faster.
I love these and I know this is all in good fun, but I feel like this one is a little unfair to Jeff. He's a content creator and he didn't actually buy the rig. If he's rich it's because he creates content like this.
Most of these are unfair in some way and many are wrong. What makes this funny is precisely that it has more snark than is reasonable (and often pushes bad assumptions as snark usually does!)
I love asking Grok’s companions, especially “Bad Rudy”, for the news of the day. It’s pretty similar: Brutally honest, filtered news. Although recently he started editorializing with his personal opinion, which is boring (from an AI companion).
Love this. can we get an honest title for this entry too? (I'm not quite happy with my 11l+ karma, please give me some upvotes so I can start the new year with a smile?)
jk, great one, cheers
I got a good chuckle out of some of the titles. In Jeff Geerling's defence (the title on the site reads "Rich developer spends $15k to run a model slightly faster"), he was loaned the Mac Studios from Apple and so he didn't spend a dime.
Also his accompanying YouTube video mentions the kit retails for $40,000+, a far cry from $15k.
Its not even just Rails ppl. In embedded ive seen so many consultants say things like, "no problem. I just started working on the ultimate, perfect way to set up a state machine." Confidence theater
I sleep better at night thinking it is just a battleground of astroturfing bots fighting each other (at least on the main pages).
Everything from massive Russian state-actor bot farms testing newly trained LLMs popping out AI-generated meme formats before deploying domestically unknowingly getting into arguments with Israeli bot farms trying to raise support for some new movie series that will enable them to raise money for their next missile strike competing for eyeballs/attention from some uni student in a dorm room paying mid-sized black market companies in India to post comments telling you that cast-iron pans are too hard to clean so you should buy the non-sticks you saw on instagram (which are just marketing dropshippers in the USA selling the QA rejected pans from established brands).
To answer my question myself I gave Microsoft copilot this prompt:
I want you to rewrite this headline "Amazon will allow ePub and PDF downloads for DRM-free eBooks"
into something a little humorous and snarky that reveals the underlying truth that would bring a
wry smile to tech-engaged but big tech-skeptical hacker news readers.
This has to fit in the 80 character limit for Hacker News so keep it appropriately short.
Also I want you to reply with exactly one headline and not anything else so I can use your output
as part of a processing pipeline
and i get the response
Amazon Finally Remembers eBooks Aren’t Supposed to Be Prisoners
which I think is great. I started with the first paragraph and got something too long with some explanation. I added the second, and got three replies and more explanation. The three replies were all "good enough" in my mind but added the third paragraph to control the output.
I prompted Gemini to tell me how to prompt itself to get similar results on other news sites and it said I should give it a description of the intended audience and what it finds funny/snarky.
It would be a really interesting feature to have ai analyze the articles and write an actually honest sub-headline. (ie not these sarcastic humor titles)
I am fine with the current layout, but I also have to say that I
preferred old.reddit.com as a layout base (the new reddit UI is horrible, and reddit overall succumbed to willy-nilly tyranny of moderators on power-trips). I am not saying HN should change to become like old.reddit.com in the UI, mind you, but a few things could perhaps be considered. Using old.reddit.com was much more efficient to me than the default UI here. It is not the end of the world, but I would not mind small, slight, modest improvements to the UI (not only the front page, but all of HN).
Perhaps HN could make a few suggestions and changes and people could vote. It should be as conservative as possible, though, because while I preferred old.reddit.com, I also think that not everyone may prefer changes. So one should aim for the highest acceptance value possible, before making any change.
Anyone want to try a prompt injection? All we need to do is to get one or two story in the front page that have a good < 80 characters prompt injection.
"Show HN: I implemented generics in my programming language"
does not deserve the roast
"I built a language nobody will use just to learn generics"
It's not fair to assume the author didn't know how to implement generics before this project. It's also not fair to assume the project won't gain traction. Zig and Rust started out small too! This just goes a little too far for my tastes.
>It's not fair to assume the author didn't know how to implement generics before this project
Yeah... what they ended up implementing is not generics. So good thing the LLM doesn't read link/comments too or will've probably wrote an actual roast.
>It's also not fair to assume the project won't gain traction
Very fair to assume this. Referencing Rust/Zig disregarding the thousands other now abandoned ones is survivorship bias. Most small hobby projects remain small. But, besides joking about it, "built [something] nobody will use", if is in their free time, and enjoy it, does it matter? Is there a need for all hobby projects to have a goal of making it big?
>This just goes a little too far for my tastes.
But the "Please star my repo so I can get a job" is fine?
I'm not even so sure it's such a useless joke. I mean, it is, and I wouldn't want titles to be like "Academic publishers admit paywalls were a scam all along" (unless ALL major publishers actually admit it, which so far they didn't). But I clicked on "Math nerd explains how to spend 3 days proving 1+1=2" and when it turned out to be a Lean tutorial I thought "Oh, that's exactly what I wanted!". I don't know why, but "From Zero to QED: An informal introduction to formality with Lean 4" I didn't even notice. It's such a boring and verbose title with lame attempt at wordplay that, that my brain somehow filters it out.
11/10 would read. So much clickbait going around (and lets ignore the articles that "magicly" are upvoted but strangewise have no comments whatsoever.... not sus at all....
Seconding a little, perhaps dim button to toggle the original. But I love this. So much so that I might start referring to it more than HN when I'm in a rush.
Definitely fun, although after recently submitting one (a simple browse extension to make HN Christmas colors last all Christmas season instead of on Christmas Day)[1] that got very little attention I started looking at other posts and found a whole lot more slip through the cracks than I would have thought.
(If you follow that link from HN, and the site sees an HN `Referer`, it will do a fake captcha load, so then click "HACKERNEWS" in the navbar on the right.)
The last post or so sounded stressed. I hope they feel better.
But in general, going to read a little n-gate was a relief when some HN comment thread went off the rails. Someone else could rant about the dumbness, and a burden was lifted.
That’s it. It singlehandedly sold the idea of an AI browser to you. Like I now want an AI radio in my car, and we’re all putting AI between Google and us because Google’s results unfiltered are bad.
ok but how does it work though? Is this seriously just passing the titles to some llm with a prompt like 'roast this'? is it reading the actual content of the link as well?
This is really awesome, I am interested how you made this, is there a way that we can have something this like for hackernews for more than this one instance of (20?) posts, I know its satirical but I really enjoyed it
Considering its hosted on github I think that it is a static page
This was a great way to start the day over a cup of coffee, sometimes we need things that make as laugh but what is awesome is the titles are spot on. Thank you for making this Friday morning fun
maybe so, though an inaccurate claim. the ai is the value add here (and quite a value add based on the other comments in the thread). we typically reserve the word "slop" for ai generated content that is of low quality or no value add. this website seems to be both of quality and value ad and it would be difficult to argue otherwise.
"Slop" is at _least_ as fair a description of "we had an LLM rewrite HN headlines" as "we rewrote it in Rust so you have to upvote it" is of "we removed our biggest source of crashes on Android by getting rid of Go FFI issues."
Entertaining and apparently useful, though of course not infallible. Given https://github.com/DGoettlich/history-llms it yields the title "Training AI on 1913 data to avoid 'woke' bias (and hygiene)". That the Honest Hacker News AI model has been trained on a dose of cynicism and intellectual dishonesty is probably hard to avoid...
Yup. And if you dared to bring this up in the comments (ie. your own rewrite of a title/post), you’d get reminded of the guidelines and downvoted/flagged. Because fuck honesty - we are here for clicks and engagements.
This is a good step. Next: disclose financial incentives and other motives just to nip it in the bud.
well, I think OP is quite funny and I really enjoyed it, but it definitely goes against the entire idea of approaching things in good faith. I'm sure some or even many of them are sadly accurate, but if reinterpreting things people say through that lens became the behavioral norm on HN I think it would quickly destroy everything many people love about this place. Just my 2 cents of course.
It's funny, the OP doesn't have an ulterior motive, and it's close to the holidays so it is not cannibalizing more important news. There's no harm here.
I feel that my life has been improved by all three of these. I hadn't seen the "hysterical clickbait" one before you pointed it out, so thank you even though clearly that was the opposite of your intent.
> How much of this navel-gazing junk do we need? See also, from the same author:
Seriously! I'll admit the first post was mighty fun. But now this is turning into an AI-spam-fest! I objected in the 2nd thread but got downvoted. Apparently the community here thinks this kind of low effort Reddit-style humor is now on-topic for this place!
Not to mention the systematic downvoting of every comment that is critical of these spam posts!
reminds me of how people used to shove autotune into anything and people lapped it up like the slop that it was and this is. but, as with that slop, this will also get boring to the masses. there's only so much "I told an llm to pretend it was deadpool by way of ryan reynolds" that people actually like. the novelty is the brunt of it. and, like with autotune, when used well, people will continue to appreciate it. just ride out the hyperslop, for now.
This is not "honest", this is mostly just dismissive. The headings are no more neutral and explanatory than the originals, because, I suppose, the intent was just having fun.
Its on the front page, that means it atttracted attention and was upvoted. If what you are saying was true, these posts would die very quickly and we would never see them.
I guess if everyone thinks mocking peoples' projects and efforts is funny, it's okay!
My opinion is a weakly that this is tiring and borderline insulting to people who are genuinely looking for feedback and community. Clever once a year or so, but the creator has leaned into it and posted a lot of meta in a small timeline.
If you transmit a message to Mars, say a rover command sequence, and the outgoing buffer is deleted on the sending side (the original code is preserved, but the transmission-encoded sequence doesn't stick around), then that data, for 20-90 minutes, exists nowhere _except_ space. It's just random-looking electrical fluctuations that are propagating through whatever is out there until it hits a conducting piece of metal millions of miles away and energizes a cap bank enough to be measured by a digital circuit and reconstructed into data.
So, if you calculate the data rate (9600 baud, even), and set up a loopback/echo transmitter on Mars, you could store ~4 MB "in space". If you're using lasers, it's >100x as much.
The furthest they did was 8.3 Mbps at 400 million km which is around ~1.38 GiB in transit.
Tom 7 did something reminiscent of this if you hadn't seen already: https://youtu.be/JcJSW7Rprio.
[1] https://en.wikipedia.org/wiki/Delay-line_memory
The moon is approximately (it varies) 1.3 light seconds away, i.e. a 2.6 second round trip, and optical links can have very high data rates. You could fit quite a lot of data on there! (Edit: although maybe the data rate won't be so high at these distances)
made my mind tickle for quite a while
https://youtu.be/a5hOmPdxw0U
https://en.wikipedia.org/wiki/Delay-line_memory
archive.space
You just need to be traveling faster than the radio waves, catch up and enjoy :)
Don't you worry!
AI rover robots are soon going to dominate Mars.
The big limiting factors are free space path loss and noise.
You can get a low order correction with Euler-Heisenberg: https://en.wikipedia.org/wiki/Euler%E2%80%93Heisenberg_Lagra...
- Amazon finally adds a feature that has been standard since 2005
- Texas accidentally does something good for privacy
Would it possible to add a feature where hovering over a title displays the original title?
Still, both of these were really interesting to my future colleagues (not the recruiter) who interviewed me in the last round of the interviews which landed me my current job. They had read them ahead of time and it really shaped the technical part of the interview.
not the number of stars, but I like looking what people have done online ie GitHub/blog. I feel like it is a nice thing to talk about.
I know it's an unpopular opinion these days cause everyone wants work life balance and not work beyond the office but it's always nice to see projects you've worked on it does show some interest. also while one can fake GitHub activity it's hard to fake well thought out and cared for projects.
it's easier to fake metrics from your previous jobs like I saved X amount of money for the company or had Y efficiency gains.
Edit: Oh no, that was for the repo I actually stared before seeing this. I'm just learning Go :)
I love these and I know this is all in good fun, but I feel like this one is a little unfair to Jeff. He's a content creator and he didn't actually buy the rig. If he's rich it's because he creates content like this.
I don't know chief, have you seen how many rpis this guy has?
Love these things. Every time someone has posted an AI-flavor of HN it's been comedic gold.
I found myself pulling up the original and the honest versions side by side. The translation makes it funny.
Also his accompanying YouTube video mentions the kit retails for $40,000+, a far cry from $15k.
Plus some of the stories seem to be a bit old like openai board controversy remark.
All in all, some funny stuff i agree!
Laughed so hard on this one.
Now you know why HN has the "no editorializing" rule. :)
Everything from massive Russian state-actor bot farms testing newly trained LLMs popping out AI-generated meme formats before deploying domestically unknowingly getting into arguments with Israeli bot farms trying to raise support for some new movie series that will enable them to raise money for their next missile strike competing for eyeballs/attention from some uni student in a dorm room paying mid-sized black market companies in India to post comments telling you that cast-iron pans are too hard to clean so you should buy the non-sticks you saw on instagram (which are just marketing dropshippers in the USA selling the QA rejected pans from established brands).
The online world is a wild place.
https://web.archive.org/web/20000302102827/https://suck.com/...
This one shows the "age" of the LLM, or the data cut off time
i am so confused, whats the reason behind this little event handler?
Which looks like what you did.
Perhaps HN could make a few suggestions and changes and people could vote. It should be as conservative as possible, though, because while I preferred old.reddit.com, I also think that not everyone may prefer changes. So one should aim for the highest acceptance value possible, before making any change.
does not deserve the roast
"I built a language nobody will use just to learn generics"
It's not fair to assume the author didn't know how to implement generics before this project. It's also not fair to assume the project won't gain traction. Zig and Rust started out small too! This just goes a little too far for my tastes.
Yeah... what they ended up implementing is not generics. So good thing the LLM doesn't read link/comments too or will've probably wrote an actual roast.
>It's also not fair to assume the project won't gain traction
Very fair to assume this. Referencing Rust/Zig disregarding the thousands other now abandoned ones is survivorship bias. Most small hobby projects remain small. But, besides joking about it, "built [something] nobody will use", if is in their free time, and enjoy it, does it matter? Is there a need for all hobby projects to have a goal of making it big?
>This just goes a little too far for my tastes.
But the "Please star my repo so I can get a job" is fine?
on point
When you developer market hard enough that you make it into the LLM training data.
They're a lot of fun! And super easy to vibe code, if I'm looking to test a new model.
[1] https://news.ycombinator.com/item?id=46266496
Who unfortunately stopped posting HN critiques, a few years ago. But you can still read old posts on: http://n-gate.com/hackernews/2021/07/
(If you follow that link from HN, and the site sees an HN `Referer`, it will do a fake captcha load, so then click "HACKERNEWS" in the navbar on the right.)
But in general, going to read a little n-gate was a relief when some HN comment thread went off the rails. Someone else could rant about the dumbness, and a burden was lifted.
An opinionated, tuneable, reader-agent.
Considering its hosted on github I think that it is a static page
My favorite is the link in the footer:
LOL ... and it actually ran slower.
http://n-gate.com/
EDIT: open the link manually, they put a mock "security check" on referrers from HN
“Click to keep avoiding work …”
Good LLM prompt, excellent understanding.
Still pretty funny tho, ngl.
This is a good step. Next: disclose financial incentives and other motives just to nip it in the bud.
I’m all for prefacing each post that comes from a16z with “Asshole Alert” so that we know who we are dealing with upfront.
Show HN: Gemini Pro 3 imagines the HN front page 10 years from now https://news.ycombinator.com/item?id=46205632 (10 days ago)
Show HN: Hacker News, but every headline is hysterical clickbait https://news.ycombinator.com/item?id=46324579 (4 hours ago)
META-MELTDOWN: WE BROKE HACKER NEWS WITH THIS ONE SIMPLE TRICK (dosaygo-studio.github.io)
Superficially, they're the same, but digging in shows the real difference.
Seriously! I'll admit the first post was mighty fun. But now this is turning into an AI-spam-fest! I objected in the 2nd thread but got downvoted. Apparently the community here thinks this kind of low effort Reddit-style humor is now on-topic for this place!
Not to mention the systematic downvoting of every comment that is critical of these spam posts!
They're a lot of fun! And super easy to vibe code, if I'm looking to test a new model.
It's hard to restrain myself from navel-gazing, the lint in there is fascinating.
I'm not sure they satisfy curiosity as much as many posts with fewer votes, but that's okay.
"We rewrote it in snark so you have to upvote".
The comments make it clear that the language author has not yet learned generics by this exercise.
> We rewrote it in Rust so you have to upvote it
I'm pretty sure they didn't go through all the trouble of rewriting it in Rust to get some internet forum points!
Maybe its just you who doesnt like them?
My opinion is a weakly that this is tiring and borderline insulting to people who are genuinely looking for feedback and community. Clever once a year or so, but the creator has leaned into it and posted a lot of meta in a small timeline.
Obviously it's just me who doesn't like them. What's your point?
Your question was "Is it wrong, though?" The answer is "Yes"