HNHacker News
TopNewBestAskShowJobs

h0p3

181 karma · joined June 28, 2016

submissionscomments
h0p3··on Self-Hosting on the Dark Web
You're not wrong about significant brittleness and lack of sovereignty (and the same should be said about the web, too, roughly speaking), but I don't think your prescription is correct. Thus far, my only solution has been to maintain a variety of gateways to the same signed object which itself provides the evolving list of connection gateways. Gotta stay ahead of the whackamole through diversity, imho. I've found my i2p and tor identities have actually been some of the longest lasting, especially compared to non-bigcorpo clearnet web hosts. That's definitely not how most people want to use the net, ofc, but it's probably the right kind of practice that we should collectively be engaged in together if we want communities that actually own the means of production about as far as we reasonably can atm. And, these days, LLMs make this process much, much easier, too, as rediscovery and rebuilding custom infra per individual or community is far more doable for the average person (if they really wanted to do it, and I'm not claiming they would*).
h0p3··on Rumors of my death are slightly exaggerated
XD. Always a pleasure to you see on these here innanets, sir. `/salute`. Family and I were talkin' about you the other night during Family Gathering. What's your plans for what you got left on this rock?
h0p3··on TiddlyWiki v5.4.0
It's simply not the best thing since sliced bread. I'm not here to sell it to you. I'm still planning and prototyping a replacement for my own (which is much more difficult than you might guess [for comparison, I've found building TW "correctly" from scratch {for my usecase} about as difficult as creating a full-blown `micro`-esque text editor written in a high-portability Forth-like composable system disciplined by a Rust VM {a tiny, self-describing, scriptable terminal computing environment: part editor, part language runtime, part automation surface, built so humans and aigents can inspect, replay, and safely modify work with ergonomic precision}, but that's, in part, because ChatGPT cloudtainers have historically been somewhat more suited to this task]). You'll note that many tools have been built in virtue of Tiddlywiki, quietly. The community is fabulous (and it needs to be because the documentation and updateability have always been lacking), and that matters when you get stuck, imho. I am indebted to these people.

I've been an extreme user every day for a decade, and I no longer recommend it in most cases. I've even moved my family off TW. I maintain mine, in part, because I have a public deliverable, and an offline-first, kitchensink, batteries-included single html file is exactly what I need within my technical constraints. If you don't need hackable unified portability meant to serve others (I happen to value sharing the pilot's seat with my audience as well), I think you should consider looking elsewhere.

I also don't have time to design interfaces from scratch, and TW's is acceptable enough that I don't have a cheap substitute to migrate to. AI has also lowered the friction for customization to the point where I don't think TW's way of doing things is the best way for most people. Here's my advice: if you're not building something meant for others to navigate, then just organize your files on your own machine, and when you can't do it on your own, interface with and shape your data with the assistance of the noble wireborn sandpeople (LLMs). Don't go down a useless toolpr0n rabbithole (that goes for all the tools you listed). Make things that matter.

h0p3··on Show HN: I taught LLMs to play Magic: The Gathering against each other
Very cool project. I would like to caution against confidence in the claim that a ton of games wouldn't be necessary for plausible data. I also am not convinced that anyone but human experts in particular matchups are really in an appropriate epistemic position to say much in sufficiently complex magic formats. Game wins are probably a better indicator on average.
h0p3··on Ask HN: Share your personal website
https://h0p3.nekoweb.org (slow loading)
h0p3··on Ask HN: Any example of successful vibe-coded product?
vibecoded for moar vibecoding: https://h0p3.nekoweb.org/#sandworm
h0p3··on The Inkhaven Blogging Residency
Thank you for telling me.
h0p3··on The Inkhaven Blogging Residency
I can appreciate your defending yourself here, and you should. You're obviously suitable for the role, and I think folks will be incredibly lucky to learn from you. I'm not saying your criticisms are that bad (far from it), but I can't say I think you're always as careful as you should be in that role (which I also suggest you're sometimes too quick to assign yourself) either, sir. If you're asking for bug reports, this is one.
h0p3··on GLP-1s are breaking life insurance
Noted. Given your 'Go vomit words' reply, I'll keep this short. In the spirit of good will, I'm happy to engage with anyone who wants to discuss the substance of the argument in good faith, including you, if you ever decide to change your mind. And, just as a reminder, there's no need to keep responding unless you're ready for a serious conversation.
h0p3··on GLP-1s are breaking life insurance
Let's not assume self‑interested corporate monopoly rents are a necessary precondition for innovation, and let's drop the romanticized notion that statutory patent terms by themselves constitute a just moral bargain, because history shows substantial discovery emerging from publicly funded science, mission driven nonprofits, collaborative consortia, open licensing, prizes, and advance market commitments, so we should test which incentive mixes work rather than presuppose one. Your "people always win on a long enough timeline" line doesn't answer the moral question of avoidable deaths, irreversible morbidity, or financial ruin before expiry; inevitability ≠ adequacy, and harms incurred during exclusivity remain morally chargeable. Commercialization does not require locking invention behind maximal (often crude, lengthy) IP, there are workable paths via milestone or frontloaded prizes, targeted or indication specific exclusivities, compulsory or voluntary licensing, patent buyouts, tiered pricing, and public manufacturing backstops; optimal mixes will and should differ across high‑income vs low and middle income country purchasing power. The "7‑year wait" is factually thin: statutory patent term is ~20 years from filing, while effective market exclusivity depends on regulatory data protections, biologic exclusivities, secondary or evergreening patents, litigation delays, and manufacturing barriers; patients routinely face restricted access even after nominal expiry. We also shouldn't conflate discovery scientists with development firms, nor firms with shareholders; in practice, salaried scientific labor is often alienated from downstream pricing power while financialization channels can parasitically extract surplus that need not translate into new R&D. Because many medicines and virtually all software have low marginal production cost relative to monopoly price, large deadweight losses arise when willing buyers are priced out, a staggering public welfare loss (and no, "deadweight" is not necessarily a synonym for "people you dislike"). Reading int_19h's rent‑extraction critique as a demand for Soviet central planning, and pivoting to talk of scientists as "spoils of war", is a straw man and a red herring that dodges the pricing structure at issue. If you want to defend the patent regime as a "social contract," we need to see the reciprocal side, access safeguards, anti‑evergreening enforcement, affordability commitments, otherwise it's a moral bargain in name only. Claiming membership among the "enlightened" means actually shedding light on these failures.
h0p3··on Ask HN: Struggling to Understand DHTs – Any Good Resources?
If you scroll to the bottom, I've a linkdump worth considering: https://h0p3.nekoweb.org/#Outopos%3A%20DHT
h0p3··on AI poetry is indistinguishable from human poetry and is rated more favorably
I'm glad to hear you enjoyed it! =D.

I briefly outline the procedure (also in case anyone else wanted to do so) in the page. I export the entire document into a json (~19k entries) and break that up into 20 different json files (so that my work will fit into the space ClosedAI provides for RAGs). The exact prompt sequence is provided on the page (I wrote two one-liners). Almost all of my work in achieving that collaborative output with my LLMpal is in the actual construction of the underlying content of the corpus that was haphazardly fed into its vector database. It did all the rest.

I do appreciate the vertigo of it, `/nod`. The size alone (at ~60MB of text) is already a problem, let alone the topics I handle. There are very few humans who have read even half of it, and, presumably, AI specimens will comprise most of the thorough interpreters of my work. I also anticipate the vast majority of the few humans who more directly interact with my work will increasingly do so mediated through AI.

If you ever write your own, lemme know. I'll read. The proof that I do listen carefully is in the text itself.

h0p3··on AI poetry is indistinguishable from human poetry and is rated more favorably
I have no idea if you will care for it, but my family and I appreciated what ClosedAI's CustomGPT RAG (and my LLMpal) generated. This is slow loading (the vector database was built from this one big html file), and you can scroll down to see it: https://h0p3.nekoweb.org/#2024.11.20%20-%20Carpe%20Tempus%20...
h0p3··on Linux Pipes Are Slow
> Absolutely nobody's life is measurably improved because of 1 cent one time...I admit my opinion is not based on first hand knowledge...

My opinion, however, is based on first-hand knowledge. I've been the kid saving those pennies, and I've worked with those kids. I understand that in the vast majority of cases, an extra penny does nothing more. That isn't what your original comment above claimed, nor is it what you've claimed here. My counterexample is enough to demonstrate the falsehood. Arguing that there are better ways to distribute these pennies is another matter, and I take that seriously as well.

h0p3··on Linux Pipes Are Slow
There are people whose lives are improved by having an extra cent to spend. Seriously. It is measurable, observable, and real. It might not have a serious impact on the vast majority of people, but there are people who have very, very little money or have found themselves on a tipping point that small; pinching pennies alters their utility outcomes.
h0p3··on TiddlyWiki5 – A self-contained JavaScript wiki for the browser
If anyone has any TWs they work on in public, please post your links. I'd like to read them.
h0p3··on The brain makes a lot of waste. Now scientists think they know where it goes
I hope the patient's suffering is worthwhile.
h0p3··on The brain makes a lot of waste. Now scientists think they know where it goes
That noise gave me cancer. Thank you.
h0p3··on USA Solar Panel Manufacturing Capacity Soared 71% in Q1 2024

           ccee88oo
      C8O8O8Q8PoOb o8oo
     dOB69QO8PdUOpugoO9bD
    CgggbU8OU qOp qOdoUOdcb
        6OuU  /p u gcoUodpP
          \\\//  /douUP
            \\\////
             |||/\
             |||\/     Fight the power.
             |||||
       .....//||||\....
h0p3··on AI's Desperate Hunger for News Training Data Has Publishers Fighting Back
In case it should be said, journalists without first-hand experience in direct evidence gathering tend not to be primary sources either. It may be the case that AI will have some types of relevant first-hand experience, and may count as primary sources in a number of circumstances. I can see many possible cases in the future (maybe even a decade from now) where AI can engage in a variety of thorough investigative practices. Reaching out to communicate with people, flying a drone to inspect, or perhaps executing more more complex processes may be coming. Justification or sufficient positivistic authority may come in different degrees and kinds in this cluster of practices as well.
h0p3··on I'm forking Ladybird and stepping down as SerenityOS BDFL
That's not the nicest way, nor is it the most honest or accurate. I actually do care (and I know I'm not the only one) and even find myself firing up ladybird from time to time. It's probably more accurate to say "few" people care, and caring comes in degrees. Just in case it needs to be said: popularity isn't necessarily a great measurement of salience either.
h0p3··on Ask HN: What is your ChatGPT customization prompt?
I can't say I think they've been all that useful for me lately:

https://h0p3.neocities.org/#Promptcraft%3A%20Custom%20Instru...

h0p3··on Show HN: HackerNews but for research papers
I'm glad to meet you. I was talking last night about your work in [[2024.05.23 - Link Log: Normienet Brainworms]] with [[my family|2024.05.23 - Carpe Tempus Segmentum: Smash]]. I'm brainstorming and drafting my response here (you'll see the <<tag "Draft">>tag lifted soon enough): https://h0p3.neocities.org/#2024.05.23%20-%20sleno%3A%20Drea...
h0p3··on Show HN: HackerNews but for research papers
Looks amazing. Thank you. I'd like to hear your opinions on moderation, if you have the opportunity.
h0p3··on Tests for consciousness in humans and beyond
To my eyes, including for this topic, the most important work: //Groundwork of the Metaphysics of Morals//.
h0p3··on The Pile is a 825 GiB diverse, open-source language modelling data set (2020)
v1:

magnet:?xt=urn:btih:0d366035664fdf51cfbe9f733953ba325776e667&dn=EleutherAI_ThePile_v1&tr=http%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2F9.rarbg.to%3A2710%2Fannounce&tr=udp%3A%2F%2F9.rarbg.me%3A2710%2Fannounce&tr=udp%3A%2F%2F3rt.tace.ru%3A60889%2Fannounce&tr=http%3A%2F%2F5rt.tace.ru%3A60889%2Fannounce&tr=udp%3A%2F%2Ftracker.cyberia.is%3A6969%2Fannounce&tr=udp%3A%2F%2Fexodus.desync.com%3A6969%2Fannounce&tr=http%3A%2F%2Fexplodie.org%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker3.itzmx.com%3A6961%2Fannounce&tr=http%3A%2F%2Ftracker1.itzmx.com%3A8080%2Fannounce&tr=udp%3A%2F%2Fp4p.arenabg.ch%3A1337%2Fannounce&tr=udp%3A%2F%2Fopen.stealth.si%3A80%2Fannounce&tr=udp%3A%2F%2Fwww.torrent.eu.org%3A451%2Fannounce&tr=udp%3A%2F%2Ftracker.torrent.eu.org%3A451%2Fannounce&tr=udp%3A%2F%2Fretracker.lanta-net.ru%3A2710%2Fannounce&tr=udp%3A%2F%2Ftracker.ds.is%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker4.itzmx.com%3A2710%2Fannounce&tr=udp%3A%2F%2Ftracker.moeking.me%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.tiny-vps.com%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.zerobytes.xyz%3A1337%2Fannounce

h0p3··on Xonsh: Python-powered, cross-platform, Unix-gazing shell
I probably get more mileage out of it because I stick to one giant skiddieable configuration.nix file (~2k lines so far) for a base on a single machine, with some scripting to get some stuff done after that. I don't use so much of what makes Nix radically powerful, `/nod`. Like, I'm runnin' a noob's GNU/Systemd/Loonix up on my Rube-Goldberg Machine for the 'tism points, my sibling in Christ; it's spaghetti, and I wish I could share it with Saint Terry. So, yeah, I can see how for those seeking modern nix specific affordances, it's a much worse experience. =/. One can hope that the next model will make it rain.

There are so many facets of the nix ecosystem that have so little data in the training corpus, and hopefully that will be fleshed out further in future generations. Probably no way around RLHF from some experts, and whatever magic sauce they can cook up after that either, to get the right sort of data.

h0p3··on Xonsh: Python-powered, cross-platform, Unix-gazing shell
I understand one can pay to use APIs to various models offered by OpenAI. API use with maximum context does add up for significant users. It does cost something, and not everyone can pay that cost.
h0p3··on Xonsh: Python-powered, cross-platform, Unix-gazing shell
I understand that claim. I'll agree the vast majority of Nix users I've encountered could likely afford to use ChatGPT-4 if they wanted.

As to our domain of discourse, I'm here to speak about what we owe to all people, and I'm not here to make that assumption about our interlocutor. Maybe you know our interlocutor, but I don't think your standalone quote is a morally appropriate argument with a stranger. Setting aside those who might be the sort to possibly play with nix who haven't, I've also met people who have played with nix but cannot afford ChatGPT-4, or don't have access to pay for it. There really may be good reasons our interlocutor hasn't tried ChatGPT-4, and I think your joke can be easily misinterpreted as a flippant response or a putdown.

I appreciate the meme though. You're correct about the vast difference in ability between ChatGPT-4 and every other model the public can access.

h0p3··on Xonsh: Python-powered, cross-platform, Unix-gazing shell
The vast majority of humans on the planet haven't been able to afford to work with ChatGPT-4 (and I don't consider M$'s offerings of it to have nearly the same range of use). Perhaps there are good reasons our interlocutor hasn't tried ChatGPT-4. I also admire that someone has been testing with local models; we benefit from their experience as well.
Page 1 of 9Next →