Some suggestions: I know none of us like "the algorithms choosing", but I think we can do better than alphabetical order. Number of clicks you see (popularity), or number of inbound links google tells you about would be good.
I also think you've gone to great effort, but it's still very light in some categories. I hope you keep going - what's your data source? Are you tracking outbound links from the ones you have indexed to find new blogs?
Thanks for the kind words. The source is user suggestions and my own searching and browsing.
I do sort of track outbound links, so that I can show which domains a blog links to most, which can sometimes give a sense of what the blog is about.
But, while I haven’t analysed the data, I suspect the links from one blog to another would be a very tiny percentage of the overall outbound links. It’s the kind of thing that might have been more interesting/useful in the olden days of blogging when more people linked to each other, and replied to each other, via blogs rather than social media.
I like the idea of some kind of algorithm minimalism, or at least parsimony; but I also think sometimes it might be appropriate? In this case, another approach would simply be randomization, which doesn't favor any name (Aaaaaron Aaaaanderson's blog :P ), this randomization can be consistent (such that you can find something you wish in linear time).
I think equally important is algorithmic transparency, that is, that the algorithm be publicly disclosed (although I think simplicity is a component of transparency: if you just dump a huge incomprehensible algorithmic mess somewhere that's not very helpful), so that you at least know what you are getting into, and better yet have some ability to choose and make educated critique of the current state of things (i.e. does the algorithm just maximize engagement like a slot machine? or does it optimize for some kind of helpfulness?).
LLMs are good at dealing with things they've seen before, not at novel things.
When novel things arise, you will either have to burn a shed ton of tokens on "reasoning", hand hold them (so you're doing advanced find and replace in this example, where you have to be incredibly precise and detailed about your language, to the point it might be quicker to just make the changes), or you have to wait until the next trained model that has seen the new pattern emerges, or quite often, all of the above.
Apologies, but your information is either outdated from lack of experience with the latest frontier models, or you don't realize the fact that 99.9% of the work you do is not novel in all capacities. Have you only used Copilot, or something? Because that's what it sounds like. Since the performance of the latest models (Opus 4.6 max-effort, gpt-5.3-Codex) is nothing short of astonishing.
Real-world example: Claude isn't familiar with the latest Zig, so I had it write a language guide for 0.15.2 (here: https://gist.github.com/pmarreck/44d95e869036027f9edf332ce9a...) which pointed out all the differences, and that's been extremely helpful in having me not even have to touch a line of code to do the updates.
On top of that, for any Zig dependency I pull in which is written to an earlier version, I have forked it and applied these updates correctly (or it has, under my guidance, really), 100% of the time.
On the off chance that guide is not in its context, it has seen the expected warning or error message, googled it, and done the correct correction 100% of the time. Which is exactly what a human would do.
Let's play the falsifiability game: Find me a real-world example of an upgrade to a newer API from the just-previous-to-that API that a modern LLM will fail to do correctly. Your choice of beer or coffee awaits you if you provide a link to it.
I’ve been making a project in zig 0.16 with Claude as a learning experiment. It’s a fairly non trivial project (BitTorrent compliant p2p downloader for model weights on top of huggingface xet) - whenever it doesn’t know the syntax or makes errors, it literally reads the standard library code to understand and fix it. The project works too!
Tbh, while impressive that it appears to work, that guide looks very tailored to the Zig stdlib subset used in your projects and also looks like a lot more work than just fixing the errors manually ;) For a large code base which would amortise the cost of this guide I still wouldn't trust the automatic update without carefully reviewing each change.
Eh, I've had good luck with porting codebases to newer versions of Bevy by pointing CC to the migration guide, and that is harder to test than a language migration (as much of the changed behaviour would have been at runtime).
I still wouldn't want to deal with that much churn in my language, but I fully believe an agent could handle the majority of, if not all of, the migration between versions.
A great idea with quite a few bugs. I would suggest watching people who have never played this game before try and play it and you'll notice a lot of things are going wrong that you've not seen before, and you're going to figure out how to make a lot of things a lot better. Really fun idea though, I'd love to try it again when it's less glitchy and the UX has had some work.
A technical question for you around the porting being a dead end:
I see from other replies that you now understand the code reasonably well and feel you can expand/extend it while keeping it in BASIC. However, I note you've also done project where you automatically ported Fortran to Lua - are you not interested in trying to do something similar for performance/maintainability reasons? Is there an advantage in keeping it in PowerBASIC?
I've wish listed the game, and look forward to playing it, it sounds like great fun - even the manual sounds like a good read.
I did exactly that originally. But here is the reality. Michael only knows BASIC and he has felt the agency to continue patching and working on the code, adding features even, since it is still BASIC. When others tried to port it to C++ be felt like he had no agency and wasn't motivated to help. So while he's still motivated to work on it, it needs to stay in a language he can work with. And thanks for wishlisting!!!
It seems to be a one-of-a-kind simulation product that could be used as part of actual financial/trading training. There's insane value here, giving the source code away for free is absolutely not a good strategy.
PSA: 99.999% of people should ignore most of the entries on lists like this.
I devour reading material. I love books - fiction, non-fiction, audio books, trade paperbacks, newly minted hardbacks, old musty stuff in a basement, all of it - and subscribe to Literary Review and Granta, and check in on London Review of Books and The Times Literary Supplement when I can. I subscribe to quality newspapers and periodicals, and I'd rather spend an evening in a bookshop with late opening hours than a nightclub. Reading is great. Everyone should do a lot more of it - it's food for the soul.
But reading lists put together by other people aren't good for you. If anything, they get in the way of you figuring out what you want to read.
Here's some simple maths: life expectancy in my home country is 83 years for females, 79 years for males. I am male, have multiple (not imminently life-threatening), health conditions, and so with a little maths I can expect to live perhaps 25 more years. Sobering. But it is reality.
If I read a book a week (which is way higher a rate than the average reading rate, and slow for a fan of reading - but I like to absorb books a little more slowly), I am going to max out at 1,300 books in the rest of my life.
Most people read a few books a year. At that rate I'd have just 75-100 books to read in the rest of my time alive. If that were my number, I should probably make each one of those books count in some way.
You should do this maths yourself, and across a few dimensions. You only have so many books, films, music gigs, vacations/holidays, restaurant visits, whatever left in your life.
As an aside, you only have so many side projects, business ideas you'll get a chance to build and test in the market, and opportunities to invest in somebody else's ideas. You should do those maths too: figure out what your error bars could look like. They're probably not as optimistic as you'd hope for.
At first, this might feel terrifying. I prefer to see it as "focusing".
Do you really want to read all 842 of the books on that list? Is this the oeuvre you want to invest a sizeable chunk of your remaining life in? Are you confident this will make you feel whole, that you will get to the end and have no regrets about making this your mission? If you yes to all these questions, and are sure: brilliant, you have found a purpose in life few others ever will. Godspeed and good luck!
For most people though, lists like this are just another todo list that create a sense of inadequacy, FOMO or regret.
In 1880, the designer William Morris said "Have nothing in your houses that you do not know to be useful, or believe to be beautiful".
Apply this to your reading lists[0]. Curate. Edit. Find what makes your heart sing or your brain grow, and dive in.
Do not worry about what other people think you "should" read. Do not read "the great classics" if they do not interest you. Safely ignore award winning writers - from Nobel laureates, to Pulitzer Prize winners, to Booker short-listed authors - unless something about that book speaks to you and you almost yearn for it.
Because when you do that, you'll realise a) most books are junk to you (but might be great for someone else), and b) that as you start to develop the habit of reading the things that you genuinely want to, it becomes a healthy, mind-nourishing obsession.
Come on in, the pages are lovely.
[0] Actually, apply this rule to everything you can in your life. It can be hard to start, but worthwhile.
For someone who wants to sell the idea of not wasting time on reading, you're trying to waste a lot of time with your comment. You wasted mine for sure so I'll hopefully help somebody else with my short comment:
----------------
- op basically says: don't pay attention to recommendation lists
- they assume you plan on reading all books recommended there or none
- they do not present an realistic alternative to finding books you might like
--------------------
Especially the last point is what I missed from your comment because it is actually a good idea to browse through other peoples recommendations to find similarities in taste.
Like in this list, I looked at the SciFi section and didn't find Neal Stephenson. Which, for me, is a sign of good quality together with other books I've already read. There are other books I have not even heard about. I might check them out and it would be a logical thing to do. Nothing wrong about it and chances are good that I might like it. They are certainly better than a wild guess.
There are 32 sci-fi books listed. Let's suppose you have read half of them, and you read 1 book each week. You've now just got your list for the next four months. Congratulations!
But wait, there's another list out there you're going to find tomorrow. And then another list, and another, and another, and they all have this same quality of having some books you've read and like, and nothing by Neal Stephenson. Then, when you're in the book store you see a book called "100 Sci-Fi Books To Read Before You Die", and you note it has these qualities but there are 80 books in there you've not read yet.
If you're busy pulling a sub-list together, you're doing the thing I'm suggesting: you're editing and curating, not just seeing the list in its own right. I'm definitely not suggesting you take wild guesses - note the magazines I read to find my own "next thing", and even that method is problematic.
I am not reading magazines, nor do I have the time to hang around book stores which0 here. don't even have qualified personal or a significant amount of Scifi literature.
So what is there left to do but follow recommendations online realistically?
Looking through the list I've found 2 books which might interest me. I'm done with the list now and won't return. If 1 of them is great, it is already a win.
Some of us lack access to good physical bookstores which are curated and allow for casual non-biased exploration. Amazon and other digital players never picked up on curation or segmentation, their store fronts are hot messes similar to digging through a random bin of books at a second hand store. To top it off they skew your opinion with customer ratings visible next to every single title.
So a list like this is a somewhat working digital alternative to browsing books in a curated bookstore.
Unlike Krasnol, I found your comment helpful. Especially, the idea that knowing what to ignore is at least as important as knowing what's out there. Ignore the haters :)
IETF have a habit of posting "fun" RFCs on the 1st April each year. Some of them are more famous for being completely daft ("avian carriers" and climbing into trees to watch 0s and 1s painted on the top of tanks being the two stand-out ones), but it doesn't mean that everything they do on that date is to be disregarded as nonsense.
I actually think this is just computer science. Why? Because the first "computer scientist" - Alan Turing - was interested in this exact same set of ideas.
The first programs he wrote for the Atlas and the Mark II ("the Baby"), seem to have been focused on a theory he had around how animals got their markings.
They look a little to me (as a non-expert in these areas, and reading them in a museum over about 15 minutes, not doing a deep analysis), like a primitive form of cellular automata algorithm. From the scrawls on the print outs, it's possible that he was playing with the space of algorithms not just the algorithms themselves.
It might be worth going back and looking at that early work he did and seeing it through this lens.
And that's my point; it's okay to create new names for sub-disciplines, as Wolfram is doing here. Because that's what we have been doing since the days of Aristotle.
The idea iiuc, is that pattern formation in animals depends on molecules diffusing through the growing system (the body) and reacting where the waves of molecules overlap.
To me , the 1952 paper is very important, since it shows up in theoretical biology a lot. Seeing generality at all these different emergence levels is really exciting to me. (and it makes me sad when others don't see it). Can you imagine? Set up a few gradients, and now you have coordinates. Put all the bits where they're supposed to go like uhhh... GLSL sort of loosly fits. How cool is THAT?
More recently I've gotten into all sorts of debates on HN by people who like Searle. Often the argument goes "Turing is all wrong, he knows nothing about biology."
Turns out towards the end of his life he was applying his knowledge to biology. Most of which experimentally verified, besides!
(ps. just to be sure: Never wondered how DNA encodes the trick? You started out as a clump of cells, all the same. How did one part decide to become the tip of your nose, and the other the tips of your toes? Segmentation controlled by Turing patterns all the way down!)
Not quite. A formal system is a system of syntactic rules defined over an alphabet of symbols. They can be mechanized in principle. Peano arithmetic is one example.
A „logical” semantics can be assigned to such a formal system, but it is not a necessary entailment of the syntax, even if such systems are typically motivated by particular semantic models. Model theory might examine how the same formal system affords different interpretations.
Such syntactic systems have computational properties, and it is how computer science kicked off historically.
> Formal Systems is the study of logical systems themselves. Ruliology is a study of what actual systems do.
Assuming that you mean the same thing by "logical systems" and "actual systems", then Ruliology must fall under Formal Systems as a sub-discipline? Since studying "what these things do" is a subset of studying "these things themselves". And grounded on it.
If not, then what's the difference between "logical" and "actual" systems?
It's worth pointing out that in France and the UK, the authorities involved are arms length independent of the political bodies - it's not like the US where if you give the President good vibes you can become head of the FBI, and all you have to do in return is whatever he says. There are statutory instruments (in France, constitutional clauses), that determine the independence of these authorities.
They are tasked - and held to account by respective legislative bodies - with implementing the law as written.
Nobody wrote a law saying "Go after Grok". There is however a law in most countries about the creation and dissemination of CSAM material and non-consensual pornography. Some of that law is relatively new (the UK only introduced some of these laws in recent years), but they all predate the current wave of AI investment.
Founders, boards of directors and their internal and external advisors could:
1. Read the law and make sure any tools they build comply
2. When told their tools don't comply take immediate and decisive action to change the tools
3. Work with law enforcement to apply the law as written
Those companies, if they find this too burdensome, have the choice of not operating in that market. By operating in that market, they both implicitly agree to the law, and are required to explicitly abide by it.
They can't then complain that the law is unfair (it's not), that it's being politicised (How? By whom? Show your working), and that this is all impossible in their home market where they are literally offering presents to the personal enrichment of the President on bended knee while he demands that ownership structures of foreign social media companies like TikTok are changed to meet the agenda of himself and his administration.
So, would the EU like more tighter speech controls? Yes, they'd like implementation of the controls on free speech enshrined in legislation created by democratically appointed representatives. The alternative - algorithms that create abusive content, of women and children in particular - are not wanted by the people of the UK, the EU, or most of the rest of the World, laws are written to that effect, and are then enforced by the authorities tasked with that enforcement.
This isn't "anti-democratic", it's literally democracy in action standing up to technocratic feudalism that is an Ayn Randian-wet dream being played out by some morons who got lucky.
> It's worth pointing out that in France and the UK, the authorities involved are arms length independent of the political bodies
As someone who has lived in (and followed current affairs) in both of these countries, this is a very idealistic and naïve view. There can be a big gap between theory and practice
> There are statutory instruments (in France, constitutional clauses), that determine the independence of these authorities.
> They are tasked - and held to account by respective legislative bodies -
It's worth nothing here that the UK doesn't have separation of powers or a supreme court (in the US sense)
European courts have repeatedly said that in France the procureur (public prosecutor) isn’t an “independent judicial authority”.
The European Court of Human Rights has reminded this point (e.g. 29 Mar 2010, appl. no. 3394/03), and the Court of Justice of the European Union reaches a very similar conclusion (2 Mar 2021, C-746/18): prosecutors are part of the executive hierarchy and can’t be treated as the neutral, independent judicial check some procedures require.
For a local observer, this is made obvious by the fact that the procureur, in France, always follows current political vibes, usually in just a few months delay (extremely fast, when you consider how slowly justice works in the country).
There's a dissonance between the claim that it's "independent of the political bodies" and "held to account by respective legislative bodies". Are the legislative bodies not political? In the sense of aren't they elected through a political process?
This is kind of a genuine question from me since I have no idea how these authorities are set up in France or the UK...
Paul I have a 200_000+ word counter with maybe 4-5x ish less karma than you. So I know my karma/word counter would come out even lower.
That's really not why I created it. A bit of it is if you might feel competent that you have written so many words in a single place. You know how to put in the efforts so you can have a blog post and create it there and create absolute gems too :D
And also Hackernews is pretty chill community. This is probably the reason why some of us are able to write so much in the first place. Personally, Hackernews has improved my grammar & gave me confidence (sort of) with interesting people to follow/learn from. I do feel like its one of the most fine places to spend time on (atleast on the internet) usually/for the most part.
I'd rather spend my time thinking about the problem and solving it, than thinking about how to get some software to stochasticaly select language that appears like it is thinking about the problem to then implement a solution I'm going to have to check carefully.
Much of the LLM hype cycle breaks down into "anyone can create software now", which TFA makes a convincing argument for being a lie, and "experts are now going to be so much more productive", which TFA - and several studies posted here in recent months - show is not actually the case.
Your walk-through is the reason why. You've not got magic for free, you've got something kinda cool that needs operational management and constant verification.
I’ve seen otherwise intelligent and capable people get so addicted to the convenience and potential of LLMs, that they start to lose their ability to slowly go through problems step by step. it’s sad.
Agreed. My work is mandating Claude Code usage this week for everyone. I spent all day today getting it to write tickets, code, and tests for something I knew how to do. I don’t understand the appeal. Telling the AI “commit those changes and then push,” then waiting for the result, takes way longer than gcmsg <commit msg> && gp.
Some suggestions: I know none of us like "the algorithms choosing", but I think we can do better than alphabetical order. Number of clicks you see (popularity), or number of inbound links google tells you about would be good.
I also think you've gone to great effort, but it's still very light in some categories. I hope you keep going - what's your data source? Are you tracking outbound links from the ones you have indexed to find new blogs?