Hacker Newsnew | past | comments | ask | show | jobs | submit | david_draco's commentslogin

Chinese Foreign Minister Wang Yi visted Europe and openly said to Kaja Kallas, the European Union's foreign-policy chief and other EU ministers: the Chinese Communist Party (CCP) does not want to see Russia defeated in Ukraine, fearing that the U.S. would then shift its full attention toward Beijing.

https://www.economist.com/international/2025/10/28/china-is-... https://www.bbc.com/news/articles/clyxk4ywppzo



Very interesting to see "Technical Analysis" in this list. I'm no expert in the field, and TA always seemed like quackery to me, but I suspect many more people believe in it than for example Cryptozoology. I personally know someone who even took a course in TA, couldn't imagine anyone taking a course in looking for Bigfoot.


Technical Analysis is a bit of a mixed bag. Some parts are fairly mainstream like saying there's a bull market in tech stocks is essentially part of it. On the other hand a lot of it is like tea leaf reading.


Could it be that the course instructor was just grifting suckers?


A browser using 32 GB for 8 tab isn't necessarily wasteful, there is a lot of caching one can do to make back/forward buttons extremely fast. It would be wasteful to not allocate that memory if it is otherwise free and unused. Usually browsers use up what is available. The more interesting metric would be whether starting another high-RAM application will lead the browser to give up most of the 32 GB.


And if I don't know, can I know?


Hastie and Tibshirani wrote a famous book on ML (https://hastie.su.domains/ElemStatLearn/), and extended GLMs into GAMs: https://en.wikipedia.org/wiki/Generalized_additive_model


I'm surprised Firefox Multi-Account Containers isn't mentioned. Seems ideal to me to keep Web Universes separate.


Unit and Integration testing is great for decreasing cognitive load too. When you are staring at an error stack trace of a complex code base, and go through mentally what could have played out to cause this, it's great to have confidence in components due to testing. Hypothesis/QuickCheck is allows dropping entire classes of worries.


Yes, exactly! If you can trust that each unit is working in all the ways covered in the tests, you can focus on the unit you are developing and not have to keep the other unit in your mind while working on it. And if there is an untested edge case you think might be the cause, you can test that edge case independently from the unit where the issue is occurring, and confirm if it produces the expected result.

Good, trusted unit tests are the difference between encapsulation reducing or increasing/complicating cognitive load. (And similar but between components for integration tests).

That being said, there will be rare times that the issue is due to something that is only an edge case due to an implementation detail several units deep, and so sometimes you do still need the full picture, but at least it lets you save doing that until you're stumped, which IMO is well worth it if the code is overall well-designed and tested.


The words "good" and "very bad" indicate that the world is less important to that person than themselves. I'd be okay with a bit of personal harm if it helps against climate change.

Ultra-processed food does not have an agreed-upon definition, and is the new "junk food" with the pretense of being more scientific. Is bread and pizza ultra-processed food? Studies do not agree on their definitions, sometimes including ingredient lists, sometimes not, sometimes it is required that the product is made in small shops with love and not in large factories. The mechanism of how ultra-processed food are supposed to cause harm remains undefined.


> Ultra-processed food does not have an agreed-upon definition

The United Nations Food and Agriculture authority have designed the NOVA classification of food[1, 2], which includes ultra-processed food as a category.

[1] https://openknowledge.fao.org/server/api/core/bitstreams/527...

[2] https://en.wikipedia.org/wiki/Nova_classification



I read "cancer" in between the lines of that comment. So the characterization of (potentially) that backdrop as "a bit of personal harm" feels wildly overassuming.


If one is truly worried about both, they don't have to eat beyond meat though. They can eat rice and beans. Eating rice and beans instead of both conventional and beyond meat is bad for beyond meat, too, I guess.


Ultra processed food is food you wouldn't be able to make at home from whole ingredients. It's easy to make bread and pizza.


I might be able to figure out how to grind wheat into flour for bread. Maybe I can squint hard enough to consider baking yeast to be a "whole ingredient". But cheese? I assume I can probably figure it out with the internet, but it is not at all obvious what goes into that. And the milk I would use almost certainly went through an industrial sterilization process that I know I am not equipped to so.


You can make ricotta in <1h with whole milk, vinegar and a bit of salt. And it's good on pizza!

But most "regular" cheeses like Swiss cheese also need rennet, ie. you need to slaughter a calf and scrape its stomach lining. You may want to make sure your downstairs neighbor is OK with the procedure before you start (offer them a veal dinner to make up for the noise?). Other than that, it's basically (unpasteurized) milk, salt and water. And time.

Yeast: take a sourdough baking class. You just need air, water and (organic) flour.


Your ignorance of the process or recipe of a food product doesn't affect the definition of ultra processed food. No amount of knowledge will let you make something like ultra processed foods at home with home equipment simply because it uses industrial processes and ingredients. Naturally there is a spectrum of processed-ness.


And yet, pizza regularly appears on lists of "ultra processed" foods. As do potato chips and ice cream, two foods that are also very easy to make at home from whole ingredients.

There is no consistent definition and people regularly bend over backward to put all "junk food" in this category.


Yeah, it's a good example of how useless "ultra processed" is as a heuristic when we can use a slightly better label like "junk food".

So, donuts are fine because they are only a few ingredients that you can make on your stove, and they're bad once a factory makes them? Maybe only because the factory uses "chemicals"?

No, it's the fried calorie-dense food that is easy to overeat while displacing nutrition from better food sources that is the problem.


Everyone here seems to be avoiding the point that ultra processed foods contain ingredients that home bakers would never use: preservatives, anti-caking agents, flavor enhancers, artificial colors and flavors. Ingredients that are not food and add little to no nutritional value.

Pizza made at home will not use such things. Your local pub that makes their own pizza will not either. Fast food or frozen pizza gets their ingredients from central suppliers in bulk, and they have no choice but to use such things in order for their products to survive the extended storage, processing, transportation, and similar delays that will occur on the way to the consumer.


It's a very convenient red herring to zoom in on some additives instead of zooming out to evaluate your dietary patterns.

Probably because we can use it to let ourselves off the hook for a bad diet. We can do things like roleplay that it's the seed oil in our Doritos making us fat, and that if it were butter then, idk, it would be a superfood or something?

It's the pepperoni, 15g sodium, 100g saturated fat, and 3000 calories of Costco pizza you just ate that's doing a number on your body, not the guar gum in the dough.

Or, how are you going to pick apart an ingredient list when you just ate a half-dozen home-cooked lard donuts? You're cool with laying down arterial plaque but you draw the line at ascorbic acid in the store-bought cream filling?

The "ulta processed" meme is a huge distraction. It's like listening to a fat guy talk about how he's very particular about the gum he chews because he stays away from "sugar alcohols". Yeah? What about the other 4999 calories of food you ate today?


> Everyone here seems to be avoiding the point that ultra processed foods contain ingredients that home bakers would never use: preservatives, anti-caking agents, flavor enhancers, artificial colors and flavors. Ingredients that are not food and add little to no nutritional value.

Precisely what is wrong with flavor enhancers? A common flavor enhancer is MSG, and using that in homemade dishes would not be that unusual. I frequently use it in many home preparations that could use more savory flavor.

Likewise with thickeners or emulsifiers such as cornstarch, xanthum gum, guar gum, etc: these are often used in many preparations at home. Just because something has 'no nutritional value' doesn't mean it doesn't have culinary value. By this same logic, spices have no nutritional value and are just flavors, which clearly doesn't pass the sniff test.


Sometimes. But if you look at papers, media coverage, and policy proposals you'll find that "has preservatives and stuff" is not actually a necessary nor sufficient requirement.


This is not true. I frequently use these spooky "ingredients" in home cooking. I use sodium citrate to make cheese sauces that don't coagulate. I use MSG if I need a source of glutamic umami. I've used various gums as thickeners.

These aren't some toxic compounds that machines put in our food. You can just go to a food supply store and grab them. MSG is just available pretty much everywhere.


You can debate semantics / definitions or make an assessment and get most of the benefit.


He is saying that the assessment is wrongly done.


I was extremely dismayed when that supermarket simulator game that got popular on Twitch called 'pizza' something along the line of 'frozen dessert pie'...

At least the way it tends to get made in the US, a sugary pastry that's stuffed full of sugar, carbs, fats and cheese? Ok yeah, my favorite foods are _all_ terrible for me and I can't eat them anymore. This makes me very sad.


Why not make pizza at home?


Pizza at home is my pet peeve. Lots of work working the dough, lots of waiting, super messy. Needs expensive oven and tons of electricity. All of this work to get one of the cheapest meal available.

Now compare to a steak - add salt, 5 minutes on pan, rest. Better than $50 steak at most restaurants.


I get that pizza at home is a whole hobby, yes. And if you don't want to do it, it's all of those things that you say.

1) My pizza oven runs on gas not electricity. Not that this is better environmentally. Some run on wood.

2) I'm getting better results than I can get in a local pizza place. Cheap pizza is not great pizza. Home pizza making has a learning curve, it's more of a niche thing than e.g. cooking a steak or burger.

For grandparent post, Pizza is of course not a "dessert", it's a savoury main course. Full of white flour, cheese fats, and salt. So also not health food.


One does not need to eat these "ultra-processed" foods when reducing meat consumption.


> The words "good" and "very bad" indicate that the world is less important to that person than themselves. I'd be okay with a bit of personal harm if it helps against climate change.

Yeah, no shit? We're not ants in a colony. I think you're pretty stupid if you're alright with harming yourself while achieving nothing. If you wanna risk your life for a cause then take direct action, eating processed slop and pretending to feel good about it is only gonna make both your world and mine shittier.


The thing is, we don’t even have good evidence that UPF is necessarily harmful. Whey protein is UPF, but is associated with positive outcomes. Mass produced wholemeal bread is UPF, but is associated with good outcomes.

I’m not convinced that the “UPF” category adds anything useful over “HFSS” at this point. Happy to be pushed off my view, but seen nothing that would do so thus far.


Looking at the code, this converts PDF pages to images, then transcribes each image. I might have expected a pdftotext post-processor. The complexity of PDF I guess ...


There is a very popular Python module called ocrmypdf. I used it to help my HOA and OCR’ing of old PDFs.

https://github.com/ocrmypdf/OCRmyPDF

No LLMs required.


20 years ago I tried in vain to get my HOA to use the virtual printer for PDF documents so they'd be searchable. The capability was built in to both Mac and Windows even way back then.

No luck. They just could not grasp it. So they kept using their process of printing out the file on paper and then scanning it back in as a PDF image file.

I finally quit trying. Now of course they've seen the light and are painstakingly OCRing all that old stuff.


Ouch! I am on the BOD so as an IT/Engineering Professional I can influence things better


It's nice, I've used it as a fallback text extraction method in an ETL flow that chugged through tens of thousands of corporate and legal PDF files.


Shell: GNU parallel, pdftotext

Python: PyPdf2, PdfMiner.six, Grobid, PyMuPdf; pytesseract (C++)

paperetl is built on grobid: https://github.com/neuml/paperetl

annotateai: https://github.com/neuml/annotateai :

> annotateai automatically annotates papers using Large Language Models (LLMs). While LLMs can summarize papers, search papers and build generative text about papers, this project focuses on providing human readers with context as they read.

pdf.js-hypothes.is: https://github.com/hypothesis/pdf.js-hypothes.is:

> This is a copy of Mozilla's PDF.js viewer with Hypothesis annotation tools added

Hypothesis is built on the W3C Web Annotations spec.

dokieli implements W3C Web Annotations and many other Linked Data Specs: https://github.com/dokieli/dokieli :

> Implements versioning and has the notion of immutable resources.

> Embedding data blocks, e.g., Turtle, N-Triples, JSON-LD, TriG (Nanopublications).

A dokieli document interface to LLMs would be basically the anti-PDF.

Rust crates: rayon handles parallel processing, pdf-rs, tesseract (C++)

pdf-rs examples/src/bin/extract_page.rs: https://github.com/pdf-rs/pdf/blob/master/examples/src/bin/e...


I imagine part of the issue is how many PDFs are just a series of images anyway.


Saw this tweet the other day that helped me understand just how crazy PDF parsing can be

https://threadreaderapp.com/thread/1955355127818358929.html


There are a few other reasons why PDF parsing is Hell! > https://unstract.com/blog/pdf-hell-and-practical-rag-applica...


Image-based extraction often preserves layout and handles PDFs with embedded fonts, scanned content, or security restrictions better than direct text extraction methods.


> there’s also an official, built-in map of the planet, carefully crafted

Is there a mathematical framework for how to optimize a map for gameplay to be most enjoyable?


I burned a heck of a lot of tokens on Claude Opus 4’s extended thoughts on this and the answer is most enjoyable!:

Taking this question seriously leads us into surprisingly rich territory! It's actually at the intersection of several fields that have been quietly revolutionizing game design.

The Alpha Centauri map works brilliantly because it balances several mathematical tensions:

* Resource distribution follows power laws that create natural chokepoints and valuable territories without being too predictable

* Distance metrics between faction starting positions that ensure interaction without immediate conflict

* Terrain connectivity that creates interesting path-finding problems and strategic depth

Here's where it gets really interesting:

Flow Theory Mathematics: Csikszentmihalyi's concept of flow has been formalized into mathematical models. The ideal difficulty curve follows something like: `D(t) = S(t) + ε`, where difficulty matches skill level plus a small challenge margin. For maps, this translates to ensuring players always have meaningful decisions at their skill level.

Information Entropy: Good maps maintain optimal information entropy - not too random (chaos), not too ordered (boring). Researchers have found sweet spots around 0.3-0.5 on normalized entropy scales for terrain variation.

Graph Theory Applications: Maps are fundamentally graphs, and metrics like:

* Betweenness centrality (identifying crucial chokepoints)

* Clustering coefficients (how "clumpy" resources are)

* Shortest path distributions (travel time variance)

...all correlate with player engagement.

The Deeper Insight: What seems naive is actually profound - by taking "fun" seriously as an optimization target, we're forced to confront what makes human decision-making satisfying. The best mathematical frameworks don't try to define fun directly, but instead optimize for decision richness - the number of meaningful, non-obvious choices available at each game state.

This is why procedural generation in modern games increasingly uses these frameworks, creating maps that aren't just random but mathematically tuned for engagement.


Most of what it said is fitting math to heuristics - “don’t put starting points too close to each other” and “have some chokepoints so there’s some battles” and “give everyone equal access to resources”

The map was likely developed iteratively via playtesting - no need even for heuristics then. Just play & ask “is it fun?”


tbh i would have been more interested in your thoughts than some AI's.


Me too! I happened to learn a few new things / aspects from asking and I thought I’d toss this into a comment


I don't know if I should trust and run this code. If it was associated to Mozilla I would. It says it is a Mozilla Builders project, but https://builders.mozilla.org/projects/ does not list it. I don't see a way to verify that localscore.ai is associated with Mozilla.


It doesn't seem they update the site often (the last 'latest' post is from December) but they reposted something claiming the same on X https://x.com/llamafile/status/1907917417118105751


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: