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

Isn't at least part of the problem with replication that journals are businesses. They're selling in part based on limited human focus, and on desire to see something novel, to see progress in one's chosen field. Replications don't fit a commercial publications goals.

Institutions could do something, surely. Require one-in-n papers be a replication. Only give prizes to replicated studies. Award prize monies split between the first two or three independent groups demonstrating a result.

The 6k citations though ... I suspect most of those instances would just assert the result if a citation wasn't available.


Not in academia myself, but I suspect the basic issue is simply that academics are judged by the number of papers they publish.

They are pushed to publish a lot, which means journals have to review a lot of stuff (and they cannot replicate findings on their own). Once a paper is published on a decent journal, other researchers may not "waste time" replicating all findings, because they also want to publish a lot. The result is papers getting popular even if no one has actually bothered to replicate the results, especially if those papers are quoted by a lot of people and/or are written by otherwise reputable people or universities.


Journals aren't really businesses in the conventional sense. They're extensions of the universities: their primary customers and often only customers are university libraries, their primary service is creating a reputation economy for academics to decide promotions.

If the flow of tax, student debt and philanthropic money were cut off, the journals would all be wiped out because there's no organic demand for what they're doing.


Gimp would have to be the extreme example of this. I used to recommend Krita to people, despite it being less appropriate for photo editing, just to avoid using 'Gimp' in work/polite scenarios.

I agree - "Birdy" is the name used with infants when talking about birds, or is a bird toy that photographers use to distract people ... which is a bit too close to the truth, perhaps.

To me it also suggests 'a toy version of Twitter'; and Twitter already had enough negativity around it for me.


Somehow I feel like GIMP's lack of popularity has more to do with its reputation for having a horrendous and impenetrable interface than its name.

At one point in the recent past there was a fork of GIMP named "Glimpse," yet weren't a sudden influx of users who were waiting for a more polite name.


I would say both.

BUT, lack of users might just be that it's too late, now. People use web-based tools like Figma, I wouldn't think a lot of people are looking for a Photoshop alternative.


Krita is doing just fine. It has the subjectively "better" name, but also the improved UX.

We're missing the last part of this quadrant with a Krita-like app that has great UX but a bad name, but the preponderance of the evidence thus far tells me that it is more likely than not that the name didn't matter, or at the very least that the UX definitely did matter while the name might not have.


Surely it's very similar, companies can't - AFAIK - be registered in USA, they're registered in a state. USA's States have different tax and legislative climates, just like EU states do.

There is actually a "European company" structure.

https://europa.eu/youreurope/business/running-business/devel...

Most notably, Airbus is an "European company".


It's not. Part of Russia is in Europe. The geographical limit between Europe and Asia is not well defined.

I think it would be similar to saying "First American chat app that...", which would be ambiguous?


It Is fair to say that "Europe" is a proxy for "European Union", like "America" is usually understood as "United States of America", without any precise geographic connotation.

Their service operates in the European Economic Area, which includes more countries than the EU and is therefore closer to the European geographic surface.


I think that "America" actually means "the USA". "American", on the other hand...

Sure, but the U.S. are a single country, while Europe is many different countries that are completely different.

I'm in Poland and can drive 2 hours and stop understanding what people are saying to me (in German and Czech).

That was my point.


> while Europe is many different countries that are completely different.

I've always found this a weird take. European (EU) countries are more similar to each other than any country outside of Europe is to any European country.

In your example, if you drive two hours to Germany or Czechia, your car will still be insured, all your bank cards will still work, the price of your mobile phone service stays the same, you'll have a good idea how health and employment systems work, and the chances are you'll be able to talk to people in English.

It remains true that the barriers the businesses face are higher, but that's not what your example was about.


> I've always found this a weird take. European (EU) countries are more similar to each other than any country outside of Europe is to any European country.

You think finland and malta are more similar to each other than sweden and norway?


We should definitely go all in on a perpetually growing population, what could possibly go wrong.

It's going to hurt, but that's what happens when people don't look to the future and do everything out of greed. The sooner we stop population growth the better.


This "greed" and lack of foresight underpins a lot of the welfare states of Europe. Many member states are doing nothing at all to account for this population shrinkage, much less plan for a world with less people.

>The EU decided to play the US and China against each other to improve its own standing, which is why the US is now moving away from the EU.

How so? What actions did the EU take?

You don't think the declaration of himself as dictator and the immediate threats against EU allies might have changed EU attitudes at all?


Your ruler is no longer following the rules of law, nor the foundational constitution. USA ended with their declaration of dictatorship and the failure of your houses/legislature/military to act against that and defend the Constitution.

I can't see how, since the end of habeas corpus, you can claim legal stability.

Your leader is World renowned for reneging on debts and is demanding bribes for companies to operate.

Isn't borrowing through the roof to pay for things like your stasi?

Daily those stasi are murdering and disappearing people seemingly attempting to foment an excuse to escalate the violence.

I don't know how that knife edge can look anything like stable to you.


It's a very grandoise (or alarmist, depending on your perspective), but this isn't super new. The US has been "unstable" with rulers breaking their own laws domestically and internationally for many decades.

Why is this getting downvoted?

My money is that influence campaigns are active on HN and try to mold the discourse. The whole internet is manipulated to hell, and HN is a prime target, you have a bunch of smart people that probably have oversized influence, how could you NOT try manipulating this place?

This is most certainly happening. A lot of US-critical articles also get flagged to death, even when they have a lot of upvotes and healthy, civilized discussion.

Yes to manipulation.

But also it asserts factually a lot that isn't true, which is why I downvoted.


Mostly an US audience here. A deeply divided country on politics. The wording doesn't help matters either.

It's unfortunate the same division tactics the US has been facing are working elsewhere too, however.


Mostly an US audience here.

I don't think mostly is true? Obviously it depends a lot on the time of day, but there are also a lot of Europeans on the site. Also, most comments here seem to be critical towards current US policy. So, I think there is quite a lot of manipulation going on, since the downvoting/flagging does not really match the comment section.


I think it's true. There is a significant audience here from other areas but this being an english language forum and one focused on tech means that the US is always going to have a dominant presence[1]. The US dominance also means that the news is highly focused on US events when it wanders out of tech which further reinforced the audience.

1. I believe Canada does have an outsized presence though!


From dang themselves: https://news.ycombinator.com/item?id=16633521

So, 32-56% US, so not mostly US.


Because it makes people uncomfortable.

It's hysteria in the addictive, viral, breathless, and self-indulgent social media flavor that permeates everything.

The excitement being blown out of proportion is hyperlocal. The system grinds on.


I've been actively moving away from USA originated products. I'm happy to see alternatives being discussed. I really don't think it's moral to fund fascist states in this way, sorry.

Yes, I'm still here, despite being told (paraphrasing) 'fuck off we don't want anyone from outside USA here'.


Fascism: 'A system of government marked by centralization of authority under a dictator, a capitalist economy subject to stringent governmental controls, violent suppression of the opposition, and typically a policy of belligerent nationalism and racism.'

Interesting because doesn't every sort of democratic state try to be 'a capitalist economy subject to stringent governmental controls'? Depends how stringent and usually not stringent enough for many on the Left and on the Right.

When tempted to use the word 'fascism', is it not better to describe the issue with which one's concerned (maybe deeply) rather than using a fit-all word and take care not to devalue the significance of the word as it was, for instance, applied in WW2 to some of the appalling atrocities that occurred in that period and those we've seen reports of recently?


A clearly corrupt dictatorial leader, lack of rule of law, violent oppression of ones own citizens, perpetual lying and use of mass media to influence the populous, violent incursions in to the affairs of other countries without UN support, threats of violent invasion against erstwhile allies, release of violent offenders and drug dealers (clearly in exchange for money), accepting of bribes more generally, acting against supposed allied interests, purging of military leaders according to political affiliations, manipulation of international markets for insider trading, refusal to address child trafficking for rape and sexual abuse (and possibly murder), release of violent insurrectionists, release of violent neo-nazi offenders (eg Proud Boys), breach of constitution, actions to prevent national and foreign journalists having access, actions to suppress proper reporting of the regimes actions, war crimes/murder of shipwrecked people, ...

I'm sure there's loads more if the question is somehow genuine?

Just spending $billions on an illegal (ie not established through constitutionally sound, democratic legislative means) military force, who travel in unmarked vehicles, conceal their identities, and target citizens in areas that politically oppose Trump are alone actions of a fascist regime. They murder, disappear, deport with no due process and act as if outside the law. Their actions, such as murder, are supported fully by the public face of the regime.

This regime is publicly supported by the billionaire owners of tech and media companies operating in USA.

Little people like me don't have much, but every £pound is a vote.


What does fascism mean to you, exactly?

It feels like one could produce a digest of the context that works very similarly but fits in the available context window - not just by getting the LLM to use succinct language, but also mathematically; like reducing a sparse matrix.

There might be an input that would produce that sort of effect, perhaps it looks like nonsense (like reading zipped data) but when the LLM attempts to do interactive in it the outcome is close to consuming the context?


``` §CONV_DIGEST§ T1:usr_query@llm-ctx-compression→math-analog(sparse-matrix|zip)?token-seq→nonsense-input→semantic-equiv-output? T2:rsp@asymmetry_problem:compress≠decompress|llm=predict¬decode→no-bijective-map|soft-prompts∈embedding-space¬token-space+require-training|gisting(ICAE)=aux-model-compress→memory-tokens|token-compress-fails:nonlinear-distributed-mapping+syntax-semantic-entanglement|works≈lossy-semantic-distill@task-specific+finetune=collapse-instruction→weights §T3:usr→design-full-python-impl§ T4:arch_blueprint→ DIR:src/context_compressor/{core/(base|result|pipeline)|compressors/(extractive|abstractive|semantic|entity_graph|soft_prompt|gisting|hybrid)|embeddings/(providers|clustering)|evaluation/(metrics|task_performance|benchmark)|models/(base|openai|anthropic|local)|utils/(tokenization|text_processing|config)} CLASSES:CompressionMethod=Enum(EXTRACTIVE|ABSTRACTIVE|SEMANTIC_CLUSTERING|ENTITY_GRAPH|SOFT_PROMPT|GISTING|HYBRID)|CompressionResult@(original_text+compressed_text+original_tokens+compressed_tokens+method+compression_ratio+metadata+soft_vectors?)|TokenCounter=Protocol(count|truncate_to_limit)|EmbeddingProvider=Protocol(embed|embed_single)|LLMBackend=Protocol(generate|get_token_limit)|ContextCompressor=ABC(token_counter+target_ratio=0.25+min_tokens=50+max_tokens?→compress:abstract)|TrainableCompressor(ContextCompressor)+(train+save+load) COMPRESSORS:extractive→(TextRank|MMR|LeadSentence)|abstractive→(LLMSummary|ChainOfDensity|HierarchicalSummary)|semantic→(ClusterCentroid|SemanticChunk|DiversityMaximizer)|entity→(EntityRelation|FactList)|soft→(SoftPrompt|PromptTuning)|gist→(GistToken|Autoencoder)|hybrid→(Cascade|Ensemble|Adaptive) EVAL:EvaluationResult@(compression_ratio+token_reduction+embedding_similarity+entailment_score+entity_recall+fact_recall+keyword_overlap+qa_accuracy?+reconstruction_bleu?)→composite_score(weights)|CompressionEvaluator(embedding_provider+llm?+nli?)→evaluate|compare_methods PIPELINE:CompressionPipeline(steps:list[Compressor])→sequential-apply|AdaptiveRouter(compressors:dict+classifier?)→content-based-routing DEPS:numpy|torch|transformers|sentence-transformers|tiktoken|networkx|sklearn|spacy|openai|anthropic|pandas|pydantic+optional(accelerate|peft|datasets|sacrebleu|rouge-score) ```

Win+shift+up

Better

If you're selling 49000 electric vehicles, and the tariff reduced from $CAN 50k (estimated cost of a new electric vehicle; 100% tariff tax) to 3k (6%), saving your customers $2.3B, that seems significant to me?

I'm only trying to give a feel for them numbers, I did check the average selling price for a new BYD


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

Search: