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

I suspect the trend would be even more extreme if multiplied by typical screen brightness over time!

i play helldivers 2 on an HDR monitor and i nearly always balk at playing daytime missions because the in game light source "sun" literally hurts my eyes to look at.

However, barring that (i mean, midday sun is nearly impossible for me, in real life, too, right?), like websites, the backgrounds look brighter on my 1080p TV i have connected to the computer, even though it is set up for the correct colors, so windows understands that letting webdevs attack my retinas is unacceptable, at least.


I'm reminded of the viral tweet along the lines of "Claude just one-shotted a 10k LOC web app from scratch, 10+ independent modules and full test coverage. None of it works, but it was beautiful nonetheless."

This is generally happening in the context of RCTs where valid causal inferences are (almost always) guaranteed via the study design, regardless of whether your analysis is frequentist or Bayesian.

Accepting Bayesian methods for RCTs is great news and leading biostatisticians like Frank Harrell have been pushing for this change for many years. What I'm most interested to see: will this actually be implemented in practice, or will it be incredibly rare and niche, like Bayesian methods are currently in most biomedical fields?

Regulators are generally really conservative. Spiegelhalter et al. already wrote a fantastic textbook on Bayesian methods for trial analysis back in 2004. It is a great synthesis, and used by statisticians from other fields. I have seen it quoted in e.g. DeepMind presentations.

Bayesian methods enable using prior information and fancy adaptive trial designs, which have the potential to make drug development much cheaper. It's also easier to factor in utility functions and look at cost:benefit. But things move slowly.

They are used in some trials, but not the norm, and require rowing against the stream. This is actually a great niche for a startup. Leveraging prior knowledge to make target discovery, pre-clinical, and clinical trials more adaptive and efficient.

Journals are also conservative. But Bayesian methods are not that niche anymore. Even mainstream journals such as Nature or Nature Genetics include Bayesian-specific items in their standard submission checklists [1]. For example, they require you to indicate prior choice and MCMC parameters.

[1] https://www.nature.com/documents/nr-reporting-summary-flat.p...


Bayesian methods are incredibly canonical in most fields I’ve been involved with (cosmology is one of the most beautiful paradises for someone looking for maybe the coolest club of Bayesian applications). I’m surprised there are still holdouts, especially in fields where the stakes are so high. There are also plenty of blog articles and classroom lessons about how frequentist trial designs kill people: if you are not allowed to deviate from your experiment design but you already have enough evidence to form a strong belief about which treatment is better, is that unethical? Maybe the reality is a bit less simplistic but ive seen many instantiations of that argument around.

If choosing a Bayesian approach in a clinical trial can reduce the number of recruited subjects, I would imagine the pharma industry is strongly incentivized to adopt it.

Moreover you can manipulate your results by disingenuous prior choices, and the smaller sample you have the stronger this effect is. I am not sold on the FDA's ability to objectively and carefully review Bayesian research designs, especially given the current administration's wanton disregard for the public good.

I would think there is less opportunity to manipulate your results with bayesian methods than with frequentist ones. Because the frequentist methods don't just require an alternate hypothesis, they depend on the exact set of outcomes possible given your experimental design. You can modify your experimental design afterwards and invisibly make your p-value be whatever you want

BOIN (Bayesian Optimal Interval) trial design is already very common in Phase I studies across therapeutic areas.

Biggest benefit I see from this guidance is support for rare disease trials, where patients are harder to find. Also regulatory bodies will be taking a closer look at stratification groups when it comes time for approval, so sponsors need to keep a super close eye on ensuring even enrollment and preventing misstrats.


Please watch https://www.youtube.com/shorts/2TWu_4gd-ik

Been used since the 90s.


The application of Bayesian probabilistic reasoning in general (as described in this video) is not the same thing as "Bayesian statistics" specifically, which usually to modeling and posterior inference using both a likelihood model and a prior model. It's a very different approach to statistical inference both in theory and in practice. This creator himself is either ignorant of this distinction or is trying to mislead his viewers in order to dunk on the FDA. It's obvious from the video comments that many people have indeed been misled as to what Bayesian statistics is and what the implications of its might be in the context of clinical trials.

Indeed, even more broadly online "Bayesian" seems to have taken on the form of "I know Bayes' Rule and think about base rates" as opposed to "Do you prefer jags or stan for MCMC?"

At a meta level, good design is a very useful tool for discussions about good design!

"While I can't provide medical advice..." is the 2025 version of "As a large language model trained by OpenAI..."

The original study: https://link.springer.com/article/10.1007/s11547-025-02161-1

It was retrospective-only, i.e. a case series on women who were known to have breast cancer, so there were zero false negatives and zero true negatives, because all patients in the study truly had cancer.

The AI system used was a ConvNet used commercially circa 2021, which is when the data for this case series were collected.


> It was retrospective-only, i.e. a case series on women who were known to have breast cancer, so there were zero false negatives and zero true negatives, because all patients in the study truly had cancer.

Well yes, that's the denominator for determining selectivity, which is what the headline claim is about.

Also, they need to set up their next paper:

> However, the retrospective, cancer-only design limits generalizability, highlighting the need for prospective multicenter screening trials for validation.


>The AI system used was a ConvNet used commercially circa 2021, which is when the data for this case series were collected.

Does this mean that newer AI systems would perform significantly differently?


Strictly in terms of architecture, CNNs are still SOTA for small data visual tasks, especially when the target is a locally specific phenomenon where global context isn't as necessary. It has good inductive bias for this.

The main known way to improve performance on tasks like this is getting more data.


Well, certainly not. We shouldn’t draw conclusions about modern AI systems from multi-generation old systems: one way or the other.

Not at all. There is no implication, implicit or explicit, that anything in the world is better or worse. It is just a statement of fact.

>better or worse.

Please quote where I used either word.


> there were zero false negatives

Wouldn't this mean that AI identitied them all has having cancer?


They did all have cancer

Yes, but then the study result should be, "AI correctly identifies 100% of breast cancers in study"

If we're saying there was a discrepancy and we're saying that all of the patients had cancer, then it would seem that there must have been some that were identified as not having cancer by AI.


2021 is an eternity in AI industry.

Edit: I have a problem with the way the title uses "AI" as a singular unchanging entity. It should really be "An AI system misses nearly...". There is no single AI and models are constantly improving - sometimes exponentially better.


I believe there's a big issue in the US of over-diagnosing breast cancer too. "Known to have breast cancer" might not be so clear cut.

That statement would benefit from a link to your source.

https://news.cancerresearchuk.org/2025/11/18/overdiagnosis-w... Is about the UK, but it's a good description of the danger of over-diagnosis.

When the treatment for a diagnosis includes radiation therapy, over-diagnosis can literally cause cancer.


But driving for Uber is much more pleasant than having a McJob. You can listen to music. You can set your own schedule. If you need more hours to make rent you can work more. If you get tired you can just go home.


>f you need more hours to make rent you can work more.

Except the part the algorithm marks you as more desperate and from now on pays you less.


Also much more dangerous, right? Since you're spending so much time on the road.


Work also gives people a status and an identity. Hard for many people to lose that, and can result in a surprising loss of social cache / status.


Part of it is the uncomfortable fact that it is not very much fun to spend time with people who have small children -- it's hard to have a real conversation for more than a few minutes, and if you yourself don't have kids, what their life revolves around right now (daycare, finding a kindergarten, etc) is just not going to be very interesting.


> the uncomfortable fact it is not very much fun to spend time with people who have small children

This is (obviously?) not a fact. I’ve had a blast hanging out with a family in Peru for the last 24 hours. I also always have a great time when I visit my sister, her husband and their little kid.

Long conversations about interesting topics are one way to have fun. And you’re right, those don’t happen as much if you just take a child free couple, some parents and maybe kids and put them in the same room.

Bigger get togethers help a lot. One kid is a handful, but six kids of varying ages can actually be easier. You can also have fun in other ways like dancing, decorating or lighting fireworks (one activity from last night).


Kids are at least as interesting as having to take care of any other living things. As strange as it may seem, if you can ask questions about someone's multiple aquariums, their cats or dogs, their horses or cows, then there's at least, probably more to ask about their children. I don't understand how raising a child can be anything but interesting. Every parent I speak to about having kids has such different philosophies, values, goals, and they're so interesting to learn about. Once the kids can talk, they themselves have so much to say!

Raising kids has to be one of the most interesting things someone can do.


Totally. If you find LLMs interesting imagine how interesting actual growing and developing intelligence is.


Eh, to an extent. But on any given day kids are usually in some phase where they only want to talk about dinosaurs or Micky Mouse or something. It gets repetitive after an hour.

I like playing with kids for a while but I won't pretend it is intellectually stimulating. Sometimes you can find something new to blow their mind though.

When they get to teen years they are capable of more interesting conversations but then often don't want to hang with adults. There is a pretty limited sweet spot of ages.


I’m not going to comment on whether I think you’re right or wrong, your opinion is legitimate.

But just to note that this and similar opinions in the thread confirm what I’ve experienced, both as a SINK/DINK and as a parent.

More often than not it’s the SINKs/DINKs that disappear off the face of the earth and lose interest in participating in the suddenly very different lives of their new parent friends, and not the other way around.


Imagine being in that "not very fun" zone continuously, day and night, for decades.

That's kind of what having kids is like. I love my kids and have great times with them, but there's also a lot of routine, endless cleaning and boredom.

People with kids probably wish they could have longer conversations. They'd happily talk about things other than kids. Sometimes that's possible - but it's very hard to predict when it will happen.

It's something I've observed since having kids - quite a lot of people I have adult relationships with simply have no interest in being near them. As a result, I just don't talk to them at all any more. It's a shame, but there's not really much I can do about it.


That’s because there is a lot of routine chores and boredom to life regardless of whether you have children.

Having kids for most people means LESS boredom, as kids are more interesting and curious and active than most adults.


I agree to some extent, but in my experience it differs a lot between kids how doable having a conversation is


It’s incredibly sad that this pathetic attitude is so wide spread.

Kids are generally much more interesting people to engage and interact with than adults.


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

Search: