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

If you give, and expect something in return, then you are not giving, that is a transaction.

I’ve written plenty of open source and I’m glad it’s going into the great training models that help everyone out.

I love AI and pay for four services and will never program without AI again.

It pleases me that my projects might be helping out.


What are their drivers like?

Fully open source? Unlikely methinks.


Last I read was Mesa support for those Adreno GPUs is still work-in-process. Link: https://lore.kernel.org/dri-devel/20250930-kaana-gpu-support...

Would be nice if Qualcomm worked with at least 1 OEM to get at least one Linux distro supported at launch.

Hypothetically if they did I'd pre order one.


AI finds nearly 2/3rds of breast cancers!

I see you're not in the marketing department! We can do better by only considering "missed" as what wouldn't also be missed by a human: AI finds 71% of breast cancer*!

*Compared to a human.


Depending how the costs of AI detection vs doctor, that genuinely might be enough to shift the math and be a net positive. If it is cheap enough to test 10x the current tested population, which would have lower, but non-zero rates of breast cancer, then[0] AI would result in more cancer detected and therefore more aggregate lives saved.

[0]presumptively


Given that every positive case needs to be verified by a doctor anyway because the patient has breast cancer, and every negative case has to be checked because it does a worse job than traditional methods... It only costs more.

Depends on the false positive rate. Hypothetically one can 'just' tune the model so false positives are low. This will increase false negatives but those are 'free' as they don't require follow ups. So long as the decrease in cost per real positive[0] goes down there's a benefit to be had.

[0] accounting for false positives, screening costs for true negatives, etc. etc.


> This will increase false negatives but those are 'free' as they don't require follow ups.

Increase in false negative rate significantly reduces survival rate and increases cost of treatment. We have huge multiplication factor here so decreasing false negative rate is the net positive option at relatively low rates.


> Depending how the costs of AI detection vs doctor, that genuinely might be enough to shift the math and be a net positive.

Based on my very superficial medical understanding, screening is already the cheap part. But every false-positive would lead to a doctor follow up at best and a biopsy at worst. Not to mention the significant psychological effects this has on a patient.

So I would counter that the potential increase of false-positive MRI scans could be enough to tip off the scale to make screening less useful


Ballmer would turn in his grave except he’s alive and well.

You can stop worshipping growth any time you like.

Capitalism and democratic systems don't work without growth. Just look at the consequences of the economy stagnating in western Europe and what that is causing on a societal and political level.

No growth means no jobs for new grads, no growth means workers have no negotiating power with their employers, no growth means young people remain poor and cannot afford families. No growth means young people with bad economic prospects seek political alternatives, and newsflash, they far-right with their magic-grift-politics are far more appealing than the far-left with their estorectic utopian far-fetched plans.

Also no growth means a worsening wealth inequality over time.


The alternative is stagnation. And if you think that zero sum economy would favor you over corporations and billionaires, I have a bridge to sell you.

And if you think humanity is making it past ~2100, I have some carbon credits to sell you: https://www.sciencedirect.com/science/article/pii/S001632871...

Growth über Alles / the race to consume everything we can in our spherical petri dish with the same level of awareness as mindless bacteria, will be the end of human civilisation, regardless of anyone's economic beliefs.


Turns out love your neighbor was incredibly valuable advice.

Climate change is simply not an extinction threat. Despite all the fearmongering, the worst case scenarios top out at "WW2" levels of devastation and loss of life.

The only credible extinction threat to humankind so far is an ASI oopise, and that's because it's an intelligent threat.


Assuming positive real return on their investments, a zero growth (or low growth) economy means the rich will own an ever growing proportion of wealth - its simple arithmetic.

Windows is doing great!

Says “Windows Central”.


Most articles I've seen from them on here have been negative enough that I'd assume bias the other way

2026 is the year of the Windows desktop! :)

Don’t let anyone tell you the right way to program a computer.

Do it in the way that makes you feel happy, or conforms to organizational standards.


The right way to program a computer:

Well


No.

There’s many contexts in which programming a computer well is not important.


Snarky.

There’s basically zero innovation in online video.

Such a pity startups can’t innovate on the content stores of the big companies.


It's actually a regression overall compared to physical media like DVDs and Blurays. No director commentaries, no behind the scenes, no silly menu games, etc. Streaming would theoretically allow for tons of this type of content to be made and connected to a film at any time but instead we have this stagnant recreation of cable TV. C'est la vie

The lack of director commentaries and behind the scenes content on streaming has always baffled me as the rights to that must be much cheaper to acquire and would result in more minutes of streaming watched for less licensing money.

It's telling that VFX subcontractors are putting out their own BTS content on YouTube now as promotional material, since the primary production companies for shows and films (with a few exceptions) have completely stopped doing this.

I miss director commentary, I loved re-watching movies with that audio track.

Is there just too much content now? Or has streaming become such a "content mill" that the creators aren't inspired enough about their own work to sit down and talk about it after it's complete?


> Is there just too much content now?

I would guess this is the reason. Before there was unlimited content or ways to entertain yourself on a screen, having additional content on a disc would have been a marketing point to make people feel like they’re getting more for their money.

But now, I doubt even 1 in 1,000 people would respond to that, since there is always something else that can be instantly switched to watching or playing, so why go through the effort?


We’ve started watching Pluribus on Apple TV and it seems like when they’re making the show Apple contractually obligates them to make a podcast about each episode. Some of them are very interesting (like costume design) and some are less so.

It was funny how the sound engineers remoted in for the podcast and had extremely low quality mics, despite it being a show with fantastic sound (really it’s an excellent show in general, just really good).


I noticed the same with Severance (also Apple TV). After every episode, there is a short director commentary/crew interviews about random episode specifics or more higher level thoughts.

I liked it quite a bit.


The Chernobyl tv show had a nice podcast that went with it as well. I think these kinds of extra features are especially nice when it is for a show based on real life. They get to point out things that may not have been 100% historically accurate due to budget/time, and also get to bring in experts to speak about things related to developing the show.

It is funny that these things often just get released on podcast platforms and aren't really integrated into the streaming service.

Especially since this show, and the shows mentioned in these parent comments are all produced by the platforms they got released on. So they also have a whole lot more control to actually integrate this extra content.

These streaming platforms often state they are competing to keep you on their platform consuming things, and it seems odd to me that they wouldn't want to try and capture people for longer with these kinds of extras. Especially since as the other user indicated, these would be much lower cost to produce and license compared to the original content. And for someone who really enjoyed what they watched it would be a pretty appealing extra to have.


DVD extras existed as an incentive for users to re-buy films they already had on VHS.

No such incentive is necessary with streaming, the format competes so well on convenience it doesn't have to invest in extra content.


Exactly. And this is why a whole dimension of collecting rare footages is quite dead now. This is why piracy through these great public trackers still matters.

Rare movies and film documentaries from the 20th century still can be found on rutracker, for example. The Russians really did create a dedicated community of archivists, with the quality varying to a certain degree depending on the uploader's reputation, but they certainly created a notorious collection of movies, even the ones relatively unknown or sometimes censored to death on western countries.


Disney+ has quite a bit of this actually. I agree though that overall most streaming services don’t offer this.

DVDs were iirc 480p which would look absolutely terrible on a modern TV.

Remember those DVD releases that had plastic edges and cardboard fronts/backs? That's an era for you. They're always max 480p, sometimes even 480i, and often single layer, dual sided. Those came out when the Sony was still making Trinitron CRTs that could barely do 720p.

Depends on the DVD. Some of them do look terrible, but some aren't too bad. Probably depends on how it was transferred and mastered and what bitrate they used on the disc.

> There’s basically zero innovation in online video.

AV2 is coming out this year.

> Such a pity startups can’t innovate on the content stores of the big companies.

What do you mean?


Can’t speak for OP but personally I’m thinking of things like the ability to actually add new features. Like what Netflix did with the Bandersnatch episode of Black Mirror years ago. Online video is extremely locked down when compared to the web.

Probably because there are over 9000 different TVs with their own proprietary apps on each one. So the easiest thing to do is just go with the lowest common denominator which is just giving you a menu to play a simple video.

For sure. But something like Bandersnatch shows that it is technically possible. Not for all devices of course. But there could be some kind of open standard companies implement and startups innovate on. But no one with power has an interest in doing that.

Can a startup not make a bandersnatch-like web page with video on it? What would stop them?

20 years ago, it was possible to seamlessly merge video clips from multiple streaming RealPlayer servers into a single composite video stream, using a static XML text file (SMIL) distributed via HTTP, with optional HTML annotation and composition.

This is technically possible today but blocked by DRM and closed apps/players. Innovation would be unlocked if 3rd party apps could create custom viewing experiences based on licensed and purchased content files downloaded locally, e.g. in your local Apple media library. The closed apps could then sherlock/upstream UX improvements that prove broadly useful.


It is not blocked by DRM but different codec. Even if you have two MP4 files, but if they were encoded differently ffmpeg will still need to do some computation to join them.

Gapless playback with MSE would require identical encoding, which is likely more prevalent in the Apple catalog than the wild west of Youtube. Client-side transcode would require DRM cooperation.

For two video streams with different encodings, swapping between two media players + prefetch can give a close approximation of a continuous video stream.


>AV2 is coming out this year.

Which has less than 48 hours to go.


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

Search: