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

What are you inferring from that? Interested to hear.

I assume they're implying the new variable aperture on the 18 Pro is more of a marketing/spec sheet thing than an actual meaningful feature

Most users will never use it, or I should say adjust the aperture, most people appear to be more interested in the video performance, and the video performance depends upon having very good software and hardware combined as one which is where Apple is able to outdo the competition, with Apple Silicon/video software, Apple appears to only care about still pictures up to a certain point, over the last five or six years. In short video is king in today’s YouTube age.

> Most users will never use it, or I should say adjust the aperture

This is like saying most users will never adjust the shutter speed. Of course they won't, because the software automatically adjusts it, just like it does on any traditional camera. For the aperture capable lens, the software adjusts the aperture automatically depending upon light and/or depth of field requirements of a scene.

And yes, photographs are still spectacularly important.

But...I don't get it regardless. You know that video also uses apertures on every such equipped device, right? Better depth of field if the environment allows it is a good thing.

> Apple appears to only care about still pictures up to a certain point

I feel like people are just saying silly bullshit now. Apple has made video a core part of their product reveals for years now. And current iPhones take outrageously good videos.


This will always be the problem with trials. Anyone with a job, often underemployed between software roles these days, will be excluded.

Still, it does benefit the hiring company, just not the applicant in these situations.


> Still, it does benefit the hiring company

Not sure it benefits the hiring company when they are artificially reducing the pool of applicants to only people who don't already have jobs.

Not that there's shame in not having a job, especially in today's market (shit happens, I'm sure lots of very qualified people are currently looking), but there are solid statistical reasons why social proof of already being employed is seen as a positive.


I think your point on “cutting down on words”, so to speak, is relative to personal writing style. I find that AI tends to write far more words than I personally do, which is sometimes useful, but in the opposite direction that it works for you.

PROMPT: make this more concise while conveying the same message

I think “cutting down on words” depends on personal writing style. AI tends to write more than I would, so I usually have the opposite problem.


Like the best art, it's left up to the viewer to decide!

Technically, Oklahoma is on the short list of locations that _have_ detonated a nuclear bomb on their enemies in wartime.

Much respect for being able to admit that you haven’t used a model in a while after saying folks probably haven’t used the models you use.

Okay I just tested 3.8 Flash (High) on some real world problems I have used Opus (High) and Sol (high) to solve. The outcome here is terrible.

One of the problems is local zoning laws regarding an expansion of my house. Comparison of annex vs extension, boundaries, precedent, costs, etc.

3.8 Flash didn't check most of the required zoning laws. It relied on parametric knowledge, which is outdated and inaccurate. It checked zero precedents. It did made a very cursory check of the boundary area, but didn't validate it, so it missed a lot of important nuance and exceptions to the boundary. Its cost estimates were wildly inaccurate. Ostensibly because it was inferring an average based on historical pricing data rather than gathering current info.

I could go on, but if I had to judge this attempt I would give it a 3/10. It's very fast, but wildly inaccurate. It's clear that the model is designed for speed over accuracy.

But don't take my word for it. [Most benchmarks show it to be significantly below frontier models like Astra.](https://llm-stats.com/models/compare/gemini-3.8-flash-vs-gpt...)

This has been a useful exercise. It's important to understand the developments taking place. I am disappointed to see that Google has made very little progress in six months relative to the frontier labs.


Gemini's search harness in the Google app is (ironically) bad so it makes the model look bad.

If you really want to compare apples to apples you need to test Gemini models against other models using the same third party search harness.

Otherwise you are largely measuring how much computation the model provider is allocating to a search harness.


I used [AI Studio.](https://aistudio.google.com/) It's possible the same issues are present there, but AI Studio is intended for serious work. I don't see why they would intentionally hobble the model's capabilities in AI Studio.

For me, this art form is one of my “saw it many times growing up and assumed they did it everywhere” experiences. It fascinated me as I grew older and realized I was just lucky enough to experience it (the later sculptures along the East Coast in the 80s-00s).

Planned obsolescence is absolutely to the benefit of corporations if those goods are being rented. Non-durable goods are inherently profitable to rent out at above the cost of outright ownership, because then the companies involved can double dip on the consumers paying the costs of producing the good as well as claiming depreciation on the good the company still technically owns. It's a similar concept to why owning durable goods and renting them out to consumers is so viable.

I imagine you're looking at it from the avenue of "if those goods last longer, they can be rented for longer", but having to manage and store more durable goods after consumers have moved on to newer and better things is a considerable cost. A good that has to be returned is often better discarded, from the company's point of view.


It certainly feels like this is an intrinsic friction point in a world where more and more of the actual code is written by AI. I know in my workflows, reading a design doc of any length is a big tradeoff of time that could be spent building. Which isn't a new problem to have, of course, but as the time from plan->code->review->deploy keeps racing downward, "reading design docs" feels like it's ripe for disruption as far as how much time and attention it takes.

> The religious retort to that…

_A_ religious retort. Also it’s not a binary “lots of kids or no kids” spectrum.


In many countries now, "over the replacement rate" or "under the replacement rate" is indeed a binary that closely tracks with religious observance. I would rather not out where I am, but here a couple having even a second child will cause some observers to at least ask if the family is part of a sect.

Okay, _a_ religious retort. I feel like you focused on a nit instead of addressing my point?

I wasn’t addressing your point, correct. Just helping you be more accurate.

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

Search: