But only once did Google actually give me what I was looking for. Every other time the Google results were the same SEO garbage I was getting with DDG.

Maybe I should try switching to Google for a full month to see if my search quality generally improves.

> After looking at the final two layers I was somewhat quick to intuit that this was some sort of password check, but wasn’t entirely sure where to go from there. I tried to reverse it, but it was proving to be difficult, and the model was far too deep. I started evaluating the structure and saw the 64 repeated sections of 84 layers that each process 4 characters at a time. Eventually I saw the addition and XOR operations, and the constants that were loaded in every cycle, and the shift amounts that differed between these otherwise identical sections.

> I thought it was an elaborate CTF cryptography challenge, where the algorithm was purposely weak and I had to figure out how to exploit it. But I repeatedly was getting very stuck in my reverse-engineering efforts. After reconsidering the structure and the format of the ‘header' I decided to take another look at existing algorithms...

Basically it took a lot of trial and error, and a lot of clever ways to look at and find patterns in the layers. Now that Jane Street has posted this dissection and 'ended' this contest I might post my notebooks and do a fuller post on it.

The trickiest part, to me, is that for about 5 of the days was spent trying to reverse-engineer the algorithm... but they did in fact use a irreversible hash function, so all that time was in vain. Basically my condensed 'solution' was to explore it enough to be able to explain it to ChatGPT, then confirm that it was the algorithm that ChatGPT suggested (hashing known works and seeing if the output matched) and then running brute force on the hash function, which was ~1000x faster to compute than the model.

I think this is true for me as well. I have two types of projects that I’ve been working on - small ones with a mix of code I wrote and AI. I have posted these, as I spent a lot of time guiding the AI, cleaning up the AI’s output, and I think the overall project has value that others can learn from and built on.

But I also have some that are almost 100% vibe-coded. First, those would take a lot of time to clean up and write documentation for to make them publishable/useful.

But also, I do think they feel “unworthy”. Even though I think they can be helpful, and I was looking for open-source versions of those things. But how valuable can it really be if I was able to vibe-code it in a few prompts? The next person looking for it will probably do the same thing I did and vibe-code their own version after a few minutes.

Comments URL: https://news.ycombinator.com/item?id=46386093

Points: 3

# Comments: 0

They won’t actually let you delete the Apple TV app, but if you move it out of the top row you will never see the ads.

My parents have an Amazon Fire TV and when I go to their house and have to use it it drives me insane. Carousels of adds large at the top, banner ads as you scroll, full rows of sponsored apps. Full screen ads for random Amazon products when you pause any show you are watching. Everything you watch on Amazon’s streaming service has minute long unskippable ads. Sometimes when you turn it on Alexa will just verbally read you ads.

It’s truly a dystopian piece of tech.

But from all reports that you can find with a quick search it seems clear that it did not sell well by Apple standards.

I would love them to bring it back and I’m not sure what it is about the Hacker News crowd that makes this phone over-represented. Maybe the tech crowd also uses laptops more, so we think of phones as our “small device” and use other devices more as appropriate?

My perception is that a huge percentage of the mass market just like OpenAI because they were the first to market and still have the most name recognition. Even my coworker who works in DevOps says “Gemini sucks, Claude sucks” even though he has never once tried either of them and has never looked at a single benchmark comparison.

I’ve noticed a new genre of AI-hype posts that don’t attempt to build anything novel, just talk about how nice and easy building novel things has become with AI.

The obvious contradiction being that if it was really so easy their posts would actually be about the cool things they built instead of just saying what they “can” do.

I wouldn’t classify this article as one since the author does actually create something of this, but LinkedIn is absolutely full of that genre of post right now.

This makes a lot of sense to me. I was also under the impression that all these lighting effects would be rather computationally expensive. This could encourage people to upgrade devices and make it hard to replicate this design on other brands’ less powerful hardware.

I also don’t use a case or screen protector on my phone fwiw

I was actually just looking at different open source tools I could run locally that would allow agents to write and execute their own Python code. Seems like this might be able to do so in 2 steps, with a file write, and then a command?

In my interpretation the author of the article is doing this almost more out of respect for those around him than himself. As a photographer he was always preoccupied with looking for a good shot rather than enjoying the company he was with.

Even when he talks about the pictures of his child’s birth he looks at it through the lens of a professional photographer - it’s not about the memories attached to the photos, it’s about the composition being ‘generic’ vs the photo saying something interesting.

I feel like this article is really more about work/life balance than taking out your phone to grab a snapshot. That’s just how I read it. Also what a sad ending.

At the same time, however, the people who need to use an LLM for this are going to be the worst at identifying the output’s weaknesses, eg just as I couldn’t write Spanish text, I also couldn’t evaluate the quality of a Spanish translation that an LLM produced. Taken to an extreme, then, students today could rely on LLMs, trust them without knowing any better, and grow to trust them for everything without knowing anything, never even able to evaluate their quality or performance.

The one area that I do disagree with the author, though, is coding. As much as I like algorithms code is written to be read by computers and I see nothing wrong with computers writing it. LLMs have saved me tons of time writing simple functions so I can speed through a lot of the boring legwork in projects and focus on the interesting stuff.

I think Miyazaki said it best: “I feel… humans have lost confidence“. I believe that LLMs can be a great tool for automating a lot of boring and repetitive work that people do every day, but thinking that they can replace the unique perspectives of people is sad.

But when I try to do more complicated math it falls short. I do have to say that Gemini Pro 2.5 is starting to get better in this area though.

But any app that a decent number of people would want to download and use probably won’t be ’vibe coded’, any polished app made by a large corporation certainly won’t be.

The author states that developers will flee Apple because of this friction, but developers will go where the customers are, and I’m not sure a meaningful number of customers care about this.

https://youtu.be/S75VTAGKQpk

I bought my first pair of AirPods very close to their initial release and everyone looked at me like I was an idiot for wearing such an ugly and expensive device that couldn’t really do anything better than a $20 pair of earbuds. Do you remember everyone making fun of them at the announcement? It wasn’t until around 2 years later that AirPods were widely worn and accepted.

I’m not sure Vision Pro will go through the same adoption curve, but I am not confident that it won’t happen by late v2 or v3.