chartered_stack

This reminds of when I was in school and we had 3 categories of notebooks.

1. A "rough" notebook to jot down notes 2. A "classwork" notebook to present those notes properly to the teacher and get them corrected and 3. A "homework" notebook to finish homework assignments that used to be given

It's quite a discipline to jot things down in the rough work book and then write out a "neat" version in the class work book.

I stopped doing it once I finished school and realized I didn't really need to do it anymore once I understood what I was writing out. But maybe it’s a habit worth returning to.

One could say: vLLM isn't a worse Llama.cpp, it's a different tool

Apple’s always done their best work when they’re the second mover.

People say Apple does its best work as a “second mover,” but that misses the actual pattern: Apple builds great products when leadership is solving their own problems.

The Mac, iPod, iPhone, and iPad weren’t just refinements of existing products. They were devices Steve Jobs personally wanted to use and couldn’t find elsewhere. The man saw the GUI at Xerox and saw how anyone could use a computer without remembering arcane commands. So he drove the development of the Mac. He was using a shitty mobile phone, saw the opportunity and had the iPhone developed. Same with the early Apple Watch (first post-Jobs new product line), which reflected Jony Ive’s fashion ambitions; once he left, it evolved into what current leadership actually uses: a high-end fitness tracker.

The stagnation we're seeing now isn’t about Apple losing its “second-mover magic.” It’s that leadership doesn’t feel an unmet need that demands a new device. None of Vision Pro, Siri, Apple Intelligence or even macOS itself anymore appear to be products the execs themselves rely on deeply, and it shows. Apple excels when it scratches its own itch and right now, it doesn’t seem to have one.

I understand the feeling. There is a huge asymmetry between individual contributors and huge profitable companies.

But I think a frame shift that might help is that you're not actually donating your time to LMAX (or whoever). You're instead contributing to make software that you've already benefited from become better. Any open source library represents many multiple developer-years that you've benefited from and are using for free. When you contribute back, you're participating in an exchange that started when you first used their library, not making a one-way donation.

They wouldn't have merged my code in if they didn't think it had some amount of value, and if they think it has value then they should pay me.

This can easily be flipped: you wouldn't have contributed if their software didn't add value to your life first and so you should pay them to use Disruptor.

Neither framing quite captures what's happening. You're not in an exchange with LMAX but maintaining a commons you're already part of. You wouldn't feel taken advantage of when you reshelve a book properly at a public library so why feel bad about this?

Like Alan Kay said about software: Simple things should be simple, complex things should be possible.

The thing is this takes a lot of resources to get right. FOSS developers simply don't have the wherewithal - money, inclination or taste - to do this. So, by default, there are no simple things. Everything's complex, everything needs training. And this is okay because the main users of FOSS software is others of a similar bend as the developers themselves.

True the unpredictability sucks right now. We're in a transition stage where the models can understand intent but cannot constrain the output within some executable space reliably.

The bridge would come from layering natural languages interfaces on top of deterministic backends that actually do the tool calling. We already have models fine-tuned to generate JSON schemas. MCP is a good example of this kind of stuff. It discovers tools and how to use them.

Of course, the real bottle neck would be running a model capable of this locally. I can't run any of models actually capable of this on a typical machine. Till then, we're effectively digital serfs.

I think we're returning to CLIs mostly because typing remains one of the fastest ways we can communicate with our computers. The traditional limitation was that CLIs required users to know exactly what they wanted the computer to do. This meant learning all commands, flags etc.

GUIs emerged to make things easier for users to tell their computers what to do. You could just look at the screen and know that File > Save would save the file instead of remembering :w or :wq. They minimized friction and were polished to no end by companies like MSFT and AAPL.

Now that technology has got to a point where our computers now can bridge the gap between what we said and what we meant reasonably well, we can go back to CLIs. We keep the speed and expressiveness of typing but without the old rigidity. I honestly can't wait for the future where we evolve interfaces to things we previously only dreamt of before.

Sure, plenty of incompetence out there, but nobody wakes up thinking they’re doing wrong. They overpromise, they play it safe. Sometimes it works, sometimes it doesn’t.

As Picard says, "It is possible to commit no mistakes and still lose. That is not a weakness; that is life"

I think this piece makes a fair and important point about LLM hype and the need to treat it as a normal technology rather than a cult movement. The over-the-top marketing and constant “AI will change everything” drumbeat can definitely obscures the more grounded, practical ways it can be used day-to-day.

That said, every major technology wave has needed a similar level of push, hype, and momentum to reach mass adoption. The Internet existed for decades before the public knew what to do with it. AOL gave such a huge push with the “You’ve got mail”, endless free trial CDs and an almost manic push to bring it into homes for it to become the foundation of modern life. The same was true of personal computers: early machines like the Apple II or IBM PC were expensive, clunky, and had little practical software. But without the evangelism, marketing, and cultural hype that surrounded them, the entire ecosystem might never have matured. So while the AI frenzy can feel excessive, some level of over-excitement may be what turns the technology from niche tools into something broadly accessible and transformative — just as it did for the web and the PC before it.

What makes you think that it's only CEOs who look like they're part of a hive mind? It's got nothing to do with capitalism or it's "creepy agents". It's simply the human condition. It's literally company/human see company/human do. One company/management loudly kangs whatever their position is and others simply follow it because they think it's either "industry standard" or it's convenient somehow to them. That's all it is - people trying to be safe.

Let me give you a technical parallel. A couple of engineers/architects from Big Co. that's hugely successful leave and go to Hot Startup. There they proselytize their One True Way because honestly that's all they know. Everybody in Hot Startup goes along with it because they are Senior Engineers from Big Co. who are now plotting the course and Big Co. is HUGE so they know what they're talking about. Now because Hot Startup is suddenly using the One True Way everybody else in the market tries to copy them because that's obviously why Hot Startup is Hot. This leads to a job market where people optimize for things used by Hot Startup. This tilts the skill set of the general tech market towards the One True Way making it gospel to a lot of people. So hiring managers who don't know the first about anything suddenly start optimizing for One True techs and ask for 20 years experience with React. They think they're doing the safe thing by using the same tech stack used by everybody else - the industry standard. Never mind that the "industry standard" changes every time it's convenient.

This is the same thing for CEOs. Oh you're having a slightly down quarter and have to answer to investors? Say you're using AI. That's the in-thing and will give you that bump to ride out the quarter. You screwed up in 2021-22 and hired a fuckton of people who are just sitting on their hands costing the company money? Say AI and get rid of them because they're not productive. It's got nothing to do with collusion or anything like that. It's just that people have mismatched expectations and things happen downstream of these unmanaged expectations.

Governments have no right or authority to stop us from being idiots if we want to be idiots.

This sounds fine in theory but it ignores the fact that gambling today isn’t just about individuals making free choices in a vacuum. There’s an active, systemic push to get people hooked. Millions (billions?) are spent on ads, algorithms, and dark patterns designed to keep people hooked. That's not freedom - that's exploitation.

With modern tech like gambling apps on your phone, 24/7 internet access, social media tie-ins the problem multiplies. You don't have to go to a casino when you have one in your pocket. The same tricks that make people lose hours on TikTok are being weaponized to make them lose their money.

Freedom matters. But if the entire system is engineered to trap people in endless dopamine hits, then society has to step in. Not to ban choice, but to create a framework that tilts people away from predatory addiction loops and toward things that actually build resilience and meaning. Otherwise “freedom” just becomes another word for “you’re on your own while other people drain you dry.”

Honestly, it would be great if it were "Gemma in Chrome" instead.

A local model capable enough to do the things that this is designed to do? Yes please.

Gemini in Chrome is a way to increase adoption. Gemma in Chrome is an innovation - a platform that allows developers to build stuff leveraging the local model. A step closer to a world where we can talk to our computers and have them do what we mean instead of what we say.

There are really two separate issues here:

A) It should be harder for non-technical users to accidentally install apps designed to harm them.

B) It should also be possible for anyone to run whatever code they want on hardware they own.

Both can be true, and platforms should support both. Ultimately, it is up to the platform to decide what they want to allow and how they protect their users.

I get why Android is tightening controls: plenty of people install shady APKs they get from random websites or Telegram/WhatsApp groups and get burned. But forcing developers to register with Google isn’t the answer. If I want to run a hobby project on my own phone, I for sure shouldn't have to jump through bureaucratic hoops.

The thing is that Google already has the mechanism to protect users: the Play Store. The real problem is that its review process is weak and flooded with low-quality and malicious apps. Fixing that would do far more good than punishing independent developers. They also don't want to open up anti-trust behavior by actually prioritizing the Play Store and saying that you shouldn't trust an app from a random Chinese App Store.

If Google wants to make Android safer, step one should be cleaning up the Play Store. Step two is making that the obvious, prioritized channel. Only after that should they even think about playing Big Brother.

I often come across tips/pointers/exhortations about how to write good design documents. I generally agree that it's an important step: not just for clarifying your own thinking, but also for communicating effectively with others.

However, these types of posts often lack are concrete examples of what a good design document actually looks like. I understand that many of these documents are proprietary and intended for internal use. Still, are there any examples of well-written design documents available publicly that learners can study to get a clearer idea of what one should look like?

I hope these OSS CC clones converge at some point.

Imo, the point of custom CLIs is that each model is trained to handle tool calls differently. In my experience, the tool call performance is wildly different (although they have started converging recently). Convergence is meaningful only when the models and their performance are commoditized and we haven't reached that stage yet.

I'd argue it's not a choice of "let's open a campus in flyover country," but a reflection of how the industry has changed.

The "older" companies were manufacturers. Even places like Mountain View and San Jose were the working-class towns with HP factories and semiconductor plants. The concentration of engineering talent (HP/Intel/Apple/Atari) is what created the affluence, especially after manufacturing itself was outsourced globally.

The newer Web 2.0 companies don't make physical things; they make software. Their most critical infrastructure isn't a factory but a dense network of developers. They go to the Bay Area, Seattle, etc., because that's where the network is. For the parts of their business that don't require that network, like customer service, they locate in less expensive regions, just as PayPal did with Nebraska. They were even the second largest employer in Nebraska iirc.

I'm a convert on this topic. I went from wanting a small phone to being unable to wait to ditch mine.

Like the OP, I switched from Android (Pixel 3a) to an iPhone SE 3 specifically for the smaller form factor. After using it for over a year, I've found the trade-offs in battery life and camera quality are too significant for my daily use.

These limitations aren't an issue when I'm at home or my desk with easy access to a charger. However, they become acute the moment I'm out for the day. For example, using GPS for navigation or connecting Bluetooth accessories becomes a liability. I can't rely on the phone to last. Also, photos are noticeably more pixelated, and the quality drop-off is clear compared to larger, contemporary phones.

This thread is evidence that the niche for small phones exists. But it's for people willing to accept these compromises by carrying a dedicated camera, a power bank, and using wired peripherals. For me and as the market suggests for most consumers, small phones just doesn't work out as reliable all-in-one devices. I'll probably wait till early next year to pick up one of the new iPhones after they iron out the initial kinks.

The Thunderbird search bar really sucks. Advanced search with the actual functionality is hidden away behind some weird menu while the big honking bar at the top of each page does basic text search and offers nothing more.

Well is the loss of training data from customers using self-hosted Llama that big a deal for OpenAI or any of the big labs at this point? Maybe in late-2022/early-2023 during the early stages of RLHF'd mass models but not today I don't think. Offerings from the big labs have pretty much settled into specific niches and people have started using them in certain ways across the board. The early land grab is over and consolidation has started.

That's a fair argument to make.

My main point was that I like the auto-translation because information retrieval was so much better. But the intent behind the video of the uploader has to be maintained and respected.

That's a fair point.

I use English while using any YT client so I didn't notice it as much. So I changed the language on my browser to German, went to a couple of American channels and now I get some of the outrage on this thread. Weirdly, some videos get their titles translated and some don't. Not only the titles but also the descriptions get translated. Honestly, I'm surprised they aren't translating the comments at this point.

I still do understand and like the feature a lot. It's a good way to push the engagement rates I guess. A simple solution would be to show a dialog with languages where the users can pick what they speak, not touch videos in those languages and translate everything else.

I actually appreciate the YouTube's auto-translate feature a lot because it allows me to search through videos in languages I don't know but still like to view videos and listen to videos in. For example, I listen to a lot of city pop and anime title songs on YouTube and a lot of them have titles in Japanese only. I absolutely would not find it as easy as I do to search through this content and listen to the music if the auto-translation feature did not exist. It just makes it easier for people who don't know the language to view videos in that language. Sure the translation quality might not be the best but it makes search a whole lot easier. This is why I find some of the comments on this thread surprising.

Having said that I am against the automatic audio translation that some people are reporting. I have not experienced it myself but that seems poorly thought out. It should be easier for people to search through items in a foreign language but that content should be served in the content originally intended.

You're right that this isn't "launched to the public" and is just a developer beta. However, I meant it in a more "outside of Apple" kind of way. I guess that should have been clearer.

The main issue I feel is that Apple's internal threshold for what quality of software is acceptable to be launched to the public has dropped a lot in the years since the last major redesign.

Yes, they iterate through versions and drop things that don't work with their design philosophy (parallax effects on iOS 7) but the first major version they released always seemed well thought out and solid from a design perspective.

I don't get that feeling from this redesign. I'm sure that this Liquid Glass redesign would look and work great next year or the year after that or even by the public launch of iOS 26. They'll fix the issues with readability, control center etc. But the fact that the first version of Liquid Glass doesn't look good is what's problematic.