https://www.sciencedirect.com/science/article/abs/pii/004724...
3,194 karma · joined April 28, 2014
https://www.sciencedirect.com/science/article/abs/pii/004724...
Given those two facts, I assumed we'd crossed that boundary already.
Personally, I use computers for work, for reading/research/writing, for news, for gaming, for creating music, and for passive entertainment. They all feel qualitatively very different. I have a hard time adding up all the hours and calling it "screen time". 20 hours reading a good book or editing an audio track is far healthier than 2 hours of doomscrolling.
> “Children are spending more and more time on screens, for everything from entertainment to homework to messaging friends,”
Yes. And it's the algorithmic curated feed that's destroying messaging. And the advertisements and garbage content of the entertainment industry. And AI largely eroding our ability to think critically. None of that has to do with screens themselves.
Maybe instead of "screen time" we should differentiate on a number of axes: Passive vs. active. Creating vs. consuming. Social vs solo. Mediated by algorithms vs mediated by social norms. Not all screen time is created equal, and some may actually be socially net positive.
I figured what the hell, let's try NixOS and see if I can set it up declaratively. At least that way if it fails, I have documentation about what exactly is failing.
I discovered the audio packages I needed easily enough. But things weren't great, there was occasional noticeable lag on the mic and midi devices. Ok, let's try to compile a realtime linux kernel... Wait there's an entire project dedicated to Nix audio setup! (https://github.com/musnix/musnix) After a few minutes integrating with my config, and about an hour to compile the kernel. Rebooted and bam. It's almost like working with analog in terms of perceivable lag.
I know this says more about Musnix than Nix, but I think it highlights the advantages of configuring your system with a real programming language. All the "audio stuff" was able to fit behind a well-designed abstraction barrier.
Going back to a "do this, now do this" tutorial feels like the dark ages. The only thing keeping me from going all-in on Nix is that it's package management strategy starts bleeding into your software project's dependency management. This is a huge problem IMO, I can't run C/C++ or Python projects the "old way" and no one is willing to upend their build system and lock it into nix. Some languages work great and allow a clean barrier between system<>app. I use Clojure, Rust, Gleam, Go, Lua on nix without any issues, it does seem to be the C/C++ ecosystem and dynamic linked shared libraries that are the problem.
Given that these regions were clearly valuable real estate - valleys with waterways near urban centers - it's very hard to make the case that these 307 flooded valleys represent all rural areas on earth. It seems like the authors conveniently skip the bias they themselves introduced. A few hundred dam sites are not representative of rural land on earth, and this paper is sketchy AF for trying to make that claim without supporting their assumption.
Furthermore, and this might be more a chip on my personal shoulder, they fail to outline their methods or do adequate sensitivity analysis. The vector footprint of the reservoirs and the raster population grid are disparate data models and there are many different approaches for determining their overlap. Do you include the center of the cell? Only 50% coverage? Any coverage? 100% coverage only? How are the data reprojected to align spatially? And given the coarse resolution and spatial arrangement (fractal coastlines, high perimeter:area), those details are not just GIS nerdery, they have massive impacts on the results. At least the code is available so it could potentially be reviewed.
If software is about meeting human demands, humans will always write its requirements, by definition. If we build another machine like LLMs, well the design of those LLMs is subject to human demands. There is no point at which we can demand perfection but not be involved in its definition.
Your plan is curiously devoid of answers to the basic questions.
"Chart goes up now, chart must go up forever" is an ideology, not a characteristic of any physical ecosystem. Certainly not an actionable plan that any serious person would follow.
Instead of an ideology, how about a real viable plan pinned to physical reality (CO2 emissions, terrawatts, acres, tons of ore, things we can measure empirically)? I keep being promised this solar future, yet fossil fuel emissions keep rising. Your ideology sounds appealing, but you've got a long way to go to make it reality.
The research you're thinking about shows that fossil fuel usage is declining as a percentage of overall energy use. That much is true. Solar is making progress relative to coal, oil, and gas.
But don't confuse the derivative of the relative difference with the scalar value of interest. Measured in physical units, fossil fuel consumption keeps increasing, as the IEA reports constantly.
It's the difference between living on a fixed income vs drawing down a savings account. And as we draw down our savings account (burn oil) to invest in our fixed income future (solar) we should not fool ourselves into thinking this means an infinite savings account. Whether we draw our savings to zero or plan wisely for fixed-income, either way we are looking at decreased energy availability, aka degrowth.
If it were just a shrinking energy/resource base, we could muddle through and build a better political system. Same with climate change. And biodiversity loss. And chemical pollution. And global nuclear warfare. And the decay of social morality. All of these things are solvable in themselves, but in aggregate I think we're running into harsh economic realities that we might not be able to solve all of it at the same time.
Well, yes, any dependency is technically a "lack of engineering". That's kinda the point - engineering capacity is limited and should be directed towards core business differentiators, not generic infrastructure.
That said, Auth is not really like all the rest of the services. It's special in that it controls all access to everything else making it the load-bearing stone in your entire architecture. Fundamental mistakes in your user auth model tend to be very very hard to unwind. So take the time to do the engineering work (thinking) even if you eventually outsource to a hosted provider - at least you'll know what you're outsourcing and the implications.
I periodically review the notes and summarize anything worth keeping into the appropriate system of record (calendar, tickets, wiki, spaced repetition, doesn't matter). Like you, I find that very little is actually worth keeping. And that's fine; the paper notes are not a system of record but more of an extension of working memory.
I wonder if the generality of the Bayesian approach is what's prevented its wide adoption? Having a prescribed algorithm ready to plug in data is mighty convenient! Frequentism lowered the barrier and let anyone run stats, but more isn't necessarily a good thing.
1. convert to an expect() - even if a panic is fine, you at least owe the future reader an explanation.
2. Use "proper" result handling, usually let Some, .ok_or, or a match.
Unwrap is a nice ergonomic cheat code to get the program compiled and leave a little TODO for yourself. You just need the discipline to circle back and clean them up, just like any debugging tool.
This. Juggling 17 tasks is not a badge of honor. It just shows the world that you have neither the focus to prioritize nor the talent to execute them efficiently. Strange flex, bro.
Put another way, nobody cares about your TODO list, everybody cares about what you've actually shipped. So actually ship things, one at a time.
Notice the critically important difference of recreating an existing design, vs using the rewrite as an opportunity to experiment on the design and the implementation (and the language, and the ...).
Vetting a new design takes time, consensus, and subjective judgement. Re-implementing an existing design is laser focused and objective.
Converting a dictionary into a list of records when you known that's what you want ... easy, mechanical, boring af, and something we should almost obviously outsource to machines. LLMs are great at this.
Deciding whether to use a dictionary or a stream of records as part of your API? You need to internalize the impacts of that decision. LLMs are generally not going to worry about those details unless you ask. And you absolutely need to ask.
You can do the same thing with many other technologies on most other operating systems. I've used, in chronological order: FreeBSD jails, VMs, Cloud-hosted VMs, Docker, K8s, and Nix flakes. WSL is probably somewhere in around K8s.
My point is, we've had the ability to run "subsystems" for decades, by different names, on every OS. WSL is cool but quite late to the game, far from being "more powerful than linux".