HN user
I'll be curious what the collecter's market looks like for Apple Vision Pro a decade from now. I imagine there are far fewer of these things out in the world than Apple hoped there would be, and I wonder how that'd impact long-term scarcity.
I'd be interested in a podcast app that allows me to subscribe to written RSS content and uses AI to (transparently) convert them to audio podcasts.
The time/attention it takes to engage with written content can be a barrier for me. The hands-free experience of listening to something is so convenient, I find I engage with audio far more often than written text.
The only major downside of audio is the challenges around notetaking, and recording the snippets of info I want to hold onto.
One thing confuses me... Why invest so much into automony on streets/roads when most (all?) US trains/trams/busses still have operators?
Elon said himself that busses have 5x the operating cost of robo-taxis (1$ vs 20¢ per hour), but failed to mention that a typical bus carries 25x people. Presumably the cost delta is due to payroll for a human operators & janitors, so if busses already have 5x max throughout per $ in a worst-case comparison, why not put more energy into driving down costs on the more efficient modality?
I imagine it would be an easier problem to solve too, as you could could define constraints on your operating environment. Dedicated travel lanes, rail, and preprogrammed routes seem like they'd massively reduce the complexity of the problem.
I could also envision a decent financial argument for moving from a customer base of general-consumers to one of cities & local govt. If a 5x operating cost reduction is feasible, I could see "public transit automation" making for a pretty compelling capital project.
IIRC the re-implementation was necessary for type-safety.
I would really love to see a "maps" app that focuses specifically on local discovery for businesses and other points of interest. Or, one that at least makes a real attempt to deliniate between getting you to a known place, and finding you new places to go.
Most mapping apps seem to blend navigation & discovery into a single experience that winds up being worse at both.
Location: Chicago, IL
Remote: Full/Hybrid
Willing to relocate: No
Technologies: Ruby/Rails, Go, JS/React, AWS, Terraform (all used professionally) w/ hobby experience in Elixir & other BEAM languages
Résumé/CV: https://bit.ly/3L7xdq7
Email: contact <at> henrynelsonfirth <dot> com
Full-stack engineer with nearly a decade of total professional experience, and 2+ years of experience working in each of the following domains:
- Front-end React projects w/ an internal npm package repo
- Back-end Ruby/Rails monoliths
- AWS/Golang serverless microservices
- Engineering Management
I also have a strong interest in Elixir and other BEAM languages, and have been staying active as a full-stack Elixir hobbyist for 5+ years while bouncing between different languages/domains professionally.
While open to all opportunities, I am particularly interested in using Elixir professionally full-time :)
(Edited for formatting)
Seems like this exists in a similar problem domain as pkl?
The only time I've had success with using AI to drive development work is for "writers block" situations where I'm staring at an empty file or using a language/tool with which I'm out of practice or simply don't have enough experience.
In these situations, giving me something that doesn't work (even if I wind up being forced to rewrite it) is actually kinda helpful. The faster I get my hands dirty and start actually trying to build the thing, the faster I usually get it done.
The alternative is historically trying to read the docs or man pages and getting overwhelmed and discouraged if they wind up being hard to grok.
I cannot fathom why it is still done this way.
Like many things in America, the government is lobbied to create an unnecessary problem by private companies who aim to profit off of solving that problem.
The nice thing is there are plenty of socially-acceptable, but still every bit as objectively valid reasons to advocate for policies aimed as reducing the use of cars in cities.
I've been feeling this pretty accutely of late.
When stresses of work/life pile up, the inability to free myself from the unceasing car noise to find some tranquility can feel downright overwhelming at times.
We do have a degree of abstraction through terraform modules, but I've found that the same copy-paste problem applies to the terraform that composes those modules together.
This is possibly (if not likely) moreso a result of creating our terraform modules in a suboptimal way due to insufficient expertise than a shortcoming of Terraform itself.
It is also largely a result of having a backlog of scheduled redactors that is longer than I'd care to admit.
Thank you so much for the explanation!
I see much more clearly how something like this could be extremely useful.
I'm having a little trouble understanding the problem(s) Pkl is trying to solve.
After reading the title, my assumption was that Pkl was yet another newer, better configuration language (a la TOML), but now that I've read the article, it sounds like it's more a language for _generating_ config.
Unless I'm mistaken, it sounds like an abstraction on top of your config files meant to help you build & re-use configuration in a more standardized way, rather than yet another config language into itself.
A problem space I'm familiar with is having a bunch of Terraform or Cloudformation configuration you want to share/repeat in multiple projects. Doing-so can get hairy quickly, as the path of least resistance is to copy-paste a bunch of config you barely understand from some other project, and then perform trial-and-error surgery to find and change a couple of lines to suit your project.
Is Pkl designed to help address that sort of problem? Or am I missing something?
I also wonder whether it's an attempt to normalize certain aspects of the form factor to ease people into what Apple sees as the future of this technology.
Namely, I wonder if the tethered battery and lack of dedicated controller exists to prepare people for a future where their headset plugs into an iPhone (or iPad or Mac), which then serves as a battery, co-processor, and input device.
I wonder how much of that is due to the specific nature of this type of product, and how it compares to those past launches.
With iPhone & iPad, you had products that people could very easily imagine using. The iPhone was the combination of three technologies people were already familiar with, the iPad was the iPhone, but bigger.
It is really difficult to imagine using a VR headset if you haven't used one before -- there simply isn't a frame of reference. And even if you have used an Oculus or a SteamVR headset, an interface built around eye & tracking is still completely novel.
I think a lack of mass-market appeal is something that is inherent to a product like this, which makes it a very different product from a marketing perspective than iPhone or iPad.
It makes sense to me that, rather than trying to copy-paste the iPhone or iPad launch, Apple would instead put out an exclusive halo product early, allow a few years for the public to build familiarity with it and settle on the core use-cases, all while they observe and iterate behind the scenes on a more accessible version for a wider audience.
My understanding is "yes, but only type errors that violate guards and pattern matching".
Jose's elixirconf keynote this year did a great job providing details on how how this new type system works:
Counterpoint: The article is titled "HTML First" not "HTML Only"
Admittedly I had the same reaction you did as I was reading the article. All I could think was how poorly these approaches would scale when the need arose for even modest ly complex state management.
While the body of the article doesn't address this, IMO the title does. I think it's generally good advice to suggest only reaching for the tools intended for complex environments when they become necessary, not before.
Writing off a shift towards walkability as a "near-impossible social solution" I think is misrepresenting the nature of the problem.
Shifting car-centric suburbs and exurbs to a walking-friendly lifestyle is a massive challenge, but that only covers 20% of the population.
The remaining 80% of the population already live in urban areas that, in many cases, are already very livable car-free, and are a few small policy changes away from taking a massive step forward.
Even a couple large municipalities legalizing accessory dwellings, abolishing parking minimums, and rezoning to allow business conversations of existing residential properties could have a sizable impact.
Additionally, many of the most expensive areas in the country are ones that are designed specifically to support walkability, which is a strong indicator of unmet demand for such areas.
As an aside, Elixir is doing some interesting, novel stuff exploring its own type system. José Valim's keynote at ElixirConf last week went into detail on the topic, so I'd keep an eye out for it on YouTube in the coming weeks if that sort of thing interests you.
Laughably incompetent.
This may be a silly question as I'm quite inexperienced w/ ML and a lot of the words in this article don't mean anything to me yet:
But, could this "model" be used for something like monitoring land use in a city? The specific example I'm thinking of is getting a percentage breakdown of what land devoted to paved surfaces (parking/roads), to vacant undeveloped lots, and to built structures. It would also be interesting to see how those percentages have changed over time.
I've been using Arc for a few months now and find myself really missing it when I use other browsers.
This is due to a number of things, like the degree of UI polish and a UX I personally find very intuitive, but I think the biggest factor comes down to their features around tab management. Namely the following:
1. Tabs closing automatically by default unless explicitly pinned 2. Being able to group pinned tabs into folders 3. Pinned tabs remembering their original state (e.g. if I pin a tab, then navigate to some other page, I can click the favicon to return to the original page I pinned) 4. The ability to rename tabs to make them more meaningful 5. Organizing tabs into an easily-collapsible sidebar 6. Having multiple spaces for different "contexts"
These features may not be particularly novel, and I'm sure they could be easily replicates w/ extensions or add-ons, but the fact that things work like this by default and are super easy to manage feels pretty fantastic.
As an example, I recently started rewriting my neovim config, and created a new space called "neovim config" so I could quarantine the tab clutter away from my normal browsing. Then I pinned the GitHub page for my old config so I could reference it easily. And lastly I created folders for "installed plugins", "required plugins" and "optional plugins" to help me organize the GitHub pages of all the plugins I've already installed, plan to install, or merely want to test out. Then if I run into any bugs or need to reference Lua syntax I can open a bunch of ephemeral tabs to stack overflow or whatever and easily clean them up once I'm done.
I know I'm probably sounding like an Arc shill at this point, but I do genuinely feel like Arc is helping me get this work done more efficiently whereas I feel like just about any other browser would be getting in my way. The simplest way I could describe it is it's almost like Trello and Safari had a baby that somehow inherited the best qualities of each, with none of the drawbacks.
The fact this thing has an M2 makes me surprised Apple didn't try to sell it as a Mac.
I feel like at the price point, this device makes much more sense as the kind of thing that could replace a laptop/desktop than as a companion to it.
If you can check connect a magic keyboard/mouse to it, this thing could conceivably be a MBA and badass multi-monitor setup rolled into one. And to me, that's really the only way this form factor makes sense.
I also find myself frequently questioning the labeling of things one doesn't like as "political".
Having a code of conduct that focuses on treating people w/ respect and discouraging personal attacks is not political, and it's strange to me that someone would label it as such.
I've been using Arc on my personal machine for a couple months now, and my favorite feature is one that surprised me:
Tab persistence is opt-in
Meaning that, unless you explicitly pin a tab, or stick it in a folder you create, it will get automatically closed after some period of time (in practice for me this meant between browsing sessions).
Initially this surprised and annoyed me, but over time it prodded me to be intentional about evaluating whether my tabs are disposable, or need to be held on to across multiple sessions, which I've found led to much tidier browsing habits.
Other features of Arc (spaces, pop-up previews, folders, tab renaming) aren't particularly novel, but the way they integrate seamlessly and function intuitively with inputs I'm used to, has made the whole thing a really pleasant experience in a way I've never been able to achieve by cobbling together extensions in other browser.
Really hope their ideas catch on and they start to see some competition because Arc is doing a great job bringing novelty while maintaining polish and I'd love to see some of this UX replicated in a non-chromium browser.
Curious how people use search these days.
For me, I'm less and less using a single search engine for all my needs, rather using a plethora of engines for different tasks.
I use DDG as my default engine to ensure I get privacy by default, and if the results aren't to my liking, I use a bang to re-run the search in a different engine.
I also like to set up custom search engines in my browser with shortcuts, so if I need to search a specific GitHub org or something I can quickly run a search there from my URL bar.
A lot of the comments here saying "DDG is dead, switch to xxx instead!" strike me as not particularly relevant with my usage patterns, but I'm wondering whether I'm just in a minority.
My anecdotal hypothesis (at least from a North American perspective) is that we've slowly adopted a culture that favors individualism and growth over community and sustainability.
We had a period of extreme wealth and growth after WWII, and chose to re-make society in ways that were fundamentally unsustainable. Then in order to keep that dream alive, we continuously borrowed from our future, and ignored those who were cut out of this dream, all to prop up an increasingly individualistic eutopia that was doomed to fail.
Now the bill is coming due for all of us. But we've gone too long isolated from our neighbors, ignoring the cries of those for whom the system was failing, and losing our sense of community that we've forgotten how to work with one another to solve our real societal problems.
The world we built ourselves is slowly deteriorating due to the natural progression of time, and we've forgotten how to fix it.
With all the talk of PostgreSQL, I'm curious if anyone has tried the offerings Fly.io has been pushing with SQLite (Litestream & LiteFS) in any serious capacity, and can comment on the experience.
Wondering whether they would be a viable alternative to Postgres in the typical "Database + Docker Container" architecture.
My own self-discipline, and being satisfied with my current lot in life.
Looking at the past few years, my project list has only grown, as I repeat the cycle of getting excited about a new idea, starting to work on it, reaching a snag and getting frustrated at the lack of progress, only to get excited about a new idea and repeat the cycle.
This year I intend to focus a little more about being generally disciplined about all the little daily things, and just being grateful for what I do have today.
My hope is that improved discipline will make sticking to one project easier, and being more content generally will help make the stakes of finishing projects feel less daunting.
And because I recognize this answer isn't totally in the spirit of the question -- I'm also working on building a personal blog site with Elixir/Phoenix/LiveView/Tailwind. I really like the programming model of these tools and how easy it is to build a scalable distributed app with Fly.io, and think building some mastery in these tools will be valuable in the long-run.