jonotime

I have never had to do this in my years of using this scheme. And if I did, I would create an email filter. Which I consider to be more standard and scalable then managing a bunch of individual aliases.

Yeah, I think I only record maybe 10% of them that actually have logins associated. For the others I just search through my email.

Just happened to me today! I was at the Verizon store and my address was verizon@... Sometimes it leads to confusion, but sometimes it leads to getting extra special treatment actually! They think I'm someone important.

I have run this for years with very little problems. And I can honestly say that have not found anyone writing to addresses I did not give them at their domain. Simple as this is, it is way to niche for companies to figure it out and exploit it. And if that really was a problem I'd just create a new subdomain.

If you are worried about privacy, get a domain just for this. Use domain privacy and dont host other things there.

Yes, some sites whitelist domains or dont allow subdomains. For those I'll use another account - or a firefox alias or something. But 9 out of 10 work fine.

I am not a fan of alias services since materializing names takes discipline. How many do you make? Maybe there is a limit of 50. When do you share them across services? My guess is many people just create 2 or 3 aliases they use for everything - which defeats the purpose. Sure, it masks your personal address, but once one gets compromised, you find it basically served as your personal address anyway.

I also dont really keep track of most of the names I use. Since most are one time things that I would never use again, like to sign a waiver or something. But I mostly stick to '{domain}@' for the names. So my nytimes account would just be nytimes@, which is predictable when I need to recover it. I used to use addy.io for this, but it was not as good since it had account limits and I had to manually manage every alias. Much easier for me to just create a mail filter to sinkhole an old name. Of course I have never really needed to do this anyway.

Pro tip for doing something like this without apple. Buy or get a cheap domain name. Create a subdomain on it and have it catch and forward all messages to you when sent to that sub. For example:

nytimes@mailsub.example.com -> jono@gmail

anything-else@mailsub.example.com -> jono@gmail

You dont even need to materialize aliases at all.

This is neat. I dont get how sync works. Is it server side? Or is there some client side oauth flow? I dont see it.

Looking forward to use all these nice AI features without using the warp account/service. So I can bring my own claude and it will show all the agent panes etc.

Trends in Ambient Development

Jono Finger Blog/site: https://www.dgt.is/ Feed: https://www.dgt.is/feed/feed.xml HN: https://news.ycombinator.com/user?id=jonocodes

Would love to use something like this for NixOS as an escape hatch, but its debian only. DistroBox always gives me trouble.

My experience has been quite the opposite. Like OP, I am a middle aged developer. I have decades of experience in the industry and with open source/side projects. I think there is so much good new software being developed by single developers (https://opensourcesecurity.io/2025/08-oss-one-person/). When there is one person its an itch to scratch, not a business plan.

I do think the type of projects have indeed changed however, so maybe OP is looking in the wrong place, or has become steeped in the corporate world (see Middle Aged).

Look at app stores, f-droid, vscode plugins, github repos, frontend frameworks, distributed databases, Linux desktops, TUI apps, text editors. There are more projects than ever. They may not be as low level or as widely used as system kernels or programming languages, but there is so much code being written for personal needs.

I actually think the fact that these are not widely used is important. Higher level constructs and AI have made programming much more accessible. Non-devs are making one off apps just for their family. Or using LLMs to quickly churn out scripts to automate home assistant or common desktop tasks.

I personally have built several projects from my long-running ideas list that I never would have had time to do before without the help of AI. Savr is one (https://github.com/jonocodes/savr). For someone like me who has built that same CRUD interface for every company, AI has been an amazing motivator to work on new interesting things and less repetitive drudgery.

The conversation here has been really interesting since it shows we all have subtly different abilities and issues around slow processing.

I am a senior/staff software engineer with 20 years of experience and who would be considered a top performer. Quick wittedness is something I rely on but technical thinking is always slow. White board interviews are nearly impossible. So finding work is ROUGH.

I wouldn't call my self the class clown, but I am often the first one to crack a joke and people look to me for lighthearted fun. I often hope this will save me in interviews as I rely on my past work and personality are my strong traits. All of the jobs I got through this, or from a personal referral. Never through a technical interview.

My way of coping with this has always been to work more hours then everyone else. It helps that I am also more passionate about what I am working on then those around me. I work on open source in my free time, while my co-workers have barely heard of Linux. If it were not for my love of the craft, I be miserable in this industry.

I first noticed this all when I was a kid and started taking more advanced classes. I’d have to take copious notes and deeply review them after class to make sure I knew what was going on. While my friends would listen once and ace the tests. I was also always behind on reading speed since I had to read things 2 or 3 times before the words really registered in my brain.

As a side note, I recently worked with a real version of the oh-so mythical 10x programmer. And he was a college intern! Running loops around the rest of us at our startup. He was doing the work of several of us older programmers. I think most of it was do to his 0-friction processing of information. Read, assimilate, code. Interesting to watch. They do exist.

For me is has been quite the opposite. Im a top performing senior/staff engineer. But always struggle to get through coding interviews.

Ah ok. Admittedly I dont host a media server so it sounds like Plex brings new challenges.

I would just prefer to not have to public expose a service for a single user. In my case when sharing an image server to family it has been easy enough to walk them through installing tailscale on their windows desktop that they use. I love adding friends and fam to my tailnet. It then also makes it easier to log in and troubleshoot their issues later too.

It looks like CFs solution for restricted public access is CF access controll, but thats still publicly exposed. Their non-public option is WARP, but that requires installation on the client machine. At that point your user setup is even harder then tailscale.

Out of curiosity why not give your sister restricted access to your tailnet instead? Then nothing is public.

There are a bunch and quite a breadth of different solutions/takes on the problem.

Here is a good recap of the current players. https://www.localfirst.fm/landscape

Awesome to see this getting more coverage. I am very interested in local first and I am working on several progressive web apps based around this. One app depends on file sync, not database sync and the best I have found is remoteStorage.js. Its not perfect, but its very much the missing piece I was often looking for.

How do you use only sockets to talk to sqlite?

Oh man. I have been working on a side project just for this purpose. The aim is to create a pocket like experience (with additional functionality like handling other media types) that is local first, unhosted, and more future-proof (no lock in).

All data is stored entirely on your device, and you have the option to sync it to your own storage provider like dropbox. This means you don't need to have the technical know-how to setup and maintain a server.

Its not usable yet, as I have rewritten it several times, but in the current iteration it is a client side PWA, so cross platform. Just started a new job so had to take a break for a bit.

Follow if you are interested (I need to update the Readme): https://github.com/jonocodes/savr

A local first read-it-later app for power users, who want to own their data. https://github.com/jonocodes/savr

+1 for local-first. there needs to be a local first bookmark/snapshot manager. where you can own your data without requiring hosting or installing things.

Can you scam the elderly into having a good time?

Location: San Francisco bay area

Remote: No. Only looking for hybrid in downtown SF.

Willing to relocate: No

Technologies: Python, Typescript, NodeJS, React, Django, Docker, REST, grpc, GraphQL, SQL, AWS, GCP

Résumé/CV: https://www.linkedin.com/in/jfinger

Email: hn0225 at m dot dgt dot is

Role: Senior or Staff Software Dev, full time

Hiring info and reviews from managers: http://dgt.is/professional

Excluded Industries: Finance, advertising, sales, marketing, retail, defense

Preferred Industries: Education, environmental good, social good, animal welfare, medical, science, public transit, open source, developer experience

Blurb:

* 15+ years as backend and full stack developer on teams of all sizes

* Enjoys mentoring engineers and giving tech talks

* Open source contributor, tech enthusiast

* Recently on front page of HN: https://news.ycombinator.com/item?id=42666851

* Will not do leet code technical interviews

Interesting. So in my use, I rarely see gpt get it right on the first pass but thats mostly due to interpretation of the question. I'm ruling out the times when it hallucinates calls to functions that dont exist.

Lets say I ask for some function that calculates some matrix math in python. It will spit out something but I dont like what it did. So I will say, now dont us any calls to that library you pulled in, and also allow for these types of inputs. Add exception handling...

So response time is important since its a conversation, no matter how correct the response is.

When you say deep seek "nailed it on the first attempt" do you mean it was without bugs? Or do you mean it worked how you imagined? Or what exactly?

Any idea what the power draw is? Resting power, vs resting with the model loaded in memory, vs full power computation. In case you may want to also run this as your desktop for basic web browsing etc when not LLMing.

I'm also kind of new to this and coming from coding with ChatGPT. Isnt the time to first token important? He is sitting there for minutes waiting for a response. Shouldnt that be a concern?

Neat. Thanks.

How do you deal with removing a package? For examplet the case where you have htop in your config, but no longer want it on your system or in your configs.

Yup, I have had that several times too.

Oh, I'm interested. Are you using it on servers, or desktop? My concern is the community is small, while Nix's has been booming.