arend321

Thanks for sharing. I hope with Android 16 and Samsung implementing AVF for DeX these issues will be ironed out.

I thought AVF has hardware virtualization where proot is user-space.

It's a very interesting project, but I don't think I'm ready to let go of all my Android apps.

I'm waiting for Android Virtualization Framework to run a full Linux distro on my smartphone with portable monitor (glasses). Already using Termux but AVF is hopefully much more performant. Maybe the Samsung S26 Ultra will have full support. I might ditch my miniPC if this works out.

I like this vibe. As a bootstrapped company making money using open source software, I have no issue paying individual devs, I sponsor multiple projects on GitHub. VC funding, however, changes the game: now a project needs to deliver 100x returns just to survive.

And you don't get surprise updates that trigger a cascading dependency hell.

Will this be monetized with the classic SSO enterprise subscription play? Would be nice if they are transparent on how they plan to make money.

The DX is quite nice, even though not well suited for existing projects as it is hard to migrate existing users. There is no easy way to keep existing sessions or do a legacy login, then migrate a user to the new better-auth supplied hashing function.

On the one hand, it's a testament to the succes of Node.js and the wildly popular npm ecosystem. On the other hand, there is work needed to make things more secure. Perhaps by adding security boundaries (like Deno, but better) or having a more well rounded standard library.

"These tactics were found in malvertising campaigns targeting users with cryptocurrency-themed lures, as well as in phishing-based social engineering schemes." That quote makes it sound less dramatic though, as the crypto field is littered with scam artists and "phishing-based social engineering" transcends Node.js.

That's so cool, thanks for the detailed reply. I agree walking around with a big bouncy screen in front is probably not going to be very ergonomic. Still I can't let the idea go completely. Going RSVP+TTS might indeed be a more viable approach, certainly something to test with the current wave of AI/LLM agents.

Waiting for the XReal One Pro to vibe code using Aider and Samsung DeX in a Termux terminal while walking my countries national parks (just something I would like to try)

Great article that reflects how I see Next.js. Currently on Pages Router but experimenting with App Router doesn't give me confidence it's the direction I want to go. Waiting for TanStack Start start to come out of beta and give it a spin.

I think this is a somewhat mainstream opinion outside the hype train. AI is helpful as a Google replacement but not ready to do serieus context sensitive coding.

Previous discussion: https://news.ycombinator.com/item?id=43317950

I'm not sure how a memory leak in Next.js, a massive project with a "move fast and break things" attitude, reflects on Node.js.

My experience with Node.js (on the backend) and Next.js (since Next.js 10) has been pretty solid, albeit being conservative, only just migrated to Next.js 15, still on pages router.

I'm not saying Node.js is the best thing since sliced bread. Memory leaks in event driven, long-running processes are a thing. I think it might be more about finding a language/runtime that clicks with your mental model. I'm interested to see how Elixir pans out for you in the long run.

I hope we're going to see a move to more cloud agnostic stacks. I'm currently running Patroni between Hetzner and Netcup, with an Amazon Lightsail VPS as a third etcd member. If Hetzner decides to flag my account, the hot standby on Netcup will take over (or the other way around, as I switch on a weekly basis). Still using Amazon with AWS SES and Route53 but these can be replaced by other parties if need be.

"I have never had a long-running node application that hasn't leaked memory. It's more a matter of if your application will be OOM killed in 2 days or 2 weeks. Debugging memory leaks in node is a nightmare"

I stopped reading after this. I'm all for using Elixir for high concurrency workloads, but this sounds like a skill issue.

I'm getting 2586 RPS with the default next.js starter app on a low budget laptop. Something isn't right here.

We need a lot more details to make sense of this benchmark. I did a default Next.js install and performed an Apache benchmark, yielding a 2586 RPS for me on a Lenovo IdeaPad Slim 3 (a low-budget laptop). Are we benchmarking the cache implementation, the JSON parser, or maybe the proxy configuration?

I think so, together with the global downturn around that time.

I see another construction boom happening in Spain. The remaining somewhat unspoiled coastline is once again being covered in concrete. Developers introduced a new time-share variant called "co-ownership" to drive demand.

https://murciatoday.com/stoadmin/how-the-rich-are-snapping-u...

Is anyone on Hacker News using YugabyteDB? While it offers PostgreSQL compatibility, I’ve found that simply converting a standard, normalized PostgreSQL database doesn’t work well in practice. Instead, you need to design your database with YugabyteDB’s distributed architecture in mind.

YugabyteDB—and CockroachDB, for that matter—seem to be somewhat niche. Despite the potential of distributed databases, you rarely hear about people using them.

This probably only works if you glue a bunch of high level, popular APIs together. It might work, but will be fragile and expensive.

I'm actually quite happy with AI crawlers. I recently found out chatgpt suggest one of my sites when asked to suggest a good, independent site that covered the topic I searched for. Especially now that for instance chatgpt is adding source links, I think we should treat AI crawlers the same as search engine crawlers.

I guess it depends on the field you're working in. When manipulating strings, await db calls, sort arrays, JavaScript (TypeScript actually) feels very intuitive and concise. But going more low level I can see it breaking apart.

Definitely looking into TanStack Start/Router, Tanner seems to be a level-headed guy. Still hard to take a left turn where the momentum is in the Next.js direction.

Google does crawl client side generated content but at a low priority and with mixed results. It's expensive and slow to render js pages so I see where they are coming from.

Some of it is confusion about what say a 404 entails, is it the endpoint that does not exist, or that say /api/users returns an empty array for the given search parameters. Some libraries even throw an error if they receive a 404. Probably ambiguity in the HTTP spec.

That's indeed the pain point. Distributing a stateless app is relatively easy. Distributing the shared file system and database over a remote, higher latency, cross-cloud setup is hard.

The effects of these weather events are much more severe due to increased development in coastal areas, often in places where houses should not be built.

https://www.eldiario.es/sociedad/spain-at-flood-risk-over-mi...

I guess I'm lucky and 10 minutes of downtime is not a problem for my customers during off-peak hours, every year or so. Upgrading has always been a painless experience with just a pg_dump and pg_restore on the upgraded db server.