I feel like reading HN sometimes there is the assumption that everyone who is a programmer by default works on web stuff (front end&#x2F;back end).I&#x27;m curious to hear about what other jobs&#x2F;domains exist outside of this and how it is working on non-web stuff.

Ask HN: Who here is not working on web apps/server code?

It&#x27;s the time of the year again, so I&#x27;d be interested hear what new (and old) ideas have come up.
Previously asked on:2024 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=423733432023 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=384676912022 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=341904212021 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=296670952020 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=249471672019 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=208998632018 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=177903062017 → https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=15148804

Ask HN: Those making $500/month on side projects in 2025 – Show and tell

- HN errored on all authenticated requests with 502 Bad Gateway. It did still respond to a limited amount of unauthenticated requests with presumably cached pages, which did not get updated. The last post on &#x2F;newest claimed &quot;0 minutes ago&quot;, but was actually much older (1:32:57 PM GMT) and not the newest post.- This status page actually identified the outage: <a href="https:&#x2F;&#x2F;hackernews.onlineornot.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;hackernews.onlineornot.com&#x2F;</a> - Pages by Hund and Statuspal did not show the outage.- The last post before the outage was <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46301823">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46301823</a> (1:39:59 PM GMT). The last comment was <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46301848">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46301848</a> (1:41:54 PM GMT).- There was an average of ~4 seconds per comment just prior to the outage. Based on this, HN likely went down at 1:41:58 PM GMT.

Tell HN: HN was down

I’ve been working in analytics and advanced analytics for about 6 years. I started in a large consultancy and later went solo, so I’ve seen both enterprise and smaller product teams up close.Something keeps bothering me. In most projects, analytics feels like infrastructure that no one is genuinely excited about. People rarely want to invest in it when building a product. It’s treated as something you should have, not something you want to have.Teams are “happy” to pay for software development, advertising, copywriting, design. Those are seen as directly useful. Analytics (GA4, event tracking, or even more structured setups like CDPs) is often perceived as background noise, necessary to keep the engine running, but not something that meaningfully moves the product forward day to day.In practice, many teams end up using only a handful of metrics to make decisions, even when a complex analytics stack exists underneath. The rest is there “just in case.”I’m curious whether others see the same pattern. Is analytics undervalued because its ROI is indirect and delayed? Or is most analytics work simply over-engineered for the actual decisions teams make? At what point does analytics shift from “necessary plumbing” to a real competitive advantage?Would love to hear perspectives from founders, engineers, and product folks who’ve built and scaled things.

Is analytics a necessary evil rather than a real value driver?

We&#x27;re a very small startup with a web-based application written in React. Every non-engineer in the company spends a lot of time making screens in Lovable and thinks we can and should just &quot;take that&quot; and drop it on the site. We fight back, but even when we just use it as a design inspiration we&#x27;re left with terrible UX, vastly different pages and a genuinely bad product that&#x27;s just good enough not to be seen as a problem by the rest of the team. Help me make the case that we need a product designer, graphic designer, and to use these tools for what they&#x27;re worth.

Ask HN: How to fight back against Lovable, Replit, etc. in enterprise products

When working on my projects and talking to investors, I often hear:
“Just post it on Hacker News or Reddit and show that people love it.”What I find strange is that Hacker News feels oddly opaque. I’ve never met anyone who can clearly explain how it works in practice. Not just the rules, but the dynamics: what’s repeatable, what’s luck, and what actually matters.By using the Kevin Bacon-number idea: I can usually get within three degrees of separation of well-known technologists like Linus Torvalds, but I can’t seem to get within three steps of someone who confidently understands how HN works.So I’m asking sincerely:
Does anyone here feel they understand Hacker News?
If so, what are the real levers, and what do people consistently misunderstand?PS: This question comes from a mix of genuine curiosity and personal frustration. I’m honestly trying to understand how HN works in practice.

Ask HN: Does anyone understand how Hacker News works?

I&#x27;m an indie developer who has built several free-to-use tool applications that I plan to monetize through advertising once they gain enough users. My dilemma is whether to open source these products now, or wait until they&#x27;re more established.

Ask HN: Should I Open Source Every Product I Build as an Indie Developer?

What are you working on?  Any new ideas that you&#x27;re thinking about?

Ask HN: What Are You Working On? (December 2025)

Is there any specific program &#x2F; college &#x2F; etc outside of tech that you would want to study if you could?

Ask HN: If you had to get a non-tech masters degree, what would you go for?

Working on automating small business finance (bookkeeping, reconciliation, basic reporting).One thing I keep noticing: compared to programming, accounting often looks like the more automatable problem:It’s rule-based
Double entry, charts of accounts, tax rules, materiality thresholds. For most day-to-day transactions you’re not inventing new logic, you’re applying existing rules.It’s verifiable
The books either balance or they don’t. Ledgers either reconcile or they don’t. There’s almost always a “ground truth” to compare against (bank feeds, statements, prior periods).It’s boring and repetitive
Same vendors, same categories, same patterns every month. Humans hate this work. Software loves it.With accounting, at least at the small-business level, most of the work feels like:normalize data from banks &#x2F; cards &#x2F; invoicesapply deterministic or configurable rulessurface exceptions for human reviewrun consistency checks and reportsThe truly hard parts (tax strategy, edge cases, messy history, talking to authorities) are a smaller fraction of the total hours but require humans. The grind is in the repetitive, rule-based stuff.

Tell HN: AI coding is sexy, but accounting is the real low-hanging target

This is something I’ve noticed about myself and I’m curious about the underlying mechanism.Whenever I receive a letter or email that looks official — from a bank, government office, school, etc. — I feel a spike of anxiety almost immediately. What’s strange is that this happens before I even know what the message actually says.In most cases, it turns out to be routine or harmless. Still, the anxious reaction comes first and sometimes lingers until I fully understand the content.I’m not looking for personal advice, but I’m curious whether this response is tied to how humans process authority, uncertainty, or perceived consequences. Is there any psychological or cognitive explanation for why the anticipation alone triggers stress, even when experience says it’s usually nothing serious?

Ask HN: Why do official-looking emails cause anxiety before I read them?

I have a media server project that I want to work on. But I&#x27;m stuck on one thing, convenience vs privacy.As the project is about spinning your own server (media server), I want to have a smoother way to have a simple account system where the user just enters an email and a password, and get the server&#x2F;ip list (everything from there is done on the actual server). For example, a user could be invited to 2 servers, and would see them in the same page, which makes things more straight-forward and a lot easier.Now, I thought a lot about it, and mostly came down to the conclusion that centralizing it is the most sane option. The data itself comes down to: email, encrypted password, encrypted IP(s) list (via key exchanges).Is there any-way to do it decentralized? I searched, even asked LLMs, but nothing felt solid (best was a Nostr suggestion) but such method would make emails, password resets painful or almost impossible. I don&#x27;t know a lot regarding this topic so its quite the challenge.What&#x27;s the point&#x2F;why not just use URL? convenience. I know, but it SUCKS having to give a parent a URL, even with some techy friends it takes a bit communicating it. I want to eliminate as much friction as possible. Also, if centralized, this has the ability that users don&#x27;t need to buy a domain, setup lets encrypt and all that which costs money and time (especially for simple&#x2F;new selfhosters); its a lot nicer and smoother and in a way provide better privacy out-of-the-box.Note, This project doesn&#x27;t even exist yet. But I&#x27;m pursuing quite soon. I also only took 1 encryption course back in college days, while I understood and was good at it, I still need to audit&#x2F;verify my method. 
It basically is: 
1. hash the password+salt in a different algorithm, save the private key from it and send the public key to the central server 
2. (media server owner wants to invite) the media server checks for a public key, encrypts the message containing all the details (IP, status, ports etc), and sends the encrypted message to the central server. 
3. The client later checks, if there&#x27;s a new message, it decrypts the ip&#x2F;info from the server and connect.Every device can login in this way and grab server list info securely. There&#x27;s gonna be some sort of way to &quot;quick connect&quot; on TVs and such, and change passwords, but I don&#x27;t want to get ahead of myself for now. I don&#x27;t think the IP&#x2F;server-info encryption suffers from any major things, but that&#x27;s the general core principle. I maybe (probably?) have missed something.The only issues I can maybe think of, is a &quot;centralized&quot; URL&#x2F;domain would be showing up all the time instead of the owner. Note, it would be designed in a way that would allow you to instead send them to your own URL&#x2F;domain and such.Anyways, let me know what would be best. btw I&#x27;m not rich but such simple &quot;auth&quot; server would probably cost like $5&#x2F;m + 2x5&#x2F;m for redundancies, shouldn&#x27;t be too bad.

Ask HN: Decentralized Auth for Information Exchange?

I struggle to find the right way to provide feedback on pull requests (PRs) that mostly consist of AI generated code. Co-workers submitting them have learned to disclose this -- I found it frustrating when they didn&#x27;t -- and now say they have reviewed and iterated on it. But often the result is still what I would describe as &quot;a big contribution off the mark&quot;, meaning a lot of code that just follows the wrong approach.Usually, when someone does a lot of work, which we used to be able to measure in lines of code, it would seem unfair to criticize them afterwards. A good development process with ticket discussions would ensure that someone doesn&#x27;t do a lot of work before there is agreement on the general approach. But now, with AI, this script no longer works, partially because it&#x27;s &quot;too easy&quot; to do it before even deciding this.So I&#x27;m asking myself and now HN: is it OK to point out when an entire PR as such is garbage and should simply be discarded? How can I tell how much &quot;brain juice&quot; a co-worker has spent on it and how attached they might be to it by now if I don&#x27;t even know whether they even know the code they submitted or not?I have to admit that I hate reviewing huge PRs and the problem with AI generated code is that often it would have been much better to find and use an existing open-source library to get the task done rather than (re-)generate a lot of code for it. But how will I know this until I&#x27;ve actually taken the time to review and understand the big, new proposed contributions? And even if I now do spend the time to actually understand the code and implied approach, how will I know which part of it reflects their genuine opinion and intellect (which I&#x27;d be hesitant to criticize) and what is AI-fluff I can rip apart without stepping on their toes? If the answer is &quot;let&#x27;s have a meeting&quot;, then I&#x27;d say the process has failed.Not sure there is a right answer here, but I would love to hear people&#x27;s take on this.

Ask HN: Etiquette giving feedback on mostly AI-generated PRs from co-workers

I&#x27;ve been on a business trip this week, and sitting on the plane, I was randomly browsing NYT and read this article which really REALLY pissed me off: https:&#x2F;&#x2F;www.nytimes.com&#x2F;2025&#x2F;12&#x2F;14&#x2F;us&#x2F;fire-department-software-private-equity.html?unlocked_article_code=1.8k8.yzeR.NwSK3PTQeXQj&amp;smid=nytcore-android-shareTLDR: Private equity is buying up all these software vendors, and preying upon vulnerable fire departments; one rural department with a total budget of $130k&#x2F;yr saw their software costs go up 3x in one year (from $4k&#x2F;yr to $12k&#x2F;yr).I want to do something. I have not worked in the emergency services domain before, so I know that I am naive. I do have experience with oncall and mission-critical software maintenance and highly available&#x2F;durable services, but when I say mission critical I mean millions-of-dollars at risk, not human lives.I want to solicit HN for very early directional advice, of any and all kinds.- Are there some existing open source projects or foundations with a mission like this, to serve and protect critical services like emergency response?- Is it feasible to build incrementally, i.e. ship small components of incremental value, and would fire departments be open to trying an open source option?- Are there things like certifications or regulatory hurdles, that might be expensive to get?- Does anyone randomly happen to have experience in this industry, and would be willing to write up their perspective?My early initial thesis is something like this: I personally cannot stomach the idea of simply passively &quot;letting&quot; private equity invade this aspect of society, to hold emergency services departments and by extension taxpayers hostage. They are essentially extorting and exploiting society&#x27;s inelastic demand for emergency fucking services. So I don&#x27;t care about making profit, I only care about actually solving the needs of the emergency departments and of protecting human life (knowing that I am naive still, I definitely do not want to overpromise and put lives at risk). But my personal feelings aside, I do think there is a useful insight here: if useful software can be made reliably and cheaply, by engineers motivated to help their communities instead of profit (or just to poke private equity in the eye), then ultimately private equity cannot compete against an adversary that doesn&#x27;t care about profit -- we actually CAN WIN, because eventually they will give up and exit these industries if they see that there is no profit to squeeze.I am looking for like-minded builders. I seriously think the software foundation could simply be a vehicle for early-retirement rich FIREd ex-techies, to add some new purpose in their lives, by using their skills to solve a targeted and well-scoped set of problems.&quot;Some men aren&#x27;t looking for anything logical, like money. They can&#x27;t be bought, bullied, reasoned, or negotiated with. Some men just want to watch the world [NOT] burn [because the fire departments should have cheap reliable emergency services software].&quot; -- Alfred Pennyworth, The Dark Knight&quot;Beware the fury of a patient man.&quot; -- John DrydenI am a patient man, with a newly gained cold fury. I am willing to write code. Or organize a software foundation. Or be an investor&#x2F;funder for someone else to write the software. Or whatever form of help I can give.Appreciate any and all help or feedback (positive or negative)!

Ask HN: Should I start a software foundation (goal: help emergency services)?

I read recently a post here in HN about this topic but I can&#x27;t find it anymore so I&#x27;m writing here my thoughts.Recently, I started a hobby project in Rust that collects some news from several websites and creates me a weekly digest.Since I&#x27;m making it for fun I started developing a scraper.However, one simpler option was to use the websites&#x27; RSS feed. So, I looked for them as well but few websites have them.Now, with the advent of Agentic AI they seem to be an old fashioned way that is needed anymore.What do you think?

Ask HN: Is RSS Still Alive?

I’m releasing *Framework Shells* (`fws`): a standalone Python package for orchestrating long-running background processes (“shells”) with *PTY*, *pipes*, and *dtach* backends.This is meant for environments where you don’t want to stand up a full supervisor stack (or don’t have one): quick multi-service prototypes, dev environments, constrained userlands, etc.### What it does* Spawn&#x2F;manage shells with:<pre><code> * **PTY**: interactive terminal sessions (resize, input, stream)
 * **Pipes**: stdin&#x2F;stdout&#x2F;stderr streams (good for daemons&#x2F;LSPs)
 * **dtach**: persistent sessions you can attach&#x2F;detach to (survives manager restarts)</code></pre>
* *Runtime isolation* (the big feature): shells are namespaced by
 `~&#x2F;.cache&#x2F;framework_shells&#x2F;runtimes&#x2F;&lt;repo_fingerprint&gt;&#x2F;&lt;runtime_id&gt;&#x2F;...`
 so two clones of the same repo can run concurrently without cross-adoption or cross-control.
* *Control surfaces*: CLI + optional FastAPI&#x2F;WS UI for listing, logs, and lifecycle actions.
* Optional *hooks* for host integration (external registries&#x2F;telemetry).### CLI quickstart```bash
# list shells
fws list# run a one-off shell (no spec)
fws run --backend pty --label demo -- bash -l -i# apply a YAML shellspec (recommended)
fws up shells.yaml# terminate shells
fws down# attach to a dtach-backed shell
fws attach &lt;shell_id&gt;# show managed shells + procfs descendants
fws tree --depth 4
```### Shellspec example```yaml
version: &quot;1&quot;
shells:
 worker:
 backend: proc
 cwd: ${ctx:PROJECT_ROOT}
 subgroups: [&quot;worker&quot;, &quot;project:${ctx:APP_ID}&quot;]
 command: [&quot;python&quot;, &quot;-m&quot;, &quot;your_module.worker&quot;, &quot;--port&quot;, &quot;${free_port}&quot;]
```### Isolation + security model (simple by default)* `FRAMEWORK_SHELLS_SECRET` derives the `runtime_id` (namespace) and API tokens.
* If the secret is set, mutating API endpoints require:<pre><code> * `Authorization: Bearer &lt;token&gt;` (or `X-Framework-Key`).</code></pre>
* If the secret is unset, auth is disabled (dev mode).Hard limit: if two runtimes share the same OS user&#x2F;UID, the OS may still allow signaling each other’s processes. The guarantee is: no cross-count&#x2F;adopt&#x2F;control *through the library’s control plane*.### Real-world usageI use this as the substrate of a full dev environment where “apps are shells” (terminals, IDE + LSP, agent&#x2F;MCP, aria2 RPC, file explorer, llama.cpp runner, etc.). Repo:```text
https:&#x2F;&#x2F;github.com&#x2F;mrsurge&#x2F;termux-extensions-2
```### Feedback I want* Does the secret&#x2F;fingerprint&#x2F;runtime isolation contract feel right?
* Any obvious foot-guns in the default API&#x2F;CLI?
* Expectations vs systemd&#x2F;supervisord&#x2F;tmux&#x2F;dtach: where would you use this?github.com&#x2F;mrsurge&#x2F;framework-shellspip install &quot;framework-shells @ git+https:&#x2F;&#x2F;github.com&#x2F;mrsurge&#x2F;framework-shells@main&quot;```bash
fws --help
```

FWS – pip-installable embedded process supervisor with PTY/pipe/dtach back ends

Hi HN — I’ve wanted to start a personal blog for a few years, but I keep hesitating.I write a lot privately (notes, mini-essays, thinking-through problems). Paul Graham’s idea that essays are a way to learn really resonates with me. But I rarely publish anything beyond occasional LinkedIn posts.My blockers:•“Nobody needs this” &#x2F; “It’s not original”•“AI can explain most topics better than I can”•A bit of fear: shipping something that feels naive or low-signalAt the same time, I read a lot of personal blogs + LinkedIn and I do get real value from them — mostly from perspective, lived experience, and clear thinking, not novelty.For those of you who blog (or used to):•What made it worth it for you?•What kinds of posts actually worked (for learning, career, network, opportunities)?•Any practical format that lowers the bar (length, cadence, themes)?•If you were starting today, what would you do differently?I’m not trying to build a media business — more like building a “public notebook” that compounds over years.

Ask HN: Is starting a personal blog still worth it in the age of AI?

I’ve been working on a small language learning app as a solo developer.I intentionally avoided gamification, streaks, subscriptions, and engagement tricks.
The goal was calm learning — fewer distractions, more focus.I’m starting to wonder if this approach is fundamentally at odds with today’s market.For those who’ve built or used learning tools:
– Does “calm” resonate, or is it too niche?
– What trade-offs have you seen when avoiding gamification?Not here to promote — genuinely looking for perspective.

Ask HN: Is building a calm, non-gamified learning app a mistake?

Here&#x27;s a video from 2004 <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=Y1J9kfDCAmU" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=Y1J9kfDCAmU</a>It&#x27;s his 100th birthday today.

Computer animator and Amiga fanatic Dick van Dyke turns 100

What are your predictions for this coming year?

Ask HN: What are your predictions for 2026?

I&#x27;ve been working on a personal project recently, rewriting an old jQuery + Django project into SvelteKit. The main work is translating the UI templates into idiomatic SvelteKit while maintaining the original styling. This includes things like using semantic HTML instead of div-spamming, not wrapping divs in divs in divs, and replacing bootstrap with minimal tailwind. It also includes some logic refactors, to maintain the original functionality but rewritten to avoid years of code debt. Things like replacing templates using boolean flags for multiple views with composable Svelte components.I&#x27;ve had a fairly steady process for doing this: look at each route defined in Django, build out my `+page.server.ts`, and then split each major section of the page into a Svelte component with a matching Storybook story. It takes a lot of time to do this, since I have to ensure I&#x27;m not just copying the template but rather recreating it in a more idiomatic style.This kind of work seems like a great use case for AI assisted programming, but I&#x27;ve failed to use it effectively. At most, I can only get Claude Code to recreate some slightly less spaghetti code in Svelte. Simple prompting just isn&#x27;t able to get AI&#x27;s code quality within 90% of what I&#x27;d write by hand. Ideally, AI could get it&#x27;s code to something I could review manually in 15-20 minutes, which would massively speed up the time spent on this project (right now it takes me 1-2 hours to properly translate a route).Do you guys have tips or suggestions on how to improve my efficiency and code quality with AI?

Ask HN: How can I get better at using AI for programming?

Noticed  how C# is underrated. About memory safety in C#. How difficult to introduce multi-paradigm memory safety approach like Rust? Ownership model for example, would it be possible to enforce practice via some-sort of meta framework?

Memory Safety in C# vs. Rust

Here’s how we’re working with LLMs at my startup.We have a monorepo with scheduled Python data workflows, two Next.js apps, and a small engineering team. We use GitHub for SCM and CI&#x2F;CD, deploy to GCP and Vercel, and lean heavily on automation.Local development:
Every engineer gets Cursor Pro (plus Bugbot), Gemini Pro, OpenAI Pro, and optionally Claude Pro. We don’t really care which model people use. In practice, LLMs are worth about 1.5 excellent junior&#x2F;mid-level engineers per engineer, so paying for multiple models is easily worth it.We rely heavily on pre-commit hooks: ty, ruff, TypeScript checks, tests across all languages, formatting, and other guards. Everything is auto-formatted. LLMs make types and tests much easier to write, though complex typing still needs some hand-holding.GitHub + Copilot workflow:
We pay for GitHub Enterprise primarily because it allows assigning issues to Copilot, which then opens a PR. Our rule is simple: if you open an issue, you assign it to Copilot. Every issue gets a code attempt attached to it.There’s no stigma around lots of PRs. We frequently delete ones we don’t use.We use Turborepo for the monorepo and are fully uv on the Python side.All coding practices are encoded in .cursor&#x2F;rules files. For example: “If you are doing database work, only edit Drizzle’s schema.ts and don’t hand-write SQL.” Cursor generally respects this, but other tools struggle to consistently read or follow these rules no matter how many agent.md-style files we add.My personal dev loop:
If I’m on the go and see a bug or have an idea, I open a GitHub issue (via Slack, mobile, or web) and assign it to Copilot. Sometimes the issue is detailed; sometimes a single sentence. Copilot opens a PR, and I review it later.If I’m at the keyboard, I start in Cursor as an agent in a Git worktree, using whatever the best model is. I iterate until I’m happy, ask the LLM to write tests, review everything, and push to GitHub. Before a human review, I let Cursor Bugbot, Copilot, and GitHub CodeQL review the code, and ask Copilot to fix anything they flag.Things that are still painful:
To really know if code works, I need to run Temporal, two Next.js apps, several Python workers, and a Node worker. Some of this is Dockerized, some isn’t. Then I need a browser to run manual checks.AFAICT, there’s no service that lets me: give a prompt, write the code, spin up all this infra, run Playwright, handle database migrations, and let me manually poke at the system. We approximate this with GitHub Actions, but that doesn’t help with manual verification or DB work.Copilot doesn’t let you choose a model when assigning an issue or during code review. The model it uses is generally bad. You can pick a model in Copilot chat, but not in issues, PRs or reviews.Cursor + worktrees + agents suck. Worktrees clone from the source repo including unstaged files, so if you want a clean agent environment, your main repo has to be clean. At times it feels simpler to just clone the repo into a new directory instead of using worktrees.What’s working well:
Because we constantly spin up agents, our monorepo setup scripts are well-tested and reliable. They also translate cleanly into CI&#x2F;CD.Roughly 25% of “open issue → Copilot PR” results are mergeable as-is. That’s not amazing, but better than zero, and it gets to ~50% with a few comments. This would be higher if Copilot followed our setup instructions more reliably or let us use stronger models.Overall, for roughly $1k&#x2F;month, we’re getting the equivalent of 1.5 additional junior&#x2F;mid engineers per engineer. Those “LLM engineers” always write tests, follow standards, produce good commit messages, and work 24&#x2F;7. There’s friction in reviewing and context-switching across agents, but it’s manageable.What are you doing for vibe coding in a production system?

Ask HN: How are you vibe coding in an established code base?

Just trying to work out if it was just my user account (used for monitoring here: <a href="https:&#x2F;&#x2F;hackernews.onlineornot.com&#x2F;incidents&#x2F;yaz-eOJeARBL" rel="nofollow">https:&#x2F;&#x2F;hackernews.onlineornot.com&#x2F;incidents&#x2F;yaz-eOJeARBL</a>) or if others saw it too.

Ask HN: Was HN just down for anyone else?

I’m using gpt-5.1-codex-max comfortably for coding and hitting the weekly limit sometimes (but a few extra credits usually cover it).I’ve heard Opus 4.5 might be better for coding. SWE-bench shows an 8% improvement but I&#x27;m having a hard time guessing what kind of effect that maps to in reality. For those who’ve switched, what changes have you seen, and how has it affected your work? Is the $100&#x2F;month upgrade worth it?

Ask HN: Claude Opus 4.5 vs. GPT 5.1 Codex Max for coding. Worth the upgrade?

I would like to start a personal blog. I think Substack is a good option, but I would like more control over styling (potentially custom components) and want to host the blog on my own website.I wanted to ask what the writing and hosting process is like for people who have a personal website and blog—do you just write markdown and then use a renderer?I would like a kind of wysiwyg editor to see exactly how the content will appear once loaded. The issue with writing in a separate editor is that the line breaks, line lengths, font, etc. never appear how they will actually look. Thanks!

Ask HN: Bloggers, how do you manage your content?

I’m curious about people’s experiences with pull request code reviewers, both human and automated.For those who have used PR review tools or bots:What did you actually find helpful?What ended up being annoying or counterproductive?At what point did you stop paying attention to the feedback?I’m particularly interested in how people balance signal vs noise in reviews, and whether summaries, inline comments, or opt-in depth work better in practice.Asking to understand real usage patterns and pain points, not to promote anything. Happy to just listen.

Who has enjoyed using PR code reviewers? What worked and what didn’t?

I was sad to discover today that all of my Dall-E image generations are gone, along with the entire https:&#x2F;&#x2F;labs.openai.com&#x2F; site. Apparently, some users received emails earlier in the year when it was about to be taken down, but I didn&#x27;t. There were quite a few images in my history that I would have liked to have saved.Maybe worse is how much this lowers my trust in OpenAI even further than it already had been. Dall-E was not a small platform; it was a cultural phenomenon accessed by hundreds of millions of users. It&#x27;s bewildering that OpenAI would so silently &quot;take it behind the shed.&quot;I&#x27;m searching, and it doesn&#x27;t seem as if there was even an HN post about the shutdown. So, it didn&#x27;t even hit this place&#x27;s radar. How many of you are hearing this for the first time?I don&#x27;t understand how a company like OpenAI could be so reckless with user data integrity and access, particularly when sunsetting a product. All of the big-dog tech platforms have fairly robust protocols for notifying users and allowing them to download their data (even with hoops to jump through). How can they hope to be one while still acting like a &quot;move fast and break things&quot; startup? I liked the thing they broke.

Ask HN: Did anyone else notice that the OpenAI Labs website was completely gone?

I read many non-fiction books, but recently noticed that only a few qualify as truly heavy, thought-provoking reads, that you literally can&#x27;t finish in a manageable time because you keep telling yourself, &quot;Wait a minute,&quot; then stop to Google something, run an experiment, or just think deeply. My current example (still unfinished) is &quot;Moonwalking with Einstein&quot; by Joshua Foer. It&#x27;s mind-blowing - the entire memory universe around us that I never properly explored before.

Ask HN: Thought-Provoking Books

So, I got Pixel 10 Pro, and I&#x27;d like to run parakeet (or whisper) model on it for voice to text. I&#x27;m building an ai dictation app (aidictation.com). I&#x27;m struggling to find a way to run this model on device. I have to reserve to use groq API, which is suboptimal.Any recommendations?