rdsubhas

And where else in law is it allowed to lie or omit material facts?

It's a bedrock foundation of the entire financial and legal world. Public and private alike. Refer SEC Rule 10b-5 as well as Private Placement Memorandum (PPM). For example, Theranos never filed an S1 but still was convicted.

Thank you for docker.

Went from the good old days of vagrant (xen/kvm) to lxc/lxd to docker, and docker truly clicked all the way from development to production.

But honestly, thanks to Docker hub even more. The CLI is only one part of the puzzle, the worldwide registry is an immense undertaking.

You're ignoring and severely underestimating social responsibility and social consequences.

For all we know an AI bot wrote the above comment as well.

This is a worrying trend on HN. Nearly for all submissions, there is one unquantified, unproven comment at the top saying it's AI, with just emotion and no proof. This is the new witch hunt, or karma farming without having to contribute anything positive to the discussion.

Yes I think the major story here is Google removing and then re-adding it back. Big companies have big egos, so this is quite a miracle.

I have a feeling now JXL is here to stay, finally.

I think you're taking it too far.

Exactly 48 hours (before and on the day of election), banning electoral advertising is the right thing, gives voters the mental space to make up their own decision. It's the silence period.

Add much as I hate the lead up to elections, I love the day before and on, where it becomes purely personal.

The DoD buys from suppliers, who buy from suppliers, and so on. Hence supply chain. Hence supply chain risk.

Alternative app stores are mandated atleast here in the EU I believe?

> "$0.23 to under $0.24 - Qwen3.8-Flash-Next"

There are only 6 categories in the site, but a category entirely created out of thin air for Qwen3.8-Flash-Next?

Literally "$0.23 to under $0.24" means EXACTLY $0.23.

grep is deterministic. Llm is probabilistic. Llm can be transferred to a tiny quantum cpu or a lower precision float.

When the author wrote Llm can be as cheap as a tool, I read it as not equivalent. They even said the Llm can be embedded into a tool.

Their point was, the higher level use case — like classification — could become as cheap as grep. Which is quite well possible.

Your article was well worth the read. Thank you for that. It's merely a HN trend where lazy people who can't read an article bad-mouth it as AI. You'll find the lazy AI comment in every article's top first or second comment now. Don't bother justifying yourself, keep it up and keep writing.

Here we go again. Commenter doesn't read an article because it doesn't fit their style. Immediately bad-mouths it as AI.

Edit: typo

> what I'd REALLY F'IN LOVE to see go away is the passwordless/magic link auth flow

This seems really naive? That's the only flow that's at the basis if you get locked out. What else, do you put people on the phone to verify people by asking their name and date of birth? That's even worse!

> Why, for you as a service provider, are passkeys not just better (or at least equivalent to) passwords?

Why you missed the last line of my answer? How did you get an impression to go back to broken passwords?

It's the stupidity of Service Providers to adopt passkeys in the first place.

As a service provider myself, I've evaluated and said "Nah" to passkeys - because it's simply increased Customer Service contacts I have to invest in, whenever a user changes or loses devices, or any of the hundreds of ways Passkeys are not portable.

And guess what, the Tech companies pushing this have zero liability for user login support or security breaches. It's always me. There is no need for me to work hard and spend CS contacts, to wall off my users to the OS or Browser vendor.

I'll simply do passwordless Email or SMS 2FA / Magic Links and own my users without the overhead of Customer contacts, thank you.

They are not selling the information. They are selling a service for easy access to that information.

These are two different things.

Note: am not an AI fanatic.

Everyone here saying about execs and so on - don't seem to catch a few nuances in this chain.

1. Bill Gates is doing the "Internal Disruptor" CEO style

2. Everyone is reacting in a way Bill Gates expects them to react

3. The results of his disruptor style is that Windows is absolutely further unbearable and horrendous, as seen today own the line.

Make no mistake: "Internal Disruptor" CEOs kill their own products and people from the inside, rather than competing or building good products outside.

It's the inverse. We created a world for introverts only in the last ~50 years, where people can spend all their time in an office desk not actually talking to anyone. This was never the case in any of human evolution before.

It's only going back. Raw urban human connections will be the only thing left to exercise.

I treat any organizational theory that has a unidimensional view - like coordination is a good-to-bad slime mold – as incompetent, or at worst, corporate sabotage.

Any active product or business development - builds product/domain/operational depth. That depth feeds back into every decision.

Without taking into account that depth, complexity, competence of Managers and Employees in that depth, headcount invested into automation, etc, just focusing unidimensionally on some behavioral or organizational methodology, IMHO, makes it worse.

It gives CxOs some cute stories to blame someone or something against. If your coordination is slow, then there is a reason behind it. Find it.

Ops here. Don't use musl if you care about disk size or compatibility either. For reference:

docker.io/node:26-alpine - 61MB

docker.io/node:26-trixie-slim: 84MB

gcr.io/distroless/nodejs26-debian13: 55MB

Before this post, we have issues with different, slower DNS resolution in musl.

In 2026, there in no reason other than self-inflicted compatibility and performance pain to use musl in production services.

It has it's place, for binary cross-OS distribution, and generally having an alternate C STL. But avoid for SaaS.

Yep. The system in general is great for treatment, but fails for diagnosis. It takes them a long time to diagnose, sometimes years. But once they manage to do it, then they bring in all the heavy artillery and the health care machine works at it's best.

Sadly, it's often too late. Looks to be the case here, appears clearly a diagnosis delay. It does not take much for a skilled system to connect "airport worker" and the right lab readings together.

I can't imagine what these people would have gone through. They would have been given the "just a flu" treatment for some days, and then waited many more days for appointments. RIP.

PS: What you've said is pretty eye opening! This would amplify a lot of the diagnosis problems.

There is a reason for it. WebP was designed for a narrow-use: optimize for web. It doesn't support HDR, limited color gamut. e.g. All the devices that you use today are taking and viewing HDR pictures, or Print workflows, webp has no role to play in it.

This is not enough for companies to invest heavily in.

JXL is broad use/ universal. High bit depth, wide gamut, different color spaces — from HDR displays, to web, to print to medical photography and so on. It would help the device that your viewing this on, right now.

It's not like WebP where the only focus is to reduce file size. The fact that it does this also, and also recompresses jpegs lossless, is just an addon, not a singular reason.

If you get a chance, ignore all the file size noise, and look at JXL for what it is and how it's being adopted by different industries.

Depth. Wider color gamut, HDR, the things our devices now capture and render.

And it supports animation by default, so the "motion photos" will be natively shareable.

Also, recompressing JPEG as JPEG-XL gives a free boost, so simply, why not. If you're sharing something in social media or anywhere, there is no reason to not convert and have the other person receive JPEG-XL by default.

My point is less about the size and so on. But what the visual (bit depth and color gamut) and functional leap of JPEG-XL. It's a format that combines and replaces the many different use cases that different formats are doing now.

JXL is one of the technologies that I hope we can fully transition over, i.e. in a couple of years nobody (even non-tech-savvy people) are sharing or copying or saving JPEGs.

How do you folks code at 40-50 tps? With an extremely lightweight harness (pi) and just 8k system and tools context, and ~40tps on qwen 3.8 27B 4-bit on low thinking mode, it still takes me nearly 30-45 mins for a basic coding session...

Does it work? yeah... But I'd pick a subscription anyday...

This is exactly my point. I've been writing with those dashes (em or en, whatever, just Alt+hyphen on mac instead of two hyphens). And now this is what people and tools use as the judging symbol for AI? That's just one example, AI text doesn't come from nowhere, it's trained on us, people.

I'm seeing a worrying trend on HN. Nearly for all articles, there is one unquantified, unproven comment at the top saying it's 100% AI — no proof, just baseless emotion of what fits the commenter's writing style. This is the new witch hunt, or virtue trolling.

Look at the later comments, they have substance oriented discussions.

HN Mods - can you please consider a policy against such comments, it's overflowing the site and is diluting discourse and value. If people don't like an article, they can simply ignore it. These articles are reaching the top because enough people consider it of value.

But when an article comes to the top, and the topmost comment and discussion thread is an unqualified witch hunt, it's getting sick.

How can one set topp and temperature in pi?

I'm on-call all the time, and the first thing I've taught myself is to not glorify it like some superhero saves the day worship. I'm both proud and (slightly) ashamed each time I handle an incident. It's not a virtue.

How many people who were writing blogs or creating open source tools before signed an on-call oath or declaration? Does that mean they're all wrong or unqualified?

On call avoidance should be rewarded more than on call superheroism. People managing software fail to realize it. Avoidance is stashing enough operational bullet points, mental checklists and manpages.

Everything is generic — because your use case and your operating environment is ONLY yours. Postgres has hundreds of parameters. Nothing is specific or made to satisfy you. If it doesn't, then you don't call the parameter as generic. You choose what to apply for your environment and expectation.

What you're saying is both a wrong reward cycle and it was also not necessary or practical before AI.

I have to disagree with this as off topic.

The blog lists ways to optimize sqlite for production, for those who have already made that choice. But these comment links are about choosing sqlite vs others and differences, which is completely unrelated and unnecessary for those who have already made that choice.

I'm seeing a worrying trend on HN. Nearly for all articles, there is one unquantified, unproven comment at the top saying it's 100% AI — no proof, just baseless emotion of what fits the commenter's writing style. This is the new witch hunt, or virtue trolling.

I read the original article and to be honest, this off topic comment definitely looks more like an AI troll bot wrote it than the article. The troll bot is instructed to do a google search and put 3 source links — no matter how irrelevant they are.

So what is it now — one emotion vs another? If an article has sections and blocks, it's AI. What about infinite poor humans like me who have been writing like this since childhood?

Since LLMs were trained on high quality blog posts, we can no longer have posts that fit the pre-AI definition of high quality? The content must be dumbed down with typis and scatter unconventional words, to prove AI trollers that it's not AI? Or a timelapse video of typing the article?

Instead how about: if you don't want to read the article because it doesn't fit your personal style expectation — just don't bother commenting?

I hope HN would ban AI trolling comments, and discuss substance.