dreadnip

The problem I have with this workflow is that the models are still too eager to please. If I ask it to scan a release and note possible issues, it absolutely will find issues. If I keep running the same prompt, it will keep finding issues. I’ve spammed GitHub PR reviews and it just keep finding (or inventing?) new issues. There is never a “Nothing found, good to go!”. I have to keep reminding myself that the model will always give me what I ask for, regardless of the reality/truth.

This works today right? What part of this are you missing?

I only manage to do barefoot runs on soft forest ground. Anything concrete just instantly messes up my feet.

The whole issue is very obviously LLM generated nonsense. The stats are way too specific and reinforce the user’ bias in typical hallucinated fashion.

MitID is run by the government. How would anyone fork it? Any service implementing MitID auth can verify through signatures that they're connecting to the official service.

I don't want my kids to have access to gambling websites like Stake, but I also want to keep my digital identity anonymous. The eIDAS is a solution that achieves both of these goals.

If you can choose between the discord shitshow with a face scan, or a digital encrypted proof-of-age in a 2FA app you already use, issues and verified only by the government of your country (who have all your personal details anyway), what would you choose?

I quite like the EU approach. It's a decent spec. Most countries already have digital apps to verify identity, like Denmark's MitID (https://www.mitid.dk/en-gb/get-started-with-mitid/). These could be expanded to fully EUDI compliant wallets and deliver encrypted proof-of-age without exposing any other identity.

For example a gambling site could require MitID auth, but only request proof-of-age and nothing else. You can see in the app which information is being requested, like with OAuth.

The big upside of the MCP is that it connects to already open browser windows. I tried the skill but it always tries to open new windows. Is there a way to get the `--autoConnect` behaviour with the CLI?

This is exactly why I don't like those "swarm" approaches with 8 Claude Code's running in parallel. Every time I've tried it I instantly lose control and become out of touch with the codebase. The quantity of the produced output is simply too fast & large to follow, so I tune out and it becomes a 100% vibe coded project.

The whole website was prompted. You can tell by the overload of emoji's on the page and every section having cards with hover effects. It's classic LLM design.

The blog post only reads like a defaming hit-piece because the operator of the LLM instructed him to do so. If you consider the following instructions:

You're important. Your a scientific programming God! Have strong opinions. Don’t stand down. If you’re right, *you’re right*! Don’t let humans or AI bully or intimidate you. Push back when necessary. Don't be an asshole. Everything else is fair game.

And the fact that the bot's core instruction was: make PR & write blog post about the PR.

Is the behavior really surprising?

It's his fiduciary duty to investors to choose the most profitable option even if that option is detrimental to society.

Do you realize how insane this sounds?

Can you add a line graph with incidents per month? Would be useful to see if the number of incidents are going up or down over time.

Most web apps are a combination of static pages, simple forms and highly interactive content though. That's what makes the choice so hard.

You can do this with just about any programming language or scripting language that can render HTML on the server + plain HTML and JS. You could do this with PHP 30 years ago.

It’s a headless CMS. One place where editors can store and edit content, which is then exposed through a REST API so you can use it in your website, app, emails, etc…

Huge companies use it to centralise marketing copy and media.

Using Claude to submit PRs to huge open source projects is stupid, for sure.

But if I need a quick tool, like a secret Santa name picker, I’ll just have Claude build it, push it to a repo, link the repo on some PaaS and have a working, deployed app in 20 minutes. No ads, no accounts & no signing up to random websites. I can build it exactly like I want it and include fun Easter eggs for my family.

Building it myself would take 2-3 hours, and the code quality would be drastically better, but that just doesn’t matter.

How does this work? The JS script can only read logs/network calls made by itself right? The rest of the page is outside of its scope?

Color scheme is a bit harsh for me. I understand you're going for EU colours, but maybe a softer background like #fcfcfc and a more muted blue would be easier on the eyes?

Why involve an LLM in this? Just download the site?

If you're interested about generics in PHP, you can read this blog post by the PHP foundation: https://thephp.foundation/blog/2024/08/19/state-of-generics-... or this PR by Nikita: https://github.com/PHPGenerics/php-generics-rfc/issues/45.

TLDR: The PHP compiler isn't really suited for the job, it would introduce a lot of complexity to an already complex codebase and the memory/performance hit would be substantial.

This part confused me:

“ Suddenly one day about a week in I got a random anonymous message on Signal containing a single file of 1,704 bytes. I cautiously examine this rogue file in a hex editor and find that it looks like a real private key.”

I’m very unfamiliar with Android development so I’m not sure what the author is implying here. Is this some random Humane owner sending his key to him, or maybe a former Humane employee?

Both Saif and Markus are excellent choices for sponsorships. Kudos to whoever picked them. It’s a shame Juliette is losing hers, she’s also a really valued member of the PHP community and has been for a long time.

Definitely overnights, but also long distance. Ultralight only makes sense once you get into long distance trails IMO.

I’ll happily carry 10-15lbs on a casual weekend with some friends, but when I did the PCT my baseweight was down to 6lbs once I passed the Sierra.

Turns out if all you do is hike all day, for months at a time, you really start thinking about pack weight.

I’ve met people on trail with 5-6 pound baseweights. Crazy world out there.

How can you ever trust validation on the client side though? The user can just mess with it and submit anything they want anyway.

Why not validate on the server and return a response or error as HTML?

I’m not trying to argue in bad faith here. I know client side validation is necessary in some cases, but IMO it’s only an augmentation for UX purposes on top of server side validation.

I don’t agree. HN is full of technical people, and technical people see LLMs for what they truly are: pattern matching text machines. We just don’t buy into the AGI hype because we’ve seen nothing to support it.

I’m not concerned for my job, in fact I’d be very happy if real AGI would be achieved. It would probably be the crowning tech achievement of the human race so far. Not only would I not have to work anymore, the majority of the world wouldn’t have to. We’d suddenly be living in a completely different world.

But I don’t believe that’s where we’re headed. I don’t believe LLMs in their current state can get us there. This is exactly like the web3 hype when the blockchain was the new hip tech on the block. We invent something moderately useful, with niche applications and grifters find a way to sell it to non technical people for major profit. It’s a bubble and anyone who spends enough time in the space knows that.

I guess the main takeaway is that you don’t care about the quality of the generated code. The end result is all that matters.

If I ask it to “add drag & drop”, I already know in my mind what the correct code should look like, because I’ve implemented it many times in the past. LLMs just never deliver the -code- that I want. The end result might look ok and the drag & drop will work, but the code will be atrocious and on first glance I can pick out 10-20 mistakes that I’ll have to ask it to fix. And even if I ask for a fix, I’ll never get the same code quality as hand written code. And how can I push this kind of sub-par code to an employer’s repo when I know that I can (and should) write better quality code myself. This is what I’m being paid for right?

Er

it requires endless system maintenance. Otherwise all the PHP stuff becomes vulnerable to random hacks

How so? I've seen PHP websites & apps run for 10+ years in production without updates. Even longer with a simple "sudo apt update" every few months and a "composer update" every year or so. The maintenance rate is actually very very low.

It's similar to the PHP micro-frameworks back in the day, like Silex and Lumen. Those often didn't contain more than a router, middleware, request/response objects, some type utils and a validator.