AussieWog93

I don't think anyone's implying that people writing software by hand, especially as a hobby, are wasting their time. Like not even if you attribute unreasonably bad faith can you extract that out.

We're living in an amazing time where an algorithm running on a GPU can create novel things by itself and it's just exciting to share what it can do.

I love reading about what people have managed to get running with Claude, because honestly it's amazing that these models can do any of the things they're now doing.

I also enjoy seeing what crazy, cool creative things people are doing.

Honestly I think part of it is just that it's cool to see software running on this platform. I remember coding some trash apps on the 3DS back in the day simply because I could. It was great seeing code running somewhere it isn't strictly "supposed to"

I know the CloudFlare CEO reads HN.

Every day I see so much cool shit from that org, that as a one man band I just can't trust because there's no billing cap/prepaid billing.

On the subject of learning something... Compare your comment to the sibling from Flashtoo. One is genuinely inquisitive and displaying willingness to learn, the other is just dunking all over someone.

Not a data centre expert, but don't they quite famously pump CO2 into the server rooms automatically if the fire alarms go off, even if there are people in the room?

Same thing happens IRL too. Not just literal scams but all kinds of ways people take advantage of one another.

Still doesn't change the arithmetic.

Honestly, I think our extremely liberal disability benefits scheme achieves this to a large extent (albeit possibly with downsides for both the recipient and society).

Look at all of the crazy stuff made by unemployed trans people/autists over the past decade, disproportionately from countries with strong welfare states.

Realistically, if Steam closed down tomorrow, within 4-6 hours there'd be a megathread on Reddit explaining how to get the games running without Steam. Within 48 hours there'd be an auto-crack tool floating around on GitHub.

I honestly never knew that, should give it another go.

The title's clickbait (from The Register, of all people - colour me shocked!).

There's nothing really groundbreaking at all in there, just "chips are expensive, and open weights models hosted locally in enterprise could displace Claude/GPT"

It might just be a me problem, but I've always been wary of regexes. They're not too bad to write, but reading them back and understanding what's actually going on can get a bit hairy. Plus, all of the subtle differences between regex libraries seems like a bit of a footgun.

Obviously they have their place, but I know a lot of the older guys seemed to love them way more than the young.

I think the main split between hardcore "GNU/Linux" guys and the general tech world is about principle vs pragmatism.

The hardcore crowd see free software as a right that needs to be fought for, whereas others see open source as nothing more than a way to modify or improve software without requiring explicit support from the publishers.

And there are many, many parts of Linux that make the software "freer" but worse.

LFSPA is a cap note, not an index fund. Hence the lower returns, haha.

Not saying it's necessarily the ideal vehicle but anything beats the banks.

LFSPA.

Not risk free, but not volatile either. Although even that underperforms compared to an index fund.

Funnily enough the two people we heard that "advice" from were Brits and Saffas. It's just bad interest yields compounding over decades.

The banks might pay 3% interest, whereas even non-volatile conservative investments like cap notes or bonds would pay something like 7%.

Plugging it into a calculator:

1.03 ^ 18 = 1.70

1.07 ^ 18 = 3.37

Example numbers, but you're effectively taking half of the money that your kid would have had on their 18th birthday, and giving it to a banker.

People in this thread are massively underestimating the level of financial illiteracy in the general population.

We've had multiple people try to convince us to set up bank accounts for our kids, so that they could accumulate interest over 18 years.

More that tried to convince me to gamble on random pump and dump shitcoins.

More still that talked about "investing" in random collectables like Funko Pops or Pokemon cards - they're not a bubble, Logan Paul told me so!

You could replace the AI with a piece of paper that says "set aside 10% of your income and invest it in an ETF" and it would outperform the financial "advice" that people receive on a daily basis.

I'm in the same boat as gp.

They seem to be doing their darndest to kill advantage #1. Even the command line "cursor ~/git" no longer opens an IDE in the ~/git folder but some random Chatgpt-esque interface that nobody asked for. Honestly, VS Code is a better "old Cursor" now than Cursor is.

#2 hasn't been much of an issue for me since Opus 4.5 came out. Anything the other models could do, Claude could do better. Of course things are starting to change now, and I've noticed Sol Ultra is specifically better at visual design stuff than Claude, but even that ~3% of cases I switch to GPT are still more easily solved with a cheap Codex sub than Cursor.

Look into the /stage-chapters skill for local review. It groups large edits logically into "chapters" (think individual Claude plan steps) that you can review one at a time, and it's local.

Love the username, very relevant!

It's only a DNS block fwiw. If you set your DNS to 1.1.1.1 or 8.8.8.8 everything works fine.

Honestly, there's a 50-50 chance that you'd be able to do this in Claude today, even without access to the source code.

Just open up a configuration window manually, tell if that the Window is open and it can pull apart the executable from there.

Biggest impediment would be Fable safeguards if it tries to decompile.

Honestly, things have really changed in the past 6 months. Opus 4.5 got most of the way there, and Fable can now literally one-shot your run off the mill CRUD app.

I use a few of them daily. While there are still bugs, there are significantly less than there used to be back when everything was being maintained in my spare time.

Still not Netflix/Google standard yet, but definitely better than the average side project or pre-AI Bangalore special.

There can be many correct ways to use a tool, while still having incorrect ways to use it.

Sometimes you need to set your oven to grill, other times fan force. But if you're complaining about getting burned then, yes, you're simply using it wrong.

Apparently Kimi K3 has 104B parameters active at a time. So at 4 bits you'd need 52GB just to hold the active params.

That said, in theory this same technique should be able to run it on a 64GB Macbook, probably at <1 tps.

Directly customer facing is a tough one, since a lot of the time chatbots are used to just fob off customers that eat up a lot of time (old people complaining etc).

I think the best experience I've had is when companies offer an MCP server that lets me get up and going with their product very quickly. Sequenzy is one such example I tried, very AI-first and works well with Codex.

If you're talking about stuff that isn't customer facing, there's all kinds of tasks that can and have been automated.

As a simple example, I run a retro games business and we've vibecoded up various controller/cartridge testing apps that have dropped the time it takes to test a box of product by around 50%. I'm planning to spin up a few Hermes agents soon to watch the prices of things and suggest improvements to our listings (e.g. noticing we haven't soldas many copies of Super Mario 64 as we should, and checking our keywords/pricing).

Obviously puts in big work with SaaS and stuff too.

EDIT: actually, Shopify's built-in agent works really well and is directly customer facing. It's a smart model (I think either a Kimi or Qwen as a base) that seems to have been fine tuned on their product itself.

You can ask it stuff like "what products have seen a decrease in sales" or "what is the value of my next payout" and it just answers them.

Before you had to manually write DB queries for this stuff, or more realistically just ignore it.

I am wary of stories like this, since they seem to appeal to what people want to be true rather than what is necessarily happening.

There are a lot of businesses quietly and conservatively implementing AI and seeing real results - it's just the crazy Jerry Springer blowups you hear about!

I could imagine a machine like this would cost something on the order of $100k, require ongoing maintenance and wouldn't really solve any problems/bottlenecks at your typical local pharmacy.

It's not like you could replace the pharmacist with a machine.

Not sure about LinkedIn, but Facebook does let you export all your data as a bit monolithic zip file.

I think everyone has weird ideas like this, but these rich, well-connected guys don't have to worry about their lives being ruined when they actually share them out loud.

I can't imagine many aerospace manufacturers turning down a contract with SpaceX because Elon said something politically incorrect. James Damore, on the other hand...

This isn't hate towards these rich guys (I think everyone should be able to say shit like this), just pointing out that we as a society hold them to different standards than we do regular people.