AussieWog93

Man, I got ready to debunk this but just questioned my life choices instead.

Why the fuck am I glued to my smartphone arguing with people posting shit like this instead of spending time with my daughters, creating new things or looking after my own body and soul.

Michael, please copy your comment verbatim into any LLM with search capabilities and ask for primary sources that confirm or deny your statements, and break down the beliefs by party.

Then, use your real flesh and meat brain to ponder where your (not other peoples' - your own) beliefs about this came from, and what their motivations for blasting them at you could be.

Goodbye HN, I think I'm done with this site for good.

I've had negative responses to small, isolated submissions that I've heavily QA'd and semi-positive responses to longer ones. I think it has less to do with the "realpolitik" of the code itself and more to do with the maintainer's viewpoints about whether AI as a whole is a positive or negative thing.

I tend to respond quite well to AI-authored or assisted PRs to my project, but to be fair we maybe only get 3-5 PRs in a good month.

I'm going to say it - I think you're just spending a lot of time around very negative people, possibly in a social media bubble.

A lot of the stuff you mentioned isn't really AI (the risk of job loss from which I agree is real), but just everyday stuff that humanity has shrugged off since the dawn of time.

I installed Ubuntu on an old laptop that I let Claude Code sysadmin. It makes it really easy to self-host open source stuff, and if there are issues it can fix them too.

Funnily enough, I've had a somewhat mixed-to-hostile response when trying to upstream the vibecoded fixes, so I suspect using an LLM to fix broken open-source software (that human maintainers don't have the time to fix themselves, nor the humility to accept an LLM-authored fix) will become more of a thing going forward too.

Oh, it's also been identifying a bunch of patterns in sales data for my business that has been increasing monthly profit consistently since last November (around $4,000 USD, every month, cumulatively so far with no sign of slowing down - could easily be $10k/mo in increased gains by end of financial year).

It's funny you say that, I do see where you're coming from.

I don't personally see cost as an issue, since I'm using it for a small business with a subsidised consumer subscription (subscription cost is negligible compared to efficiency returns from the last 6 months), but I have found myself going to GPT over Claude for specific niche tasks - namely visual web design and SVG graphics. It does a good job of more things than just pelicans - its N64 controllers are good too!

I run a few small businesses and quite regularly get unsolicited emails/calls.

There's absolutely a real, qualitative difference between contact from a real human being who respects everyone's time and has done actual research, and someone who is just playing the numbers game and blasting as many people as they can (regardless of how "warm" the leads are).

It's a bit of a spectrum, but if someone has taken the time to handwrite and research an email or call, I'll at least try to give them the courtesy of a response.

I've had the opposite experience giving claude code SSH/ADB access to my devices. Not sure if it's the model itself or the harness, but I haven't had to manually do a sysadmin task in months.

I was about to "correct" you but then read the username - has the best model for agentic coding/sysadmin not just been Claude every single day since Opus 4.5 released last November?

Godwin's new law!

Usually I will ask the LLM to commit only the work it just did, in case the worktree is dirty. It also tends to write actual in-depth commit messages too.

I think literally anyone would stop reading these randomly long long bash commands at all after prompting Claude to do some basic sysadmin task and getting 15 of them in a row.

For installation, maybe. sometimes.

(The biggest issue I'm aware of is that if you have an old version of the app indexed anywhere on your root drive - for example in ~/qa/app-v1.02-DO-NOT-DELETE - it will literally stomp over that instead of installing the app to /Applications, even if the developer manually specified the install location as /Applications in the manifest.)

But for uninstallation, not a chance. There's no Mac equivalent to Add or Remove Programs on Windows - once you install a PKG good luck uninstalling it.

I used to really enjoy this aspect of older games like Warcraft 3 and TF2.

You'd get the one guy on your server who was just in another league compared to everyone else, and it was so satisfying to get the kill on him. And sometimes, it would be you pubstomping.

Then again, back then I don't think people cared as much about winning and losing.

I mean, let's be real for a second. It's cybersecurity.

Yes, it does skew male, but you're probably looking at 10x the standard population density of Autistics, and 20x the density of trans women.

I'm sure there are other "sociologically valid" ways you can measure this too, but there is absolutely a correlation between having the type of brain that is able to perform offensive cybersecurity operations and mental illness/social rejection/isolation.

> It’s very clear that vast majority of the business do not generate additional value from incremental intelligence gain from these models.

It's been 2 months since Fable was released to the general, man. Nobody knows what's going on inside of these companies except the people at the coal face.

Not really, no. My point is that a $100/mo LLM subscription is generating $5000/mo in value, so even a negligible difference in performance wipes out the cost savings completely.

Same reason it makes sense to assign a team of humans that cost $100k/mo to a product that brings in $5M/mo, rather than one human with 5 Claude Max subs.

The cost is a rounding error.

We are obviously limiting access to be read-only for anything customer-facing (it will be able to put recommendations in the dashboard but not actually change things directly) but I think the point of GP's comment was literally just to get a reaction.

Yeah, fair. If we were talking $2,000/mo vs $200 then the maths starts looking very different.

To put actual numbers on it, since using AI to start solving all kinds of bottlenecks/inefficiencies in our small business, we've seen monthly net profit go up by around $4,000 USD. These are semi-permanent fixes, and the tech is only partially deployed. I am the only one using it, and I only use it part time.

We've just spun up our first Hermes agent, with direct API access to our main inventory system and that's expected to find another few grand per month in misallocation/inefficiency.

I wouldn't be surprised if we were doing more like $10k/mo higher in 6-9 months' time.

When you're talking about numbers like this, the fact that one AI is $100/mo and another is $10/mo or $40/mo doesn't matter. They could make GLM-5.2, or any other Opus 4.5-class model free and it still wouldn't make sense to deploy in a commercial context.

The other angle I'd approach things from is that Opus 4.5 (and I'd agree with you that that model was the saddle point) was "good enough" for the types of things we were asking it to do back then, but as the models have become more capable the tasks we're asking them to do have also expanded with it.

I know I've personally gone from "hey can fix this race condition with a Redis mutex" 6 months ago to "Independently redesign this full embedded USB stack and QA it end-to-end, working around a specific Kernel bug in macOS Tahoe that requires decompilation to find the source of, while keeping in mind the constraints of our 8-bit AVR chip from 2011" now.

But that said, yes, maybe in 5 years' time we will reach an "intelligence saturation" where the average person won't be able to even conceive of how to use the new SOTA.

I think we put up with Fable's occasional hiccups because there's nothing better at the moment.

I use Claude Code semi-heavily for my small business, and the $100/mo I pay for that is a rounding error compared to the value it provides.

If I can avoid spending an hour or two "massaging" the output from a lower-end model once, or it avoids introducing one load-bearing (sorry, couldn't resist) bug, then that's the entire $100 right there.

Hell, you could argue that the best "coding model" that we have at the moment is the human brain, and people will gladly pay $10,000/mo for one of them.

Arguing over $20 vs $100 for something that actually puts in work just seems insane to me.

I think the implicit assumption is that if you explained to the owner of the website what the scraper was doing, they would be more than happy to allow it.

The bot was also being generally courteous, only scraping the 2-3 pages needed to get the information for the human who spun up the LLM.

I don't think GP was at all suggesting he should be able to act however he wants, just because he's a paying customer.

I run an ecommerce site and I'm actually more worried about bots being rejected than accepted because of stuff like GP. We get a decent amount of business through ChatGPT.

This is a sensible choice.

Their newer models are heavily trained out of the "whoopsie doopsie I deleted prod, and all backups too" behaviour from last year.

I've been running it for months in auto mode, heavily, getting it to do sysadmin tasks via SSH across multiple servers for both myself and a client, and there's not even been a whiff of anything dumb/catastrophic -quite the opposite, in fact.

I'd even say it's more careful than a lot of humans. It's extremely anal about standard "hygiene" stuff like not leaving plaintext secrets lying around, and creating post-deploy scripts to confirm that every file/dir is created with the correct permissions.

There was one time I carelessly suggested uploading (my own) private data to a random public endpoint when testing OCR options and the model actually stopped, explained the risks and refused to continue until I confirmed I understood. I decided not to.

I'm not saying it's perfect, and I'm sure HN being HN there'll be someone who responds with an example of their agent doing something dumb/dangerous (give dates/models/context if so, I'm curious!), but I think on balance it's currently more sensible, and more cybersecurity-minded than the bottom 80% of IT professionals.

This is all true for Claude, I don't know much about Codex but it seems a lot less heavily trained for this kind of stuff.

Honestly, this is a problem that has been solved by LLMs. Just point Claude Code to the server over SSH and say "We need Immich to be served over HTTPS".

You could theoretically upstream the fixed compose.yaml if you felt like it, but the project seemed unusually hostile towards the one fix I tried to upstream.

>I am developing a mobile app, competing with companies with millions in revenue and entire dev teams. I know it is viable. Others lag in effective adoption, their opinion is likely to change soon with even more capable models.

To add to that, as an ex-software dev that mostly works in a semi-unrelated field now (and can only code as a small part of my job): I think a lot of software devs are underestimating what someone with reasonable technical skills and specialised domain knowledge is able to create with AI.

I honestly wouldn't be surprised if, in 5 years' time, the majority of software used by (e.g.) potato farmers was primarily created by other potato farmers. The code might still be less elegant but I think it will be easier for the potato farmer to iterate with an AI than outsourcing to a dev firm.

I don't see the AI bubble popping any time soon (as in, not in terms of physical datacenter buildout - stock prices could fall). Even if progress hits a brick wall, we still have multiple years of just expanding out the models we have to more people in more industries.

Could we not just make bigger wafers, if the technology called for it?

I don't know whether you're looking at small-to-medium batch indie stuff, or big customers, but if it's the former I'd reach out to Josh Lifton from Crowd Supply. Surprisingly accessible guy and this would be a great fit for them.

I'd say it's more "Downloading LimeWire Pro from LimeWire" than actual theft.

I mean, if you could get Opus or even Sonnet 4.5 at 1000+ tok/s exploring the codebase, they would probably change that setting back.

But either way, I think GP's overall sentiment of "delegating intelligence-saturated tasks to an outdated but fast subagent" makes a lot of sense.

I read the paste, it got the etymology wrong, no? Schlong comes from shlang (snake), not shlemp (is this even a word? I don't speak Yiddish but couldn't find it on Google).

Oxford also claim that its first recorded use was from the 60s, not the 20s; https://www.oed.com/dictionary/schlong_n?tl=true