VadimPR
VadimPR on Hacker News
I am on a paid plan for Google's AI offering, and I did just about ten prompts to Argon before I hit my 5 hour limit. I'm not holding my breath this will be a model that is practical to use now either.
AGENTS.md, which they are good enough at following. Of course it can be overridden but this stopped most folks from sending slop grenades.
We've solved this in our FOSS project by requiring all PRs to have been tested and manually signed off by a human. It significantly decreased the amount of spammy LLM made PRs, while empowering the core contributors to move faster.
I'm using Linux for exactly the opposite reason - it's less hassle than other OS's. And by that I mean every OS breaks, Linux is just simpler to fix when it does (and these days - Claude fix do it for me).
It's not mandatory to customise Linux and mess with it the moment you get your hands on it - you can also run a stock Ubuntu install which will be a great working & gaming workhorse that doesn't add hassle to your life.
Antigravity which lacks an auto approve mode? Not really comparable to Claude Code when you're looking to run a team of agents from my experience.
We still maintain phpbb forums, but activity on them is close to nil when compared to the Discord server. This is for a MUD client too, not exactly a mainstream audience!
The hard part, I heard, was closing a UK company!
They can't convince others, so I shopped around until I found a willing bank. This is due to introduced KYC requirements and the harsh penalties associated with them - so banks are preferring to err on the side of caution.
Same story goes for opening a personal account.
Before Europe gets lumped in as one country, founding a company in Netherlands and Sweden, speaking from personal experience, is a breeze.
Although Sweden is a bit strange in the fact that banks have as much equal say as the government authority does in you starting a company, and if they don't want you as a customer, they can simply deny the right for your company to start!
How can you practically use this in software if you're to deploy this to users? Asking a user to create and enter their own API key is a bar too high for good UX.
The release post is AI-written with little human oversight and it shows.
Claude is also surprisingly good at analyzing system issues on a Linux system and solving them!
The question is where do you go?
I find that fascinating as well! I hope tech will give us the answer in our lifetimes.
I had the same discussion with the color of a river in Albania with my wife. The test says my boundary is a bluer than 85% of the pop - sounds about right!
Appreciate the honesty from the team.
At the same time, personally I find prioritizing quality over quantity of output to be a better personal strategy. Ten partially buggy features really aren't as good as three quality ones.
I'm dissatisfied about the TAP leak as well! I was affected, and like you, didn't even receive a notification - nevermind compensation for having leaked my personal data to the dark web enabling all sorts of shenanigans that make my personal life difficult.
What would mathematics have to do with internal company politics, a soft/people skill demanding job?
Only on macOS though? This doesn't seem to work on Linux. Neither does Claude Cowork, not officially.
Can it work with comments in sheets as well? When I looked into it, that seemed like a limitation.
US should first implement a national identifier that can be used for healthcare purposes before implementing age verification, that would be a lot more helpful.
If someone is willing to put in the work in governance, FOSS projects would be willing to fund it - at least Mudlet would be. We get income from Patreon to cover the costs.
I'm in Europe and ended up creating an organization since I have my own company, but they messed up the verification of one of the legitimate documents, and there was no way to reach them once they made that mistake. Frustrating, and definitely a lost customer for them.
A year ago I used Azure Trusted Signing to codesign FOSS software that I distribute for Windows. It was the cheapest way to give away free software on that platform.
A couple of months ago I needed to renew the certificate because it expired, and I ran into the same issue as the author here - verification failed, and they refused to accept any documentation I would give them. Very frustrating experience, especially since there no human support available at all, for a product I was willing to pay and use!
We ended up getting our certificate sourced from https://signpath.org and have been grateful to them ever since.
I've been doing something similar by letting Claude run in a Virtualbox VM. It's easy to use, no issues with observability, and the attack & damage surface is far less of an issue.
I'm not one to believe the Silicon Valley hype usually (GPT-2 being too dangerous to release, AI giving us UBI, and so on), but having run Claude Opus 4.6 against my codebase (a MUD client) over the weekend, I can believe this assessment.
Opus alone did a good job of identifying security issues in my software, as it did with Firefox [1] and Linux [2]. A next-generation frontier model being able to find even more issues sounds believable.
That said, this is script kiddies vs sql injections all over again. Everyone will need to get their basic security up on the new level and it will become the new normal. And, given how intelligence agencies are sitting on a ton of zero-days already, this will actually help the general public by levelling out the playing field once again.
1 - https://www.anthropic.com/news/mozilla-firefox-security
2 - https://neuronad.com/ai-news/claude-code-unearthed-a-23-year...
Wouldn't this affect quality of output negatively?
Thanks to chain of thought, actually having the LLM be explicit in its output allows it to have more quality.