You keep mentioning that in this thread, but a11y on a Mac is considered the gold standard.

Security on a Mac, the same (SIP, Keychain, Secure Enclave, great tools for fleet management)

What specifically is in violation of "#a11y or security evaluation frameworks"?

#a11y in Mac used to be a gold standard. And FedPack in the 2000s made MacOS-X a good alternative to the confusing jumble tha was windows security. This is not the case in 2025.

Source: "trust me bro"

2.5 flash and 2.5 Pro were just sitting back and watching?

The problem with Google is that someone had to show them how to make a product out of the thing, which Open AI did.

Then Anthropic taught them to make a more specific product out of there models

In every aspect, they're just playing catch up, and playing me too.

Models are only part of the solution

I have "unlimited" access to both Gemini 2.5 Pro and Claude 4.5 Sonnet through work.

From my experience, both are capable and can solve nearly all the same complex programming requests, but time and time again Gemini spits out reams and reams of code so over engineered, that totally works, but I would never want to have to interact with.

When looking at the code, you can't tell why it looks "gross", but then you ask Claude to do the same task in the same repo (I use Cline, it's just a dropdown change) and the code also works, but there's a lot less of it and it has a more "elegant" feeling to it.

I know that isn't easy to capture in benchmarks, but I hope Gemini 3.0 has improved in this regard

I have the same experience with Gemini, that it’s incredibly accurate but puts in defensive code and error handling to a fault. It’s pretty easy to just tell it “go easy on the defensive code” / “give me the punchy version” and it cleans it up

Yes the defensive code is something that most models seem to struggle with - even Claude 4.5 Sonnet, even after explicitly prompting it not to - still adds pointless null checks and fallbacks in scripting languages where that something being null won't have any problems apart from an error being logged. I get this particularly when writing Angelscript for Unreal. This isn't surprising since as a niche language there's a lack of training data and the syntax is very similar to Unreal C++, which does crash to desktop when accessing a null reference.

    but I would never want to have to interact with

That is its job security ;)

I can relate to this, it's doing exactly what I want, but it ain't pretty.

It's fine though if you take the time to learn what it's doing and write a nicer version of it yourself

I have had a similar experience vibe coding with Copilot (ChatGPT) in VSCode, against the Gemini API. I wanted to create a dad joke generator and then have it also create a comic styled 4 cel interpretation of the joke. Simple, right? I was able to easily get it to create the joke, but it repeatedly failed on the API call for the image generation. What started as perhaps 100 lines of total code in two files ended up being about 1500 LOC with an enormous built-in self-testing mechanism ... and it still didn't work.

Hmm... it basically says that real power in the US isn’t with the government anymore, but with giant tech companies and the rich weirdos behind them.

These rich weirdos run the show through apps, platforms, and algorithms that decide what people see, do, and believe.

It’s not that democracy broke, it just quietly got replaced by a system where the rules are written by whoever owns the server.

Me: I built this insane 3D multilingual AI avatar platform with real-time conversation, full enterprise stack, 47 languages, and Unreal Engine

You: Can bypass US Treasury sanctions and open a Stripe account

Ha, fair.

Yeah, the skills mismatch is real. I can build distributed systems with 4-server architecture, but I can't open a Stripe account.

That's literally the post. "I'm good at X, terrible at Y (geography), need someone who has Y."

If you know anyone who's great at sales but wishes they had a technical co-founder, send them my way :)

It's a very unique cofounder pitch! I hope you bootstrap this idea as far as you can.

You should encourage any potential founder to talk with a lawyer, though, too!

Are you even allowed to have github or Gmail accounts? From my understanding those are US companies and not allowed to provide any services to Iranian nationals.

It does pattern match to the Spanish Prisoner scam, with extra FinCEN violations thrown in.

He does have a LinkedIn, and his GH user has a real looking GitHub pages site associated with it.

Please don't violate US Treasury dep sanctions just to be a founder for a business with 0 customer validation, you might as well be helping a Nigerian prince.

Seek counsel with a lawyer versed on this first, rather than the HN comments section or your own exuberance.

An extra reason for the victim not to go to the cops!

A better analogy would be the "Trade Offer" meme where someone offers a share of the profits, but your offer is that you are willing to violate serious international sanctions.

Honestly I would, but I’m already helping this Nigerian prince.

This post reads less like an engineering deep dive and more like a Xeon product brochure that wandered into a video game blog. They casually name-drop every Intel optimization short of tattooing "Hyperthreaded" on their foreheads.

well of course they would. they bought all intel hardware. and they are making one of the most perfromant multiplayer servers ever. they should be mentioning every optimization possible. if they had amd thread ripper servers they would mention all those features too.

I think one thing they need to focus on is making extensions easier to create.

I'm a big fan of Intelli-J, but VS Code is eating their lunch with AI agent plugins. I'm assuming it's because it's much easier to create them on that platform.

The same pluginm from the same team, on Intelli-J always works much slower, feels clunkier and is prone to crashing.

It's a shame

Ah, I'm glad I'm not missing out by using Visual Studio. It's probably even slower in VS, I'm sure. I've been dual wielding VS for building/running and VSCode/Cursor/claude-code/codex-cli for AI editing, and it's a much better experience. Although if it wasn't so shitty in VS, I would do it all in the same UI.

EDIT: I'm especially pissed at Microsoft for not improving VS's rendering speed. They are all about making its backend services async, spinned out to external processes to be able to use the latest dotnet version, while the main IDE process is stuck in NET 48 seemingly forever, with zero improvements made to rendering performance. Perhaps the Vello Sharp news I'm hearing recently will kick MS into gears (one can hope).

Which plugins have you compared? I think plugin devs of this generation are only target vscode clones because if it gives them coverage of vibe coders who are happy to pay even for dubious quality.

> about US$2.5 billion on stock-based compensation, nearly double the amount from the first half of last year.

Wow! 2.5B in stock based compensation

Wouldn't it stand to reason that they would provide a sort of utility for a collegue as they read through the generated code.

Comments are often the best tool for explaining why a bit of code is formulated how it is, or explaining why a more obvious alternate implementation is a dead end.

An example of this: assume you live in a world where the formula for the circumference of a circle has not been derived. You end up deriving the formula yourself and write a function which returns 2piradius. This is as simple as it gets, not hacky at all, and you would /definitely/ want to include a comment explaining how you arrived at your weird and arbitrary-looking "3.1415" constant.

This is awesome! Downloaded it and played with it and as a non Kagi user, I'm impressed!

But please please let me select text!

I wanted to look up a word, so I double taped it to select it, then I realized non of the text in the articles is selectable