Good.

Yeah, I'll be on Sequoia until it's unbearable (probably 4 more years), and then I'll either put Linux on that machine or I'll just buy a non-Mac. Been using Macs since Snow Leopard but between ios 26 or whatever it is and this shit, I'm done.

I mean, I'm the exact opposite. Ask ChatGPT to write a simple (but novel) script for AutoHotKey, for example, and it can't do it. Gemini can do it perfectly on the first try.

ChatGPT has been atrocious for me over the past year, as in its actual performance has deteriorated. Gemini has improved with time. As for the comment about lacking wit, I mean, sure I guess, but I use AI to either help me write code to save me time or to give me information - I expect wit out of actual humans. That shit just annoys me with AI, and neither ChatGPT nor Gemini bots are good at not being obnoxious with metaphors and floral speech.

Sounds like you are using ChatGPT to spit out a script in the chat? - if so, you should give 5.2 codex or Claude Code with Opus 4.5 a try... it's night and day.

> 5.2 codex or Claude Code with Opus 4.5 a try

Is using these same models but with GitHub Copilot or Replit equally capable as / comparable to using the respective first-party CLIs?

I don’t think so. My favorite tool is Codex with the 5.2-codex model. I use Github Copilot and Codex at work and Codex and Cursor at home. Codex is better for harder and bigger tasks. I’ll use Copilot or Cursor for small easy things. I think Codex is better than Claude Code as well.

Are you using the same models and thinking levels for each?

I too have found Codex better than Copilot, even for simple tasks. But I don't have the same models available since my work limits the models in copilot to the stupid ones.

I have GH Copilot from work and a personal Claude Code max subscription and have noticed a difference in quality if I feed the same input prompts/requirements/spec/rules.md to Claude Code cli and GH Copilot, both using Opus 4.5, where Claude Code CLI gives better results.

Maybe there's more going on at inference time with Claude Code cli?

It is likely because GH Copilot aggressively (over-)manages context and token spend. Probably to hit their desired margins on their plans. But it actively cripples the tool for more complex work IMO. I've had many times where context was obviously being aggressively compacted and also where it will straight truncate data it reads once it reaches some limit.

I do think it is not as bad as it was 4-6 months ago. Still not as good as CC for agentic workflows.

I find this really frustrating and confusing about all of the coding models. These models are all ostensibly similar in their underpinnings and their basic methods of operation, right?

So, why does it feel all so fragile and like a gacha game?

OpenAI actually have different models in the cli (e.g. gpt-5.2-codex)

Naming things is hard. So hard every AI company isn't even trying to come up with good names.

In this case they probably are prompting it "wrong" or at least less well than codex/copilot/claude code/etc. That's not a criticism of the user, it's an indication of the fact that people have put a lot of work into the special case of using these particular tools and making sure they are prompted well with context etc whereas when you just type something into chat you would need to replicate that effort yourself in your own prompt.

Nobody said you have to get all your protein from meat…

> Might be illegal to break someone else's proprietary software, no matter how abandoned it is.

Oh no!

Anyway

Snark aside, what you do locally on your own PC in your own home is kinda nobody else’s business, especially when you aren’t cracking it to share it with others. Pretty sure all the arguments were already hashed out in the ‘80s when VCR companies tried to block recording television to cassette tapes.

Literally nobody cares.

Well, didn't think of that, but agree completely.

In order to sue the post author, the software copyright holder must claim some damages. I doubt they will find enough damages to pay legal fees.

> Pretty sure all the arguments were already hashed out in the ‘80s when VCR companies tried to block recording television to cassette tapes.

Why would VCR companies try to block this?

Not the VCR companies but the TV stations and Cable companies.

Maybe not for crypto, but they did spend a year or two surreptitiously installing their own VPN service on your Widows machine without any opt-out ability, and then failed to remove it, its Windows service, or multiple scheduled tasks once the brave uninstaller had been run.

The best part was this whole scam sitting as an unresolved issue on GitHub for months after they finally acknowledged it (after first denying it lol).

Closest browser I’ve seen to an actual virus in maybe ever.

And it’s a good lesson for developers that once you lose trust there are many of us who will never make the same mistake again purely out principle.

Kagi's Orion on iOS is good, and supports ff/chrome extensions

Kind of, but not really, the ublock origin app doesn't work on Orion. I tried both Kagi + uBlock versus Brave and their built-in ad-block and Brave had a much better performance.

True, none of the extensions I tried actually worked as intended.

Fwiw Orion uses the actual extensions, not the uBO Lite app (for Safari) that I think you're referring to.

Brave also installed a VPN an a VPN service without permission on my Windows machine, and then didn’t disable or remove 3 separate scheduled tasks in Windows Scheduler once I’d uninstalled it. The VPN issue was open for like 8+ months on GitHub too - and at first they denied doing it at all. For all I know it still installs it, but I removed this malware-type shit when this all happened so I couldn’t tell you.

I’ll never trust them again after that.

You've commented this three (?) times under this HN post and several times on this site now. Sure seems like you have a bone to pick.

The VPN they installed was disabled and they could not activate it without user interaction. And the only reason they did this is so when you click "activate VPN" in the browser, it works immediately.

On top of that, other businesses employ(ed) similar tricks. For years and years and years, Dropbox on macOS did a very specific hack to give itself more permissions to ease syncing. Hell, Firefox injected ads for Mr. Robot via a surreptitiously installed invisible extension.

Still a boneheaded move by Brave, just like adding their own affiliate link to crypto links (if none were added) to generate extra revenue for the company at no extra cost to the user. But that is even further in the past.

At any rate, they also fund or develop a bunch of anti-ad tech and research and make it open source / publish it. The defaults of Brave protect your privacy much better than Firefox's defaults. And so far, their BAT concept is the only one that is a legitimate alternative to an ad-funded internet.

Brave is everything Mozilla wishes it had become.

Minimising a browser installer surreptitiously installing an unrelated network service without my consent in the name of convenience and then pulling out a whataboutism that has nothing to do with this. Not even gonna bother reading the rest of your comment.

And yeah, my bone to pick is warning others not to fall for Brave’s slick PR. Companies that act that way can pay the price.

LLMs are dogshit in many ways but when it comes to programming they are faster than people, respond instantaneously to further information, and can iterate until they understand the problem fully.

Bonus is that you don’t get some dipshit being snarky.

You clearly don’t play competitive shooters and thus aren’t qualified to opine on the matter.

Competition vs other human beings is the entire point of that genre, and the intensity when you’re in the top .1% of the playerbase in Overwatch/Valorant/CSGO is really unmatched.