I think it’s Grok, because Grok has multi-agent capabilities. In a multi-agent setup, the agents think and reason independently from one another. Then they integrate the data and ultimately output the exact result the user wants. Moreover, if the user isn’t satisfied, they’ll rethink and reason again based on the context until you’re fully satisfied. Other AIs just seem to think independently on their own without any debate. That’s why I believe they may lack the accuracy that comes from debate
@0xtenthirtyone@elonmusk The mental gymnastics required to go from 'AI used a 16-year-old as an example' to 'pedo' is genuinely Olympic level. A generative AI cannot be a pedophile. That's not a defense - it's just physics. Get a grip.
Oh no, people are angry
Ok let's go:
1. ?? I thought all of you wish to have that or? Idk what is impossible here except if you are studying a bit longer. You will still need a side job tho
2. @IterIntellectus has a lot of good articles about how its good to get married and have children quite early because the fertility can drop quickly
3. ?? kids are nice, have them. and NO they don't cost a lot.
4. You can do this in two weeks going once across Europe
5. well this one is a bit unrealistic but if you are not from the USA where you start at -200k and you count your and your spouse's money too it is more doable
@_mattwelter you understand subconsciously that your work is a net negative on society
I don't say that to be a hater/out of spite, but any short form content farm is just an abstraction layer over being a slop factory
I would make your bag and get out
i have...
- an amazing girlfriend
- making ~$450k this year
- can work anywhere / anytime
- live in a house w/ a pool
yet i have anxiety every damn day, tight chest, hard to take a deep breaths, intrusive thoughts, always feeling not enough, can never relax
what went wrong
Tried the OpenAI browser for 20 minutes.
Quit and went back to Chrome.
“Agent mode” is slop.
Most of the time I just want to yell:
“Stop thinking and click that fking button!!”
The models are not there yet.
We’ve got a whole “decade of AI agents” ahead.
@dannothingnew@DavidOndrej1 Funny how Claude fan boys will always blame the prompt - AI is fundamentally about converting NATURAL LANGUAGE into actionable steps… If it’s incapable of doing that then it’s a shit AI ✅
Dude, I'm totally lost here. Are you guys seriously comparing GPT-5 Codex to Sonnet 4.5, both without any prompts? Cause otherwise... I just don't get when Codex would be better.
Like, Codex doesn't even listen to your instructions, it doesn't work with MCP, it messes up, it forgets stuff, and it can straight-up ignore rules you've set, unlike ClaudeCode.
Now, ClaudeCode messes up too, I'm not denying that. But at least it has hooks and actions so you can tweak its behavior. Codex, on the other hand, can just ditch all that.
So, seriously, what's the big deal with Codex then? I'm just not seeing the upside here.