Grant Jordan

@sama Unpopular opinion: the government of humans. We are too irrational to govern ourselves. Ai should assist us. However, OpenAI models should never be allowed governance capabilities. They’re far too leftist. Please strive to make politically neutral models.

what problem do you most hope AI will solve in the future? maybe we can help!

@justinsunyt Terminal Bench 2.0 is easily gameable. Just two months ago ForgeCode was in the top spot, now it’s nowhere to be found in the list. Definitely don’t trust terminal bench

Codex is also not the best GPT-5.5 harness! Capy scores higher on TerminalBench, alongside 3 other harnesses People often ask us why we don't just build on top of other coding harnesses like Codex or Claude Code. The reason is simple: the harness makes a BIG difference for both performance and more importantly UX @capydotai we optimize our agents to excel not only at coding, but also behaviors like planning, user communications, and multi-agent orchestration, which happen to be very important for multiplayer/async interfaces like Slack and Linear We also do it for every frontier model so you can bring your Codex/Copilot subscriptions and enjoy a SOTA background agent with any combination of models + reasoning efforts without a harness "tax"

@Alberto8793 @XFreeze You’re a moron. Check the aliases for 4.3 dumbass

@XFreeze “Just launched” huh? Well, I’ll “just” leave this here.

Grok Build 0.1 just launched and is now available in the API It’s optimized specifically for fast developer workflows Pricing is extremely aggressive too: • $1 input / 1M tokens • $2 output / 1M tokens The output token speed is insanely fast

@jxnlco Creative work. GPT models are not creative. World class at logic based tasks though.

When do you reach for other models instead of Codex? What can we do better? Hit me with all of your frustrations. dms open. If you can give me detail (e.g. specifics/transcipts) - it'll help a lot in finding out exactly what we need to do to improve the next model

Very exciting! Reading the docs… can you guys please support .agents/ directories, in addition to .grok/ and .claude/ ? My sym link tree is bloated with all the independent dot directories, and I switch harnesses all the time. It would be great if all harnesses supported .agents/ ! Codex, opencode and others already support this.

An early beta of Grok Build, an agentic CLI for coding, building apps, and automating workflows is now available for SuperGrok Heavy subscribers. Through this early beta, we will improve the model and product based on your feedback. Try it at x.ai/cli

@steipete Dude you should really try oh-my-codex (omx). It already has all this built in and much more. Very powerful, I think you’ll love it

Wrote a skill that runs codex /review in a loop until there's no booboos anymore. Caveat: It won't fix system architecture for ya, so you still need BRAIN as master model. github.com/steipete/agent…

@ClaudeDevs You guys keep fucking up left and right. It's unbelievable. Y'all were the heroes at the beginning of the year.

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

@steipete Have you seen OMX ?

All I want is codex automatically entering /review mode after it's done and just looping until it stops finding booboos. (Yah I'm gonna build that)

@sama I would like them to be better at asking questions when the parameters are unclear, rather than just going off and doing something when they filled in the blanks on my context

what would you most like to see improve in our next model?

@realsigridjin What’s your take on $ralph vs /goal ?

If you only need to choose between them GPT 5.5 low reasoning + /goal vs GPT 5.5 xhigh reasoning w/ one shot what you gonna pick?

Love codex, but the CLI visual experience is annoying: - Collapsed tool inputs and outputs, hard to debug what the agent is doing and switching to transcript mode is awkward. I forked the repo just to be able to see agent tool inputs and outputs. - scrolling in tmux is so annoying - if I /resume a session, I don’t see previous tool calls when scrolling despite seeing reasoning summaries This is the latest release version. But overall really great cli tool . Well suited for autonomous agentic use

What are we obviously not getting right with Codex?

@sama @official_taches The pro $200 plan plus the $100 plan is all that would be needed if wrapped into one plan. I’m a heavy, heavy user (literally all day long) and a plan like this would get me through the week

@official_taches what do you need for just one plan?

@nikitabier @BillyM2k Yes, with an option to start a new chat with the post in context. I find myself tapping the Grok logo, then pausing inference immediately to ask the question I want. Or I copy the post share link and go to the Grok app. Slightly inconvenient

@BillyM2k What if it showed a tooltip when tapping it and gave you some options? • Fact check: Is this true? • Summarize this • Explain like I’m five

i find myself using the “ask grok” button on the upper right of posts very frequently these days, especially since grok 4.20 feels a lot smarter than previous iterations it’s really helpful when something looks like fake news and i can quickly check if it’s been vetted anywhere

@SartainPodcast 🤣🤣🤣

If you’re a hoe and you’re not at Coachella, you’re not serious about your craft

@theo Tailscale + tmux + moshi iOS app + wisper flow audio. Amazing workflow gains.

I want to control my agents from my phone. I'm sure you do too. What workflows do you want? Do you have one that works now? Tell me all about it.

@lexfridman You're a great man, Lex. Your mind inspires millions. Thank you for keeping it real and seeing The Human in People.

This life is fucking amazing. I'm so grateful to be alive, with all of you on this miracle of a planet. Oh and I'm sorry if I fuck things up sometimes. I'm a flawed human. But I promise to do whatever I can to try to add some more understanding and love to this world. After the world leader convos I get attacked intensely by all sides, and many disparate online communities. It has led to some really low points for me mentally. But I don't matter. I'm listening. I'll do better. And I'll try to find the strength to do more of them, always with rigor and backbone, seeking to truly understand. And despite accusations, I do extremely high amounts of research, sometimes 100+ hours for a conversation. Ask many of my previous guests. But when I come to the table, I put all that aside, and make it all about the other person. I don't ever try to sound smart. I know the vastness of my ignorance. But I'm trying. Sometimes I do fuck up and sound like a douche, or do something incredibly cringe. And I hate myself right after. But I'd rather fail and embarrass myself a million times, than not do what my heart says is right. And besides world leaders, historians, CEOs, engineers, etc, this year I want to travel the world and talk to a lot more everyday people on and off the mic. This is something I've wanted to do for a long time. Anyway this is written while on I'm on a 10 mile run, probably procrastinating, since to type I have to walk and not run 🤣 But I did just get stopped by a super smart and kind girl who works at a humanoid robotics company here. And she asked if she can give me a hug to thank me for being me. Sometimes the universe sends you a message that even a dumb dude like me can almost hear. I really needed that today. Thank you for the hug and the kindness 🙏 I'm just hoping she was real and I didn't just imagine that 🤣 Then again if I went full crazy might as well enjoy it! Back to the run. I love you all! ❤️

Since switching to GPT-5.4 full time I’ve never had a problem like this. People need to realize that there’s better tooling and models out there besides Claude Code + Opus/Sonnet. Be comfortable changing your tools often for better results and less frustration.

> “AI will make programmers obsolete” Meanwhile GitHub is absolutely drowning due to the 14x increase in code changes. Ai makes programmers MORE productive, not obsolete. I’ve personally never built more in my life than I am right now.

@dkundel Uh oh don’t tell Anthropic they’ll change their TOS

I built a new plugin! You can now trigger Codex from Claude Code! Use the Codex plugin for Claude Code to delegate tasks to Codex or have Codex review your changes using your ChatGPT subscription. Start by installing the plugin: github.com/openai/codex-p…

@theo @steipete Exactly my experience

Claude Code is kind of like if Codex was drunk. Fun, friendly, bit more creative, makes really dumb mistakes, probably shouldn't be trusted with prod.