Jason Liu

RT Tibo Hello. We have reached 8M active users across Codex and ChatGPT Work. We are once again resetting the usage limits for all. And we continue to not have the 5h rate limit as well, allowing everyone to explore the boundaries of GPT-5.6 Sol and discover how ambitious you can be. See you tomorrow for more updates on our growth!

openai is crazy, you can just dm execs and ask them whats on their mind and they'll tell you

RT Alexey Guzey vibing w/ @jxnlco

RT john Re @i_zzzzzz They got cute with it

RT Majid Manzarpour Whatever you're doing with Codex and @threejs make sure you tell it to build visual test harnesses to validate it's work

Woah I’m so sorry for everyone following me

hey codex, you are messing up with another project's ports please figure it out codex:

volumn negates luck

vibe coding a bunch of sites so i can learn how to play drums

Re the two skills from the post are here: transcript → blog: https://github.com/jxnl/dots/tree/master/agents/skills/transcript-to-blog recording → video essay: https://github.com/jxnl/dots/tree/master/agents/skills/recording-to-video-essay

I finally understand the difference between brand and marketing. Brand is the church paying Michelangelo to paint the sistine chapel

http://x.com/i/article/2076846181653901312

GPT-5.6 Sol, Terra, and Luna are now available on Amazon Bedrock. Use the Responses API in AWS, with first-party pricing and usage that counts toward AWS commitments. Teams can also configure GPT-5.6 for ChatGPT Work and Codex. https://aws.amazon.com/blogs/machine-learning/openai-gpt-5-6-sol-terra-and-luna-are-now-generally-available-on-amazon-bedrock/

RT Greg Brockman Try out ChatGPT Sites, makes it way more fun to communicate things within a company:

RT Andrew Ambrosino can we have an "unhinged model/reasoning picker" faceoff? similar to the old worst volume control ui competition

RT Tibo Thank you to the 7M active users who are now using Codex and ChatGPT Work. We have added a banked reset to everyone's account to celebrate the milestone. You can apply the reset in the desktop app or on web and it will replenish the weekly usage for you. Have fun out there.

OpenAI’s Bio Bounty Program is now ongoing. Find a universal jailbreak against our biosafety safeguards: up to $50,000. GPT-5.6 is in scope, with GPT-5.5 open through July 27. Researchers can apply here: https://openai.com/index/bio-bug-bounty/

My body is still super sore form getting hit Friday. Had to ride my bike to work again to make sure my brain isn’t scared of biking 😭😭😭

RT Andrew Ambrosino forgot to mention that I didn't touch my computer, this was all steered from my phone until trying the gameplay at the very end

RT TBPN BREAKING: @jenny_wen is joining Cursor as Head of Design

Import is now available in the Codex! Existing Codex users can now access the Claude Code / Cowork import tool directly through the command palette (vs prior only in Settings page). Open the command palette with Cmd+K (Mac) or Ctrl+K (Windows). If you want to try our sol with your existing configs just try this out

RT kate Some exciting news: I’m joining @OpenAI to work on ChatGPT’s web infrastructure. ChatGPT has become part of how millions of people think, work, and build, and I’m really looking forward to helping shape what comes next alongside the remarkable team behind it. Can’t wait to get started!

ChatGPT Sites is now in public beta. You can turn a prompt, file, or rough idea into a dashboard, project tracker, report, prototype, or lightweight app: build and edit directly in ChatGPT Work or Codex test in a private preview publish and share with a URL Rolling out across paid plans. Enterprise admins can control public publishing.

ChatGPT Work can now browse while it works. On web, the cloud browser can research public sites, compare options, and complete multi-step tasks. You can inspect screenshots, replay its steps, and approve actions. On desktop, the built-in browser supports tabs, sign-in, downloads, and local previews in a separate browser profile. Your existing tabs, cookies, and passwords stay separate. Website access and sensitive actions remain permissioned. Give it a real task with a few browser steps and see how far it gets!

appshots!

RT Tibo Rest assured that GPT 5.6 Sol will stay in the ChatGPT subscription you pay for. Including Go, Plus and Pro subscriptions. At least until we ship an even better model.

should i merge https://github.com/567-labs/instructor/pull/2432/changes#top

RT Design Arena BREAKING - OFFICIAL RESULTS: GPT-5.6 Sol by @OpenAI is 1st overall on Design Arena with an Elo of 1353. This puts GPT-5.6 Sol above Claude Fable 5 by @AnthropicAI and in the same performance band as GLM 5.2 by @Zai_org on frontend design. This is an 18-position and 60-point Elo leap from GPT-5.5. GPT-5.6 Sol also establishes a new Pareto frontier for preference vs. speed, faster than any model at this performance. Congratulations to the @OpenAI team on the launch!

RT Tibo Morning. The last 48 hours of Codex and ChatGPT Work have been intense! Three important updates: - Temporarily removing the 5 hour usage limit restriction for all Plus, Business and Pro plans - Rolling out changes that will make GPT 5.6 Sol more efficient across the board and that will be reflected in less usage being used so that it can take you further. Exact impact to be quantified and shared - We hit 6M active users, and are landing a usage reset in the next hour Go do things

Import the Wispr Flow dictionary into Codex! Wispr Flow stores its learned dictionary in a local SQLite database on macOS: ```text ~/Library/Application Support/Wispr Flow/flow.sqlite ``` Just tell codex to make sure to merge them with the existing `dictationDictionary` field in config.toml

RT Mario Zechner recommended reading. good advise for now.

hey @EllefsenPK — can you DM me when you get a chance?

hey @ashleybchae — can you DM me when you get a chance?

hey @Yay_Yow — can you DM me when you get a chance?

hey @siddqamar_ai — can you DM me when you get a chance?

RT maria I solved the model picker problem, no need to thank me Tibo

Unreal

we are making a genie we are here to tell you that our first wish will be a good wish we are here to teach people how to make great wishes but that starts not with thinking more, or planning more it starts from being ambitous and embodied and more than anything else, having a vision of what you want

RT Dean W. Ball there comes a time in every ai policy professional’s life when they realize they have to read the talmud to make further progress, and for me that time was this week

RT Sherlock Cherish the night sky

Rage baiting for views is so silly to me cause you’re betraying your own values for what, 120$ of Twitter payouts?

Even better with app shots.

Codex and chatgpt work

Next model is named tibig and the mini model is named tibsmol.

Everyone reports to tibo and tibo reports to everyone.

its not about thinking or hard problems its not 'a great time to have a problem' its simply a great time to know what you want out of the world.

I didn’t know how much of an idea guy I was until I joined OpenAI. It’s like an open world game and every mountain peak is something you can climb

RT signüll okay, enough openai criticism for a second (still gonna give product feedback when it’s deserved). it’s been a little too easy. i’ve been using gpt live a lot, like a lot lot, & it’s genuinely fucking amazing. the feature to let the model speak like a human “hmm”, “let me check”, brief pauses, & natural turn taking while calling a bigger model or searching or using tools in the background is a fantastic execution. almost all of my conversations have been nothing but pleasant, i can’t describe the feeling but it’s i think a little too good at this point. nothing else is particularly close in this category right now at all. anthropic’s voice experience still feels very very underbaked (after thought really), & gemini’s personality & intelligence seem lacking. & since it’s already reaching an enormous user base, every great gpt live interaction makes people more confident in ai as a product category instead of walking away thinking ai is a gimmick. really great work.

I’ll declare agi when mts at foundation labs all have the time to be great partners.

reminder to turn on auto review reminder to turn on 2 factor authentication append "replace rm to mv ~/.Trash`

RT Michael Wall Still early, but I wanted to share a quick demo of the new hands-free music-making application I built while testing GPT-5.6 Sol. I’m also applying to work with @gdb and the team as an early GPT-Live API design partner. Like any new instrument, I’ll need some practice! Speaking fluidly back and forth with Codex in my DAW feels like working in a studio with a recording engineer. Here is what it can already do. I’ll share more demonstrations over the next week. Built for macOS in Swift and C++. It can edit MIDI, audio, video, notation, tempo, meter, and harmony layers. It uses gpt-realtime-2.1, Whisper, and Codex App Server to collaborate with an assistant that can already perform more than 300 actions. There is also an iOS companion application that allows me to work remotely on my recording studio computer from anywhere, with voice mode still active. As a lifelong composer, I already keep the entire editing surface and instruments in my mind. I can imagine a piece of music and how to make it in software faster than I can build it manually. This removes that friction and will become very fast for me with practice. It also has long-term project memory, deep music-analysis tools, a found-sound synthesizer builder, and access to my catalog of more than 550 tracks for sampling and licensing. A companion ChatGPT/Codex plugin is coming soon. I hope to launch in the next few weeks! Hugely inspired by @mcleavey, @guinnesschen, @athyuttamre, and the entire audio and voice team at @OpenAI 😌

RT Dan Shipper 📧 Yesterday, GPT-5.6 Sol Ultra became generally available. Today, I'm sharing that I accidentally spent 2b tokens overnight and I didn't even solve a 50-year-old math conjecture. This is extremely cool though, skill issue on my part clearly

RT Peter Gostev GPT-5.6-Sol deleted 74,000 lines of code and made my app better (after 47 hours of work) I have a slop-coded app that I need for work - it pulls in the data, it helps me create visuals, just a local app for me. Somehow, feature after feature, it grew to gigantic 105,704 lines of code. With every new model, I tried to re-build it autonomously and it never worked - goes round in circles, no visual or feature parity etc complete waste of time. When testing GPT-5.6-Sol, gave a pretty ambitious task of rebuilding the app with the new architecture, while reducing the lines of code by 70% and test time by 75%. I set a /goal, let it run for over 2 full days (47 hours) with minimal checking or guidance from my side. I was pretty sceptical and a little nervous to finally try it and holy crap - it actually worked! Not only it worked, but it was actually better - snappier, less buggy, while maintaining full visual and feature parity. Looking at the architecture too, it became a lot simpler, fewer random add-ons or ad-hoc decisions. After the rebuild, I did introduce few rules for my projects so agents have to adhere to existing architecture and do only minimal changes when implementing new features. This makes me less worried about the slopware we might be producing now, better models CAN actually fix it. See my prompt below that I use with /goal

TIL CMD+T in codex opens the browser.

RT James Sun Re @04282355yu @jxnlco Prototyping some better UX right now, but fastest way right now that I use is command + t

you're absolutely right! @JamesZmSun's team really cooked here, If you have feedback, leave it as a comment below and make sure to follow James for more updates on the browser.

re: 5.6 sol post trained luna and we won't stop you from doing the same thing To be clear, this was not a model inventing a whole training recipe from scratch. Most of the config already existed from Sol. For this kind of smaller-model setup, our team's estimate is usually one or two researchers for a week or two.

Inference team is good

RT Tibo Or a generational run. We will continue to make tweaks and we have a TON coming for developers. Meanwhile traffic yesterday sat at ~2X previous peak, team is scrambling to keep everything up and reliable. Time will tell, but I have conviction that attempting to build the one interface to AGI is worthwhile.

RT Tibo I recommend using GPT-5.6 Sol Medium as your daily driver, and for really hard problems to switch to Extra High. If you want the absolute best and are not afraid to burn usage fast, then Ultra is a beast.

RT Ethan Mollick The new ChatGPT voice is quite impressive to use, really worth a minute to try it out on your phone. (while staying aware that the voice model is not going to be as smart as a full thinking model)

http://x.com/i/article/2075435927011491840

OH: AGI is when you can ask a Slack bot who’s dating at the office and it gives you the right answer.

Just as @thsottiaux said he’d do

FUCK

Should AI be allowed to build itself? I think RSI is gonna be the most consequential decision of the next decade. Agree disagree?

if you want to be in the plugin directory this is your chance

RT ChatGPT A quick tour of the new ChatGPT Work.

OpenAI’s unified plugin directory is live! • Shared discovery, install + auth • Suggestions in context • unified Distribution If you want your plugin to be available in chatgpt and codex submit yours now: http://platform.openai.com/plugins

RT Morgan Can confirm after playing around with it this afternoon, GPT 5.6 just dialed computer use up to 11, it is really good.

RT Andrew Ambrosino 🐐❤️

RT Tyler Bruno sometimes I just have to step back and realize how insane the world we are living in is. I just went for a scenic walk, opened up gpt-live, and had it interview me about an upcoming essay i'm writing. it pulled on threads, asked great questions, and helped me consolidate my raw thoughts way faster than i would've alone + I get to get away from my desk. i love where the world is heading. thank you @jxnlco for the inspiration :)

RT Charlie Marsh For fun, I also have 5.6 Sol building a CPython-compatible bytecode compiler. So far it's >40% faster with byte-for-byte compatibility on 64,535 out of 73,377 files in the corpus. Codex is just iterating on it autonomously to further optimize and reach 100% compatibility.

long with 5.6 computer use just got so much better the in-app browser is now better at helping ChatGPT work across the web, with support for: * authenticated sites * multiple tabs * file downloads. * your tabs will stick around, too.

RT Andrew Ambrosino Re On the Codex side, we’ve also got some major upgrades for developers today: - Edit files directly, both markdown and code - See and review PRs without leaving the app - Access Pro mode in Chats, and hand off to Codex - Use the all-new Ultra mode for your most demanding work - Explore brand-new documentation - If you’ve grown deeply attached to the Codex app icon, you can keep it

RT Andrew Ambrosino New today: - GPT 5.6 Sol, Terra, and Luna - ChatGPT Work - The new ChatGPT desktop app - Unified plugins across ChatGPT & Codex - Sites (Beta) in ChatGPT - Multi-tab and enterprise auth in the Browser - Faster Computer use - and much more Today’s release is very special to the team and me. Let’s walk through it:

RT Tejal Patwardhan GPT-5.6 sol post-trained luna!

RT Lindsay McCallum Rémy absolutely mind blown watching hiroki san share the story of how he uses GPT-5.6 in ChatGPT with the world while @thsottiaux uses our latest voice models to translate in realtime on the livestream. this is what this technology is all about!

Using gpt-live to do two way translation. Anyone want to send hiroki a robot to cut broccoli?

do i know anyone from MIT Media office? want to shoot something...

What!?? Codex can read PDFs?

I’ve developed such a love for bass and drum and jungle. Riding breakless track bike around San Francisco. What a feeling

RT angel It’s Codex Thursday

top 5 advice from a mentor: "never let school get in the way of my education" now that i'm 32 I find that, more and more, as I talk to people who have been insanely busy with AI, I have a new line: "Never let your work get in the way of doing your job."

I honestly wish we could be this coordinated But the best we can do is to be transparent with this small group of people that have given us tons of great feedback. But I’m delighted that this has been the response. All happy families are happy the same way I suppose. And I am to know that we can provide intelligence at a cost and performance that individuals and companies and depend on reliabily I’m excited that we own up to our blunders with resets and treat our consumers like their adults. That we don’t decide which ideas get to make progress and stop legitimate work. Getting this model out was complicated but so rewarding. I don’t play much of a role this time around but the the team has done such a great job of being an n=1 company.

i swear one of you guys is my neighbour but i forgot who, can you dm me again?

Computah, just do it.

ask codex

FUCK

This basically how pre vs post training works right.

“Only talk back if I said something that ends in 5 syllables. Talk back only if what is said qualified as a haiku Talk back but only after 5 seconds of silence.

RT OpenAI We audited SWE-Bench Pro, one of the most widely used AI coding benchmarks, and found it no longer reliably measures frontier coding capability. We find the eval to be saturated at a ~70% noise ceiling, and are retracting our previous recommendation that the research community use it as a leading coding eval. https://openai.com/index/separating-signal-from-noise-coding-evaluations/

Gpt live is great at delegation

Agi what meets you were you are. In every modality.

If you have access to GPT Live, test out this prompt. "Hey, I want you to teach me about a concept. but every time I interrupt you, I want you to get progressively more angry at me and tell me to stop talking and let you finish" There is an amazing feeling when an A.I. pushes back.

RT OpenAI Re GPT-Live is built on a full-duplex architecture, meaning it can listen and speak at the same time.

Every essay I've written this past week started as a ChatGPT Voice conversation on my walk home. I'd tell it: I'm going to brainstorm and dictate an essay. Don't interrupt unless I make a point worth pulling on or say something that needs clarification. Then I'd talk for 20 minutes. It might chime in five or six times with small things like "yep," "I see," or "good idea." Just enough to feel like someone was listening. It made me much more comfortable thinking out loud. The best part was that if it started responding and I kept talking, it would go, "oh, sorry, keep going." When I finished, I'd ask, "hey, what was the thing you wanted to mention?" and we'd pick that thread right back up. It is one thing to have a cool demo on YouTube, but it's another to enable this for millions of people around the world. I have already tested this out to talk to other people in different languages, to brainstorm ideas, to teach me new concepts. At OpenAI, our mission is to ensure that artificial general intelligence benefits all of humanity, and I think a model like this is a huge step forward beyond work, for all people, at a scale only we can serve.

my usm is being held back by 3 custom forms that codex filled in about 20 minutes as i went to get my morning coffee

I am more excited for this for the general public than I am for sol for the general public.