Dan Crewger

@waiting4agi @tszzl What's your set up?

@tszzl Seriously? I do everything through phone only

people are walking around with their laptops slightly ajar to keep their agents running

@simonw amazing interview with @lennysan! I'm a Product guy, and this was hands down the most entertaining and informative piece of content by a Tech guy (or girl) I've consumed.

@thsottiaux Not being able to scope chats to projects/workspaces. See github.com/openai/codex/i…

It’s the little things that matter, what are some small papercuts you have noticed in Codex? We’ll fix as many as possible in the next week.

@Tommy_fjc @programmerByDay @AIandDesign @vercel @rauchg Cool I didn't get either email

@programmerByDay @AIandDesign @vercel @rauchg This email to customers which I also got:

Dear @vercel, Do hackers have ALL MY SHIT and do I have to spend all night fixing it or not? I did not receive an email. Can we please have some TRANSPARENCY here?

@trq212 Great idea! Will be an amazing resource for many! I'd love you to do some with people who have high level tech understanding but no coding skills (typical PM profile)

I want to do some streams where I work with non-technical people using Claude Code to figure out how they might be able to improve their process. My feeling is that just a few tips could make a big difference in efficiency. Any mutuals interested?

@CRSolomonWrites @OpenAI A lot of people asked for this. The fact that you can't imagine something not important to you being important to someone else fits the keep4o theme very well.

@OpenAI Nobody asked for this. Just stop killing the models people love. #QuitGPT #keep51 #keep4o

We’re updating our ChatGPT Pro and Plus subscriptions to better support the growing use of Codex. We’re introducing a new $100/month Pro tier. This new tier offers 5x more Codex usage than Plus and is best for longer, high-effort Codex sessions. In ChatGPT, this new Pro tier still offers access to all Pro features, including the exclusive Pro model and unlimited access to Instant and Thinking models. To celebrate the launch, we’re increasing Codex usage for a limited time through May 31st so that Pro $100 subscribers get up to 10x usage of ChatGPT Plus on Codex to build your most ambitious ideas.

@lydiahallie Why don't YOU ship a fix that sets the compact window to 200k? You must be losing subscribers left and right. This is in your hands. A large portion of users won't read this message.

Digging into reports, most of the fastest burn came down to a few token-heavy patterns. Some tips: • Sonnet 4.6 is the better default on Pro. Opus burns roughly twice as fast. Switch at session start. • Lower the effort level or turn off extended thinking when you don't need deep reasoning. Switch at session start. • Start fresh instead of resuming large sessions that have been idle ~1h • Cap your context window, long sessions cost more CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 We're rolling out more efficiency improvements, make sure you're on the latest version. If a small session is still eating a huge chunk of your limit in a way that seems unreasonable, run /feedback and we'll investigate

Thank you to everyone who spent time sending us feedback and reports. We've investigated and we're sorry this has been a bad experience. Here's what we found:

@westerbamos @corbucorbu @lydiahallie Lol well he doesn't clear context windows because he doesn't have to pay...

@corbucorbu @lydiahallie And they were pushing it pretty heavily… x.com/trq212/status/…

@UzoPamela1 @lydiahallie Have the previous one produce a summary, feed that in. No other way. If Claude needs to know about something, that something consumes tokens.

@lydiahallie If I start a new session, then it's a whole new conversation. Is it possible to do that while maintaining the flow of our previous conversation but in a new chat?

@lydiahallie Seems I have been doing a lot of this right, probably why I never hit the limits in unreasonable ways. However, on reddit users report a fresh chat with "Hi" ate sth like 11% with hardly any skills or MCPs. That can't really be explained by anything you shared can it?

@rohanvarma @dakshgup Don't forget to dogfeed!

@dakshgup Will ensure this user need doesn't go unmet

If we made /slow mode in Codex, would you use it? What for? (Slower inference at a cheaper cost)

@trq212 What I would LOVE is a combination of this and Codex Web. Having to run a mini pc at home feels very outdated (at least for coding).

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

@trq212 I went through the setup this weekend, let Claude guide me through it. Took forever, it felt like Claude wasn't fully aware of the guide and kept bumping into walls. Great feature though now that I got it to run. Already looking for mini PCs.

This is literally what working with Claude Code feels like at times. I think I need to get better in scoping permissions.

@nlw Love the show, but I disagree on something from the recent AI and jobs episode: efficiency and profitability is not a means to an end (fulfilling human desires) in shareholder capitalism. On an idealistic level, yes, but that's not how the system is actually set up.

@LolOut160461 @donaldtusk Because EU pays more 😉

@donaldtusk It’s getting ridiculous at this point. We’re sitting at the table trying to protect European interests while someone is literally live-texting the Kremlin from under the desk. If you’re that cozy with Moscow, why even stay in the EU? It’s pure sabotage

The news that Orbán’s people inform Moscow about EU Council meetings in every detail shouldn’t come as a surprise to anyone. We’ve had our suspicions about that for a long time. That’s one reason why I take the floor only when strictly necessary and say just as much as necessary.

@svpino @dave_alive Does /loop work locally? I would have assumed it runs in the cloud so it doesn't matter if your computer is running or not.

@dave_alive I didn’t ask to schedule a task on a data center.

Claude is retarded. All of these models are. I wanted to schedule a skill every day at 8:00 am. Claude decided to schedule it at 7:57 am "to avoid the on-the-dot" surge. I SAID 8:00 AM! Do the darn thing the way I asked you to do it! You gotta be crazy to trust these models.

@levelsio Honestly I'm perplexed by the webmcp hype. I'm a non-dev so maybe I just don't get it, but wouldn't it be much leaner to just embed proper API documentation for the APIs the website is already using, rather than implementing a second way to achieve essentially the same thing?

Thank god MCP is dead Just as useless of an idea as LLMs.txt was It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs

@steipete Meanwhile, all my codex conversations are stuck in thinking mode for the past hour.

Oh right i realize hell freezes over, we reached the point where app > cli That combined with the speed increase also means less windows necessary, codex goes brrrr now! developers.openai.com/codex/app/

@peterthiel wrote this 12 years ago. Shows how crazy fast the world is changing, if even he couldn't fathom us being where we are now, a mere decade ago.