This literally just landed for me and holy moly, such a little difference when you first look at it but such a huge difference in terms of context management in practice. The fact that it predicts when having multiple chat threads coordinating with each other is what got me, that's super impressive context relaying.

we made a simple thing, people internally *really* like it, and we hope you like it too the little composer that could, from planning the launch with @sharifshameem and @andhudhow:

So I tested GPT-6.1 Sol on real work: 2 repos, 105 planted bugs. Find and fix what you can. Unlike GPT-6 Sol, which was just a nerfed GPT-5.6 Terra, this one is real. The results: - GPT-6 Astra (max): 45 for $33 - GPT-6.1 Sol (max): 44 for $6.56 - GPT-5.6 Sol (max): 43.5 for $95.35 - Opus 5.5 (max): 41.7 for $58.53 - GPT-6 Sol (max): 29.3 for $9.33 With a model like this, the 50% cut to the $200 plan doesn't matter. n=1. More runs and results at more effort levels dropping in this thread over the next few hours 🧵

Just went for a long walk while using ChatGPT Voice with plugins on my phone, and I think it's a more consequential update than it might seem at first glance. On my walk, I talked through my calendar, email, and to-do list. I had ChatGPT read through my emails and triage them. Then, we added items to my to-do list. Then, we reviewed my calendar. By the end of the walk, I'd finished getting organized, and never had to look at a screen once. I've been waiting for this a long time. Yes, you can kind of do this in the ChatGPT desktop app. And I think in Codex remote on the mobile app. But I've never gotten either to work very smoothly or reliably. And I often want to do this when I'm not near my laptop, like when I'm out for a walk, and the ChatGPT desktop app for whatever reason won't connect. This feels like the future. It feels like a true assistant. You have a natural conversation, and it does stuff for you. It disappears into the background. You're not walking down the street with a screen in your face. You're not contorting your thumbs to type sentences and scroll menus. You're just talking, conversationally, and stuff's getting done. My sense is: • Screens are going away for many uses other than visual content consumption. They won't disappear entirely. But they also won't dominate the way they do now, and definitely not for things that can be done verbally. • Apps that aren't agent-native and plugin-optimized will die. I use Google Workspace and Todoist primarily because they're so well-integrated with ChatGPT. I stopped using apps, such as many of Apple's, because with few exceptions they're poorly agent-integrated. (X is a big exception, but I've hacked some integration via IFTTT.) • Those apps, however, will become more valued for what they let agents do than for their in-app experiences. Not all will become essentially databases, but many will. It also seems inevitable to me that at some point OpenAI and other AI app providers roll out native calendar, email, and to-do functionality that's maximally agent-optimized. • We'll have our audio agents on all the time. We may have them on mute sometimes (I discovered you can do this with ChatGPT Voice by clicking an AirPod), but at great cost, because you'll want be able to just say something like "summarize that conversation I just had and then draft the presentation." It will be like having an assistant following you around when you want them to. • We need new hardware. For example, I want a camera in my AirPods not to take pictures, but to have the context of what's around me. Like, "What kind of tree is that?" Or, "That's a cool shirt. Find out where they bought it." (I have this in my Meta Ray-Bans, but people are now very sensitive to those devices because they do take pictures.) I also want ChatGPT able to be always listening or awoken with "hey, Chat." • The individual thread model starts to break down when you're treating ChatGPT like an always-on assistant. We need to have the infinite continuous thread from which ChatGPT can delegate work to subthreads, the central assistant model that Muse is built around. This makes even more sense when you start treating your AI apps like assistants. They should be juggling all those individual threads on your behalf. More to come I'm sure as I continue to experiment.

One of the most underrated features in the ChatGPT desktop app: Appshots. Appshots take the context on your screen and bring it into the desktop app, allowing you to share instead of describe. You’ll be surprised at how useful it is. To get started, press both Command keys on macOS, or both alt keys on Windows.

this if you really want to use a multi-agent setup, ask a side chat to audit your usage patterns. for example, having your main agent repeatedly message the same subagent is unnecessarily expensive. the context keeps growing and wastes input tokens even when cached if your main agent keeps asking a subagent to call a tool, the delegation is unnecessary. just let the main agent call the tool astra is smart. send it this tweet and ask it to audit things

I would like to clarify a few things: 1) The screenshot is my reaching out to Levent to coordinate our releases. I hope it’s clear from the message that we came in with the best possible intentions. 2) I never ever asked for Levent to be removed from authorship of his own work (as indicated by my text). I was surprised to learn during the call with Tristan that they had only solved Euler and not Navier-Stokes; after learning this we brainstormed possible paths forward. One option we discussed was that Tristan could be the lead author on a rewrite of OpenAI’s Navier-Stokes proof. It is in that context that I said “it would be simpler if Levent was not an Anthropic employee” because I felt it would be inappropriate for an Anthropic employee to author OpenAI’s work. Importantly it was admitted that internal Anthropic models had been used in their proof of Euler blowup; I therefore felt I could not consider Levent to be an independent academic. Another option I wanted to propose (but got cut short) is to offer access to our internal model so that they could try to finish their proof and bridge the gap between Euler and NS. Again I did not know how to navigate giving access to internal OpenAI IP to an Anthropic employee. 3) To reiterate it plainly: as my text clearly indicates, and as I said during our call, OpenAI's intention was to do everything possible to celebrate their mathematical achievements and the heroic efforts that they made on Euler. In the call I was immediately met with a litany of slander, including direct threats that if we were to announce Navier-Stokes he would immediately go to the press with a barrage of unfounded accusations. I refuted all these accusations but he replied “there is nothing you can do, I simply do not trust you”. I was confused why one would turn an incredible source for celebration (of their achievements!) into such bickering, which is when I said that I did not understand why one would risk their career [over unfounded accusations]. Genuinely, at that moment, I was trying to care for him and do a last ditch attempt to get a chance to give them all the credits that they deserve. I deeply apologize for this extremely poor choice of words, it is the opposite of what I was trying to convey. (I should say that I retracted them on the spot by the way.) 4) Overall, on a personal level, it was incredibly difficult to have these conversations. Levent refused to attend any of the meetings despite my repeated asking. As Sholto Douglas said, there will need to be coordination between Anthropic and OpenAI in the future; I felt I was doing a proxy negotiation with Anthropic while the Anthropic employee refused to directly participate.

A series of false and inflammatory allegations against me are currently circulating on social channels. To clarify, I came into the discussion following academic norms, and I'm disappointed that it has come to this. Anyone who knows me knows that academic standards are of the highest importance to me. Will have more to say tomorrow.