Morphexe

@RealWilliamHsu @ben_ai_eng my man I have been using medium for a while now , and absolutely demolish my 20x limits in a day in a half

@ben_ai_eng Same here, I even switched to Sol High instead of xHigh.

Is anyone else burning through Codex 5x Pro weekly limits unusually fast? I used two full weekly limits in one day doing normal engineering work, mostly on GPT-5.5 Medium/Low and GPT-5.6 Sol Low. I tested older models to see whether Sol was the issue, but GPT-5.5 seems to drain limits just as quickly. For comparison, I’ve used Claude Max similarly for a week and only consumed about 20%. Feels like there could be a compaction or usage accounting issue, or the plans are simply structured very differently. Anyone else testing or seeing this? 👀

@min_aws @unity I was soooooo happy when I read this.... :(

@unity I read this as we are making our Editor and engine source available... Guys you won't get the vibes.

We’re opening the Editor to EVERYONE 🤝 For 20 years, the Unity Editor is where a whole generation of games has taken shape. But a game isn’t built in the Editor alone. That’s why we’re building an API and set of services that open a Unity production to the tools, roles, and agents around it - no full Editor access required. Preview the real game, automate from the terminal, and use whatever tool fits the job. Now available in closed beta.

@deanwball This is such a bad take on the subject is suprising coming from you.I thought you were actually praising. The rest smells like fear, because it seems almost on par, clearly lacks some RL, but I am fully expecting for them to get super competitive soon.

Some observations on Kimi: 1. It's a very good model! I don't think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also seemed very token hungry. It's not obvious to me that this model is actually that cheap to run. 2. I am personally surprised the Chinese state continues to allow the open sourcing of models this good, given potential risks. To be clear, I *myself* might be fine with models presenting this level of marginal risk being open weight, but I am surprised that China is fine with it. I suspect the reason they are is 75% explained by strategic blindness/lack of AGI-pilledness (the CCP is very Yann Lecun-y in its views of AI). The other 25% or so is their lack of compute for customer inference (making China's open-weight strategy an unintended byproduct of US export controls) and the normal Chinese strategy of aggressive exports. For the companies, as opposed to the government, the decision to open source is partially ideological and partially because they are behind, and they know that very few people would pay for sub-frontier models from China. 3. Open-weight models are inherently decelerationist, and I'm continually surprised to see the so-called "accelerationists" so excited about open-weight models. I suspect the reason they are is that they know open-weight models are effectively ungovernable, and they simply like the overall cloak of ungovernability open-weight models create over the whole of AI. It's not a bad strategy; it reminds me of James Scott's recounting of the hill people in "the art of not being governed." Still, in the end, open-weight models deter further AI capex. 4. One probable outcome of an open-weight-model-dominant world is full AI communism, which is precisely what China proposes: rather than a market product, AI is a "public good" which will ultimately be provided by the state as a kind of "digital public infrastructure." This future strikes me as a dystopian hellscape, but I've never met an open-weight models advocate who doesn't ultimately concede this is where things end. You'd be surprised how many 'accelerationists' lobbied me, while I was in government, to support an eleven or twelve-figure federally funded data center so that startups could train models at a subsidy and then give them away for free. There was no other way for AI to progress, they said. Perhaps this is the logical end state of things. Nonetheless, I find myself surprised to see supposed accelerationists excited about such an outcome. I think many of them just don't know what they're doing. Many accelerationists do not view the creation and serving of frontier models as a legitimate business. 5. I would guess that the Trump Administration will at some point realize that their best strategy here would be to create large amounts of regulatory risk around the use of open-weight Chinese models. You don't need to "ban open source" (one of the dumber motifs of AI policy discussion). You just need to direct every agency to issue soft law that creates FUD. "A Federal Reserve Advisory Bulletin found that there may be backdoors in Chinese AI models." It needn't be that well justified. You just create enough regulatory risk that every regulated enterprise backs off. You probably don't want to create so much regulatory risk that you scare off the hyperscalers from serving Chinese models; this will just drive startups to sketchier providers. There's a happy middle ground here. I'd assume they will do some version of this. 6. It's probably true that open-weight models of this capability make the world a bit more dangerous, but not so much more that you'll really notice. At some point the models will be capable enough that you will notice. "A nonliving, invisible, dangerous, and infinitely self-replicating agent escaped from a Chinese lab," you say? Color me shocked.

@thsottiaux Computer use absolutely locks windows. if chatgpt is using the computer and you take control, just pause for a bit. Its not the first time I had to reboot the machine because it got stuck and no mouse or keyboard worked.

How can we make ChatGPT better. It’s already on your phone, but what do you wish would happen more or less of when you open the app. How can it bring you more joy?

@elonmusk Cant wait for you to get your own robot there :D

Robot fights are fun 😂

@thsottiaux So what?? Am I running Codex 5.6 Locally now ? This is a 4090RTX btw lol Maybe you need to talk with @theo and @jullerino To fix the Rendering performance on the app :) @OpenAIDevs Any idea?

@amankumarhq And this is exactly what you have with subscriptions. DIscount they use your code. API prices your code is private :)

Claude and Codex should offer a two-tier pricing model: 1. Full price = your code stays totally private 2. Discounted = you let them train on your code

@hotstoneregent @thsottiaux Same lol

@thsottiaux I just used the reset card.🤡

Oops... I did it again. Enjoy reset usage limits for all paid users for Codex and ChatGPT Work. Super grateful for an incredible team who is iterating at lightspeed and keeping the infra up as we scale faster than ever. Enjoy the weekend!

@thsottiaux @SamiFathi_ @OpenAI @sama I NEEED SLEEEP ....

@SamiFathi_ @OpenAI @sama Will give signs of life soon

I am getting seriously concerned about the health and future of @OpenAI @sama and @thsottiaux. It's been over 24 hours since a reset. Are we okay?

@thsottiaux please leave this offfff! This is by far the best thing ever :D If people struggle with hitting limits, just make this toggleable per account/user. This is fantastic.

We've had no 5h limit in Codex plus and pro for a few days. Do you think it is better or are you finding it difficult to manage the usage included in the weekly limit effectively? If we were to make this different, what should it look like in an ideal world?

@MyMateMorris @mweinbach According to him he is on the 40$ sub. so not that much

@mweinbach Wow. impressive! What does your 60% monthly Kimi usage mean in real dollars used?

It finally finished, here's the final output. Used 60% of my monthly Kimi usage on it macos27.kimi.page

@theo I just blew the 100$ sub in a stupid couple of tests, but its actually good.

What are the best harnesses to use with Kimi K3 right now? It heard it's pretty good at orchestration, where's the best place to see that?

@kalin_t Any tech posts about how you achieved this ?

from up high you cant really appreciate how much distance is being covered, but from down at ground level you can feel it what should with do with this infinite world? single D&D adventure sandbox?

@theo @bedesqui As it should be....

@bedesqui The transparency here implies you’re not reading any of the code, respect

if this REAL 500 lines workflow finishes well in pi @theo must let ben make a podcast episode about pi yes, it took almost 10 minutes to write the workflow itself.

I have a confession to make... I have been struggling to use all my limits in CODEX the last few days... I think I am going to have crashout when this bonanza stops @thsottiaux

@theo Because.... I bought two machines to run remote agents just like a youtuber told me.... And, I will be going on holidays in a few weeks, and tokens need to be burnt.

T3 Connect and the T3 Code mobile app are ready for a small wave of testers. Reply below with why I should invite you to test

@theo @thsottiaux Just go for t-codex :D

@thsottiaux I like how you think

Now we're getting somewhere

@sandislonjsak Yes, Terra is great to do research on your code base - and on mid / high does fantastic implementations. Luna is great for one ofs, summaries, and other simpler tasks. knowing when to use them makes wonders

Amazing how OpenAI released Sol, Terra and Luna and no one cares about the last two. I didn’t even bother trying them out. Have you guys even tried them?

@kalin_t I am interested in the rendering, are you meshing the chunks, or doing raymarch / lattice ? Just rendering code, or collisions as well ? Any tech details? would love to know what it used to do this :D

some more terrain variety thanks gpt-5.6-sol xhigh pure C code, all CPU side generation, barely any CPU load here

@CommandCodeAI I could give a shot - would love to compare it :D

Who wants access to Kimi K3 model?